Bezos AI Play, Future of Airports, CAA Fund | David Reger, Mike Wior, Ariel Cohen, Jeff Tatarchuk, Matt Joseph, Ade Ajao, Jeremy Fraenkel

Let me tell you about Figma.

I pushed the Figma button.

Agents, meet the canvas. Your AI agents now create and modify your Figma files with design system context. You're watching TVPN.

That's a happy accident.

Today is Thursday, June 11th, 2026. We're live from the TVPN Ultra Room, the temple of technology, the fortress of finance, the capital of capital. Let me also tell you about ramp.com. Time is money, save both. Easy to use corporate cards, bill pay, accounting, and a whole lot more all in one place. Gotta adjust my microphone over here. Well, Bezos is getting very serious about the AI game. $12 billion raised for Prometheus. It's just been 6 months since he emerged from stealth with $6.2 billion in funding. The industrial AI startup, as they're calling it, raised another $12 billion in Series B. It's just 1,000 times bigger than a normal Series A and a Series B. You used to go from $6 million Series A, $12 $1 million Series B. Now it's $6 billion Series A, $12 billion Series B. Serious dollars, but AI is expensive. The valuation, $41 billion. Back in November, there was very little disclosed about Prometheus. We wrote about it a little. There was some speculation on the time.

$41?

$41 billion. Bigger than— what's a $40 billion company?

And this is a business that is Buying existing companies.

What? That's the thing.

We don't really know.

We don't really know. But it's a lot of money. Could be buying companies and trying to transform them. Are around $30B to $40B.

Well, being able to raise that much capital to buy businesses with that little dilution is a pretty remarkable feat that pretty much only Jeff Bezos could pull off.

Yeah. So the Wall Street Journal has some more details and Mostly it just gives you, takes you a little bit more inside the mind of Jeff Bezos and where he's thinking, like how he's thinking about the shape of the AI impact. The headline is very clear. Bezos bats down AI job loss fears while launching new venture. New startup Prometheus seeks to build an artificial general engineer. AGE is the new buzzword. You thought we were at We thought we were running out of them. We got AI, AGI, ASI, RSI, AIFDE, personal superintelligence, meta superintelligence, TBD, continual learning.

Standard intelligence. Yep. Irregular intelligence.

Non-standard, ineffable intelligence.

The irregular intelligence, the irregular superintelligence company of the dog patch.

A dog? Something like that. But artificial general engineer is the new phrase, the new term. And he wants it to be able to design and manufacture complex physical products. So Jeff Bezos expects artificial intelligence to create a labor shortage in the economy. Interesting. There's some correlation between like liquid net worth and your projection—

And your optimism.

—of employment, right? Yeah. So he expects a labor shortage, rejecting fears fears that fast-evolving technology will put humans out of work as he launches a new venture aimed at making engineers more productive. The billionaire founder of Amazon.com, and don't forget Blue Origin, is co-leading a new AI business called Prometheus, which plans to build an artificial general engineer that can design and manufacture complex physical products such as a jet engine. And we just saw some cool demos. I saw people vibe coding a Boeing 747. Obviously it's just the outside. It's modeling it. It can pull different reference points from CAD files on the internet. But lots of cool demos, many powered by Fable 5. I've seen some GPT-5 stuff as well, where people try and go and see the classic pelican on a bicycle in SVG. Well, you can see how you can draw a line from pelican on a bicycle in SVG to functional jet engine in CAD that then we can actually go manufacture with step-by-step instructions for each part. Not that far away, certainly on the trend path, and that's what he's investing against. He said in an interview that the company's goal is to, quote, empower engineers and make innovation easier and faster so smaller teams can do much bigger things on much shorter time cycles. That'd be good. It takes forever to build a jet engine. It takes forever to build a plane. We've talked to multiple founders that have been trying to build new plane designs for decades and haven't really been able to scale them up, although they've made incredible gains and traction and demos. Actually getting new planes on the runway safely approved has been far too long. If AI speeds it up by 10%, that's like years. It's crazy. The Amazon and Blue Origin co-founder joins a group of longtime technology leaders that have jumped into the AI fray. You got Uber co-founder Travis Kalanick, Coinbase co-founder Brian Armstrong, Robinhood co-founder Vlad Tenev, Also Brian Chesky over at Airbnb. These folks are all building new companies around the AI boom, among others. Prometheus, which said it is valued at about $41 billion, has raised $12 billion from investors including Bezos, JPMorgan Chase, Goldman Sachs, and BlackRock. Went straight to the big banks. Yeah. Like, the invest— VCs are just like, yeah.

Well, yeah, I mean, there's another world where doing 6 On $40 billion is just like your IPO.

Yeah, totally. Yeah, that is IPO numbers. That's, yeah, that's a crazy amount of money. Yeah, you got to go to the banks for this. They've hired 150 people across San Francisco, London, and Zurich. Interesting. Despite Silicon Valley's enthusiasm, enthusiasm for AI's promise, the broader population has become very wary, fearing widespread job losses. Some of the pessimism about AI among young people is the opposite of reality, Bezos said. And we've been seeing this in the jobs numbers, like the The economy is running hot, inflation is high, but also unemployment is very low, lower than expected. We're adding a lot of jobs. AI will mean fewer workers are needed in jobs that exist today, but it will also create far more opportunities and increase productivity, he said. That's a very white-pilling statement. We'll see how it plays out. Bezos argued that if AI makes it cheaper, faster, and easier to invent things, employment will ultimately rise because even though you're shrinking the number of people needed by 10x, the technology will create more jobs. More than 10x as many opportunities. And you certainly saw that with Amazon, the number of store workers in physical retail stores probably shrunk, but the number of entrepreneurs building products, selling them on Amazon, in and around the Amazon ecosystem, not just in the warehouse, but in the broader Amazon ecosystem, probably increased. And we certainly saw that in the broad job numbers in the United States. Like, the labor market has grown unequivocally since 1999. We've added jobs all over the place. As people have shifted around. There's going to be two— there are going to— there's going to be two-earner income households where one earner drops out of the labor pool because there's going to be so much productivity, he said. That's an interesting take. He's saying that there's going to be so much more productivity that some people will choose not to work. And, of course, that won't necessarily show up in the unemployment rate because the unemployment rate only measures people that are actively looking for jobs. So you could have a world where the workforce is shrinking in terms of the number of people that have typical jobs is declining, but the unemployment rate is also very low because you have more single earner households, one person working, making more than enough money to support the whole family. Sort of an older school mindset, harkens back to, you know, the, the, maybe apocryphal stories about the mid-20th century, but certainly an optimistic view nonetheless.

Yeah, it's notable NEET levels have been rising since 2021.

That's not employed in education or training, not employment.

So that trend started well before the current AI paradigm. Yeah, but wouldn't be surprised if it continues, hopefully for the reasons that Bezos is talking about. I think there's gonna be a lot of focus on his summer plans, John. Yeah, I think there's gonna be more scrutiny than ever. You know, this is a guy that can hardly jet around the Mediterranean on a big boat without being photographed every single day. It's true. And this was prior to these back-to-back, you know, multibillion-dollar rounds. And so going into the summer, I think people are going to be paying extra attention, right? Is he doing back-to-back days at the beach club or is he keeping it to a 4-hour window on Saturday, locking back in by Saturday night, getting a good night's sleep, spending his Sunday, you know, preparing for the, for the coming week, right?

He does have a co-CEO. Maybe they've split it up. Most people think co-CEO is like they're both involved in every decision. Maybe it's more of like a he takes the winter and fall quarters and the other co-CEO takes the summer and spring? Like, if you divide it up that way, no one's ever tried that with a co-CEO. Just like, you run the company for 6 months, then I run the company for 6 months, and you're completely off on vacation, maybe?

Yeah, yeah, the alternative is, you know, Bezos would say like, Friday from 7 PM until 2 AM on Saturday, I'm not gonna be super available, super online. Uh, and then also from 2 AM till around 11 AM, uh, I'll be fully, fully offline. Okay. Um, then there's about a 1-hour window on Saturday where you can take like a lunch break or whatever you need to do, but then from 1 PM until 2 AM Sunday, I'm gonna be offline again.

Uh, but back, you know, maybe if you're co-CEOs, you should just do, I, I'm CEO of TVPN on the odd hours and you're on the even hours. Yeah. And so if there's a business decision—

Is that the secret to nonstop productivity?

Interlace them. Yeah. So, oh, it's 3:00 AM, call Jordi. Oh, it's 4:00 AM, call John. Co-CEO is working out. Bezos is co-CEO of the new company along with Vik Bajaj, who, a Google veteran and former healthcare founder and investor. They started Prometheus in late 2024. So people have sort of like pegged this as like a— Overnight success. Late in the cycle. It's, oh, it's 2026 and they're jumping in. They've been working on this for 2 full years. And obviously, Jeff Bezos has an incredible amount of AI and robotics experience, both in the recommender systems in Amazon and Kiva Robotics, which is probably like the— you could make the argument that it's like the Instagram of the robotics boom. It's like one of the most successful acquisitions. We're talking to the founder of a robotics company at 11:45 today. So his co-CEO is an adjunct professor at Stanford School of Medicine who previously co-founded Google's life sciences division. No big deal. That's a huge accomplishment. And the company's goal is to build AI systems that can assist throughout the entire engineering process. That ranges from designing products and predicting their performance to manufacturing them. We're trying to approach that problem end to end. They also held talks. This was the big one. They held talks to raise a $100 billion fund. Which is a lot of money. We'll talk about SpaceX in a bit, but SpaceX is looking for $70 billion, and it's a huge business with so much attention. This is an entirely new thing, and they wanna put $100 billion to work, but of course, Bezos has plenty of money to backstop it, anchor it, do whatever he needs, and he has massive connections. So the Wall Street Journal reported this, and they said the goal was with that fund to acquire manufacturing businesses and deploy deploy AI across their operations. And remember we did that whole list of like, what could they actually buy for $100 billion? What would be a good buy? And there were some very interesting, like you could buy like Goodyear, for example, and work on tire manufacturing. And I think I was wrong in all my predictions, but it was an interesting exercise to just take a little tour of the American manufacturing industry.

But my read on it is some combination of like a foundation model that allows you to generate full product, full products basically, like product specs, designs, et cetera, to high fidelity. And then something like a SendCutSend style infrastructure where you can just make that product on the fly and get it delivered. And that is the moment we're in right now where on your computer you can make at least an image of anything you want in the entire world., but then there hasn't been that much of a speed up on the other side, right? And so closing the loop there is gonna be very, very cool.

Wow. This company Prometheus is bigger than AIG, the insurance company, bigger than Chipotle, bigger than CBRE, the real estate company.

Are you on companiesmarketcap.com, John?

I'm not, I'm on ChatGPT. Bigger than Baidu? Wow. That is crazy.

John, do you think that they're gonna be punished for their hubris? Hubris and chained for eternity with a—

The nominal determinism of Prometheus is sort of crazy, right?

Yeah. But chained up for eternity, eagle comes, eats his liver every day.

Isn't Prometheus like the third time an AI company has done this? Because there was the Prometheus data center at Meta and then there was a company called Icarus too.

Icarus.

Yeah, I got in trouble for that one because I was talking to a New Yorker reporter about that and sort of overstepped a little bit because I was like, they have a great team. I'm rooting for them. But yeah, so I think we need another company be named Hercules, right?

Because Hercules saves Prometheus. He frees him from the chains. Okay. But then, uh, Hercules— company called Hercules.

That's a good company name. I like that one.

Yeah, that's right. I don't know.

Um, but, uh, yeah, so if raised, it would be the latest in a series of investors buying established companies and using AI to improve productivity and margins. We've seen a couple of these funds, uh, roll out the AI roll-ups, AI modernization, private equity firms. Obviously private equity Microsoft themselves are doing this, but no one's come in with $100 billion. You're in a completely different tier from all the other companies that are doing this. We've had many of them on the show. AI's advances have paved the way for, quote, a multitude of golden ages, not just one, happening simultaneously, Bezos said. He said this is the best time to start a company.

I love the optimism. That's crazy. Strategy, MicroStrategy, and Prometheus are roughly the same, valued the same, as well as CoinShares.

MicroStrategy is in sort of a rough spot, right?

Isn't Bitcoin down? No, it's a good— I mean, it's an interesting bet, right? What do you want to buy? What do you want to own at $41 billion? Yeah. MicroStrategy is at $42 technically.

Yeah. You could buy Reddit. That's $33 billion. Anyway, the next story we got to talk about is gamers are rising up. Gamers are rising up. You heard that the government was trying to put gyms in airports. Well, the gamers had something to say. They're pushing back. Airport lounges are catering to gamers. The airport lounges where video games are more important than drinks. Travelers can kill time, play everything from Mario Kart to Minecraft before their flight. We got to do this. Get a little rust going next time we're flying. Let's do it. The first thing David Lee—

Have airports— Before we dive into this. Has TSA gotten better, more consistent? I have a hot take on this.

I think TSA might be goated.

I think they're pretty dialed.

I think it's amazing. I think it's good.

Out of the last maybe 30 flights, I have not waited more than 30 minutes in a security line. And that makes— commercial is obviously still— Torture.

Torture. It's sort of a— it's sort of a—

a Promethean punishment. Humiliation ritual. But that being said, being able to plan on a 30-minute security line makes commercial travel pretty— Yeah.

I mean, it was the butt of every comedian's joke because it's so universal. Everyone's been through TSA. It's very slow. Take your shoes off. There's a million little places for comedy gold that's highly relatable.

You don't have to take your shoes off. Yeah, you don't have to even take your laptop out.

No, it's, it's getting back. And there's like, I mean, I guess it's getting more surveillance-y because they take a photo of you like 17 times on the way through.

You can refuse that though. You can.

Yeah, you can just say no.

Nick does that. You can just say no, but then don't.

Yeah, you can just say no.

They don't. They used to about, I would, I will say about 2 years ago, they would tell you more frequently like it's optional. Yeah, it's optional. Okay. Now they're just like photo now. Okay. And you can read the text and it says like you can opt out.

Okay. Can you opt out of the whole thing? Is it all just like a high-agency test? You can just be like, actually, I just want to go around. I don't even want to go through the metal detector. And they're like, yeah, right this way, sir. No one's ever thought to break out of the Matrix, but you reached the final level. You reached based reality.

Good job. Continue. Yeah. But truly being able to plan on it's going to take me like 30 minutes to get through. You can like comfortably show up 45 minutes before your flight. Yeah. Reliable. Make it every time. Every time. Every time. Which means that you're pushing it even more to the limit. Oftentimes, I will be getting through security, and Jon will text and be like, just got my car. I'll be there. I'll be parking in 45 minutes. And I'm like, I thought I was cutting it a little bit close. But mandate of heaven, you always make it.

I always make it. And if you're not missing flights at least like 10% of the time, you're actually setting yourself— you're not cutting it close enough. No. You have to be taking some level of risk or else you're just wasting time.

No. Yes. No, because oftentimes—

Hey, you went to an investor and you were like, your goal is to never invest in a company that goes down. Oh, you're a VC, but if you get a single zero in your portfolio, you can never invest ever again. Would you do that? No way. No way. No way. I'm not engaging. No way. So you're going to have some zeros in your portfolio. You're going to miss some flights. But on average, your DPI is going to be quite good, which is expected.

Yeah, yeah. Comparing that— This is where the argument goes.

I can just have as much fun at the airport. So I don't mind showing up 5 hours early.

No, no. But OK, if you miss your flight, and then the next flight is like 3 hours. Yeah, yeah. —then like, it doesn't, it doesn't math out because like, if you just get your flight 10 minutes or like a little bit earlier, then it'll like match. Like, I think it makes sense.

So basically, like, let's say I fly once a week and I save an hour with my strategy. That's 50 hours a year and I only miss one flight. You're not saving an hour. I saved 3 hours.

And you're not, you're not like saving the time. I am. I am. You were probably at your— just like having another coffee at the hotel. And/or like doing something, having a conversation that you could just have.

No, I was literally, before I go to the airport, I am literally locked in at the clock factory, okay? I'm locked in at the clock factory. I'm being the most productive I've ever been. Making clocks. I'm making clocks, clock by clock. And you think that that time's useless.

Well, okay, now you can be gaming at the airport.

You should be getting there way earlier. I should get there, I should get, if I want to prestige on COD, You should get there maybe like 72 hours early.

Yeah, you're just like, yeah, I got to get up at 2:00 AM this morning. Like I got like a, I got a super early flight. They're like, oh, what time is your flight? Like 11:00 AM. 11:00 AM. Like, yeah.

Oh, is this a coincidence GTA 6 dropped and you're going to the airport 10 hours early? Well, now you can because airport lounges are starting to cater to gamers. Gamers are the most important constituency. We're going to have a gamer president one day. So the first thing that David Lee and Cassie Yang noticed when they popped into the new lounge in Terminal 1 at Minneapolis-St. Paul International Airport on Friday wasn't a robot bartender, which they have apparently. I believe Ben will be traveling through Minneapolis-St. Paul. We should give him a little gaming challenge, see if he can get to prestige on COD before he hits his flight. It was something only parents in this digital device era could appreciate, unfettered access to their phones. Their kids, 8 and 12, were busy playing Mario Kart and Crash Bandicoot at the lounge's video game stations to commandeer their phones for some pre-flight screen time. The Portal Lounge is, by design, not your grandfather's airport lounge. Yes, there is a small buffet and bartenders, robot and real. That pour that lounge libation staple, an Old Fashioned. But the focus is on entertainment. They're gaming. It's the latest concept from a husband and wife team on a mission to amp up airport downtime. Emma and Jordan Walbridge, with backgrounds in hotels and medical device sales respectively, opened their first airport video game lounge at Dallas-Fort Worth International Airport in 2018, called Gameway. There are now 11 of them at airports across the country and 2 more due to open this year. Portal in Minneapolis is the latest iteration of a concept with a few more bells and whistles. They zeroed in on video games after regularly visiting lounges on business trips and finding them too staid. I always felt like I was in a library, Emma says, with some guests slushing others and throwing— shushing others and throwing sides.

Picture makes me feel like I am a television screen.

What do you mean? Oh, because they're looking at you like you're— With the control. You're the game. Yeah, interesting. She's played Mario Kart for years. Jordi has played Halo, which is getting a refresh soon with Campaign Evolved.

I think this is gonna be more successful than like the whole airport gyms movement. I think so. Like, every once in a while somebody will post online, "Why don't airports have gyms?" And there's been a few moments where it would have been nice to have a gym. But imagine a world where, you know, even just 5% of your flight had just gotten off of an hour on the treadmill. Yeah. And was getting on pretty sweaty.

Pretty brutal. Pretty heavy. Some heavy breathing going on.

I don't know. Golf simulators in airports. That could be pretty neat. Done that before. Yeah.

You've actually done that. You've actually played simulated golf. Interesting. Racer Martin. What a crazy name for a 19-year-old chef from Arizona. R-A-Y-C-E-R. That's a cool name. Racer. He plays Apex Legends at the Gameway Lounge at Chicago's Midway Airport. I plunked down at the gaming station and embarrassed myself playing Mario Kart over a lunch of Diet Coke and Ruffles cheddar and sour cream chips. The patient Gameway manager who used to work at GameStop assured me that he— that it just takes practice and encouraged me to play at home. Another employee did her best to introduce me to Minecraft even though I gave up after 5 minutes. This Wall Street Journal reporter is clearly learning the ropes of these games. Video gaming is the main attraction at the much bigger Portal Lounge in Minneapolis which opened over Memorial Day weekend in the D Concourse. But it's just one attraction. The food is more lounge-like with build-your-own bowls. That's a lot of bowls. Buffalo cauliflower salad fixes a couple dessert. Adding to the futuristic feel, a robot bartender takes the sec— the center stage pouring drinks. It's from the same company as the makers of the Bionic Bar on some Royal Caribbean ships. At almost 4,000 square feet, Portal isn't on par with, say, an American Express Centurion Lounge. It's more like one of those smaller escape lounges. RC Cola is on tap. There's music, outlets galore, and plenty of seating. But no dedicated bathroom. The gaming lounge is in the back and bright gameway.

Anyway, you can go check it out. Seems— If we let John keep going here, he would just talk about airport gaming lounges.

No, no, I would talk about Railway because Railway is the all-in-one intelligent cloud provider. Use your favorite agent to deploy web app servers and databases and more while Railway automatically takes care of scaling.

I would like a Railway lounge in every airport.

You joke, but a Vibe coding lounge is coming. Like these gaming PCs, they can do local inference. You can fire up your agents. You can be doing all sorts of stuff. I like it. I'm surprised. Yeah, I wonder what the cybersecurity stack is like there. It's got to be sort of tricky to have someone give them free access to a powerful PC. They're going to install some weird stuff.

Unfortunately, the lovable airport lounge is coming. It's like, have 5 minutes, build an app.

Build a website.

Build the next hit app. Yes. Yes.

Uh, well, I think the next hit app might be Halo: Campaign Evolved. Let's watch this video from Blue Thunder. What do you think? Are you a Halo guy? Do you ever play Halo?

Not really. Never played Halo. Never captured my heart.

But I never got into the whole hardcore multiplayer scenario. But it does look good, doesn't it? Look, uh, like fully refreshed. Looks very modern, AAA. Uh, Blue Thunder says this is—

yeah, what if they made a—

not the best look FPS games he's ever seen.

What if they made an airport-sized lounge? Sort of like a facility just for 1v1ing on Rust. That'd be great. You could potentially get, you know, 100,000 people all in the same place. LAN party matched up. Yeah. This could have been, this could have been like a one-shot output of an LLM. Yeah.

When I saw this, I wasn't sure if it was AI video or vibe-coded version of Halo. Did you see that thing where someone was trying to build GTA 6, like vibe-code GTA 6? You saw this? It's such a funny demo. I mean, you don't want to hate on it because like it is like the tools are very powerful and, and you do imagine that over time you'll be able to. But the first effort video was very funny because—

Well, we got it right. Halo. We got an ex-Halo 3 pro in the chat, Alex. He says will be a conservative pickup of nostalgic non-PvP gamers. Don't expect it to do numbers.

Interesting. X Halo 3 Pro. What was that like? Interesting. Yeah. Day 1 of building GTA 6 still feels fake. Typing that out loud. Upgraded to Cloud Pro Max 20x just for this. Spent a couple hours getting the whole project structured and pushed to the repo sandbox. It's up and running. I mean, this is actually where you'd start. Like, it's not that unreasonable, but it just like, it's funny because it doesn't look anything like any GTA game. Wait, what? You haven't seen this? You haven't seen this? So Matt Schumer, former guest of the show and creator of the viral essay, Something Big Is Coming, where he predicted— Something Big Is Happening, where he predicted that AI would cause 10% unemployment by April and compared it, likened it to COVID. Wait, that was his prediction? Yeah, he said that it was going to be like a COVID-like takeoff in the economy. He didn't map the economic numbers directly, but we did. We were like, okay, like what was COVID actually like in 2020? And we went from February, which is when he wrote his essay, February 2026, saying AI is real, it's gonna change things, it's big, which a lot of people have been saying for a long time. And I was like, I agree, but I don't think in April, May, like the NBA is gonna be shutting down because of AI, or I don't think that the unemployment rate is gonna spike to 10% overnight. Like there might be little pockets of layoffs related to AI and there might be some—

If anybody wanted to make AI less popular, shutting down the NBA Finals because of how good the models have gotten.

Just all the NBA players are like, we got to invest. We got to— we don't have to—

not even invest. They got to invest. They got to invest their time.

Oh, they got to vibe code. Yeah, maybe they don't want to get left behind. No, I don't think they'll be left behind. But Matt Schumer said someone should set up a community-funded Fable run with a prompt like /loop until you've created a GTA 6 caliber open world game with a quality and scope surpassing what is shown in the initial trailers. And he says, happy to write the prompt if someone organizes this, which some people were making fun of. But it started. Zhiwen is doing this. He's 25. He likes Claude and Codex, tech, AI startups. I'm rooting for this. I think this is going to be very educational for Zhiwen, although, yes, this video is very funny because it doesn't, it doesn't look like GTA yet. But, you know, if you can get to anything that's at all playable or interesting, even in a month, even in a year, take as long as you want if you can do it. If you can get, you know, a million people to pay $10 for this and it only cost you $5 million in inference and you did it as a solo dev, like, that's a $5 million profit. Like, that's a great outcome. Like, I love indie devs. I love that world and I like that these tools are opening up more opportunity for people people to go mess around and build cool things. It's got to be really entertaining to just think about building, uh, like a vibe-coded game. Some of the games look really good already. Uh, of course, like, even if Strauss is gonna be pretty worried— no, um, did you, did you hear him talking? The Prime Agent had a good reaction video to David Sener talking to Strauss Zelnick, uh, and he's like, finally, like a normal AI CEO, because he, he basically says, like, we have Mobile games, we have a mobile games division. We have— we make billions of dollars from it. We make a lot of money from mobile games. And there are direct clones of our games that launch every day. Like, the software has been good enough that you can clone our games. They launch every day. The business is much more complicated than that. The IP matters. The distribution matters. The functionality of the game matters. There's so much more to it than just a carbon copy. So he says, yes, you could just go vibe code GTA 6. People are still going to play our game because of all the other strengths of the Take-Two business. And I agree. I thought it was a very good take. Anyway, let me tell you about the New York Stock Exchange. Want to change the world? Raise capital at the New York Stock Exchange. Just do it. Some big IPOs happening this week, next week. It is IPO season, but it's also trucking season. The 4-year US trucking slump is officially over, says the Wall Street Journal. It's over at last. Truckers who held on through almost 4 years of slumping freight rates finally have something to celebrate. It feels like a breath of air for an industry that maybe felt like they were running out of air, said Webb Estes, President and chief operating officer of Estes Express Lines, a Richmond, Virginia-based carrier with revenue of about $6 billion. Freight rates are finally climbing. Good news for them. Trucking executives are calling to an end to one of the longest freight downturns in carriers' memory. They say rates have risen to more sustainable levels after a period of low earnings coupled with Trump administration crackdowns on immigrant drivers pushed many carriers out of the market. Logistic Managers Index, a monthly survey of supply chain managers, showed transportation prices increased in May at the fastest rate for any metric in the report's 10-year history. So if you scroll down, yeah, that chart is crazy. So good news for the truckers. Of course, this is going to push up, you know, prices and inflation potentially. But it is good to have a healthy trucking industry, a healthy dry van spot rate, which I'm sure you've been following intently. Dry van spot rates for the week of June. Were up 5,200 basis points. Let's go. 52% year over year, excluding fuel surcharges, according to a report. So the fuel prices get passed through, but the economy is doing well. And so prices for dry van spot rates are increasing. As he said, is a stronger freight market over the last few months. The nationwide company is increasing its fleet of more than 10,500 trucks. And expanding its pool of about 11,000 drivers.

What is driving this?

This has been a supply-driven freight recovery as opposed to a demand-driven freight recovery. Turnaround comes after 4 years in which carriers have been squeezed by a combination of too many trucks on the road and not enough loads. Drivers rushed into the industry during the pandemic, so there was a boom in online ordering, shipping, all the different things happened during COVID Consumer demand drove freight rates to record highs. Rates plummeted in 2022, forcing hundreds of thousands of smaller carriers out of business. Many truckers held on even as costs for drivers, equipment, and insurance rose. The exodus accelerated over the past year, and trucking specialists say the industry has finally found an equilibrium where the supply of trucks is low enough to lift rates, which will, of course, bring more trucks into the market and Rates will adjust, of course, but higher fuel prices due to the war have pushed up trucking companies' expenses. But operators are largely passing along those costs to their customers. So anyway, let me tell you about Shopify. Shopify is the commerce platform that lets you grow with your business and lets you sell in seconds online, in-store, on mobile, on social, on marketplaces, and now with agents.

We have some breaking news. What's breaking news? Ty Morse says on Saturday we finished construction of set number 2. Set number 2. So Ty had built a set. Yes. For his interview with Elon Musk. Yes. It sounds like they built it in a place that they weren't— they were only able to occupy for like a week.

7 days. So they shifted. 1-hour notice. And that was on June 6th. That's how this is gripping. This was a trending story on X. He says, for the past 3 months, I've been asking myself, the simple question, what would Elon do? Elon would build the rocket and put it on the launch stand before he got approval. He would ask, what's the biggest bottleneck to achieving my mission? And then go about systematically undoing the bottlenecks until he and his team had made the impossible happen. A maniacal sense of urgency is our operating principle. And so he has built yet another set that I think can last a full month. And hopefully he lands the interview. He's incredibly pumped up for this. And the meta story of him building this set has done better than most podcasts. There's more views on—

I want to read, John, I want to read the summary of the book about the story of Tim Worst getting Elon on his podcast.

Like 7,500 people, 2 million people watch the video about the set. And so you do have this dynamic where the meta story is more interesting or it like, it's by default more interesting because the interview hasn't actually happened. But there are plenty of Elon interviews at this point. Like he does a lot of interviews that don't go super viral. Like he did this interview with Jamie Dimon and I think that's fully recorded. I think you can just search it and watch it. But I don't know that that's something that like everyone tech saw and everyone in tech talked about. But this is a story people are talking about, the Meta story. And people love to see bold moves. People love to see building and all these crazy things and risk-taking and all of that is on display. And so Ty Morse is getting a lot of attention.

So good luck to him. Godspeed, Ty. It is a beautiful set. It is. It is.

Texas is America's new center of gravity, says The Economist, Exxon's reincorporation. They were in Texas, they left Texas, they went back to Texas. There's one more feather in the state's cowboy hat, says The Economist. Just when I thought I got him to— just— this is a hilarious, hilarious song lyric. Just when I thought I got him to fall in love with Tennessee, I should have known better than to take him back to Abilene. Abilene, Texas. They were singing about data centers in Ella Langley's "Choosin' Texas," a song that has ranked near the top of America's music charts throughout 2026. Some of those listens may be coming from policymakers elsewhere in the country. Much like Miss Langley, they are choo— they are losing to Texas. On May 27th, ExxonMobil shareholders approved a plan to cut the oil giant's ties with New Jersey and reincorporate in Texas, where it has long had its headquarters. It's not alone. According to CBRE, a property firm, at least 140— 184 companies shifted their headquarters to Austin, Dallas, or Houston between 2020 and 2025. Among them Tesla, a car company, and Caterpillar, a maker of construction equipment. Texas is steadily establishing itself as America's— America Inc.'s new center of gravity. No state receives more business investment or is adding more people to its population. From 2020 to 2025, it created roughly a fifth of all net new jobs in the country. Wow. In early— in the early 2020s, Texas was luring remote workers fleeing high taxes, exorbitant house prices, and bad policies in America's coastal metropolises while benefiting from the Biden administration's subsidies for green energy and chipmaking facilities. Now, the state's dominance in energy has made it a major beneficiary of the data center boom. Meanwhile, its technology and finance ecosystems have been deepening. This summer, it will ring its first standalone burst, this Texas Stock Exchange, joining outposts of the New York Stock Exchange and NASDAQ already operating in the state. Donald Trump has called the New York Stock Exchange's new branch an unbelievably bad thing for his hometown of New York, even if his social media venture was the first business to list on it. So that's a very funny thing, but he was obviously giving that in the context of like New York is defined by its exchanges, and in the context of just New York, the city, losing out to Texas, or New York State losing out to Texas in any way is bad for that state, I suppose. The state's appeal is to yuppies, to yuppies is growing with every stream of a country song. I've been listening to a lot of country songs, Joe Weisenthal. Plays in a country band, blues band. Uh, you know, who knows, maybe he'll move to Texas one day.

People ask, uh, dream guest. Yeah. And I tell them dream performance. Oh yeah, Joe Wiesenthal acoustic set. Unplugged. Unplugged. Tiny desk. The tiny, tiny, tiny desk. What if we set him up with a slat wall set?

We got to get Roger from Conde Nast. He can jam.

Got to get— but a little slat wall set.

Yeah, and you can just play an SM7B podcast microphone on the guitar. That might be the trick. Might be the trick. To understand its ascendancy, start in Houston, heart of the Texan energy industry. Its oil and gas barons have been raking in profits as a result of the Iran War, but over the past years, the state has also become a hub for green energy. This year, it's expected to build 2/5 of all utility-scale solar in America. A technology for which its wide open flatlands are ideal. Interesting. We got to talk about the Buzz Lightyear GT3 RS. We got to get to that before our first guest arrives. Let's scroll down and find the latest, the latest and greatest from Porsche and the Pixar Company. They're teaming up for the launch of a custom GT3 RS. Porsche CEO first, he said The company will not produce an electric 911, but they will produce the Buzz Lightyear GT3 RS. Can we play sound?

It also says Space Ranger, Lightyear. My vents are going to be purple, and Porsche even commissioned Goodyear to make the tires say Lightyear, just like they do in the movie Cars. I get more painted graphics here on my roof. My roof fins are going to be painted. My intakes are going to be painted. My entire wing is painted, and it even has a Space Ranger logo. Painted on the back. As soon as you walk into the car, I get a plaque that says "To Infinity and Beyond." I have Lightyear-branded floor mats. I get carbon bucket seats, but this time the carbon is actually purple. I get a Space Ranger logo on my headrest, another one here on my roof, suede on my steering wheel, more on my dash, a 1 of 1 plaque, and even my key has the Space Ranger logo. There's also a Woody version that's based on the 911 Carrera T. They made the paint look like actual denim.

I get— and Woody kind of got Bullseye. Buzz gets it, gets an RS, and Woody gets a T.

Yeah, these gorgeous denim and leather badge embossed on the headrest. I get a 1 of 1 plaque, wood on my dash, a leather steering wheel, and they should have made it the ST.

Missed opportunity.

This car painted in Jesse White Metallic. I get painted, and instead of saying Targa, it says Jesse In Jesse's car, you get a plaque that says, "Yeehaw!" I get cow print floor mats. Wow. These denim and red leather seats with Jesse's badge pocket.

Do you do one for the T-Rex? T-Rex, the T-Rex? You gotta do the off-road Cayenne Turbo. So which one would you take?

This guy's video, like his format, really, really exposes my ADHD. Oh really? Because of the editing or? Yeah, I'm like, wow, this video, they really, this video could be like a 10-minute video.

Yeah, he's condensed it down. It's crazy how much he's condensed it. He goes through everything, and I think every line is like written, and then, and then he says it multiple times.

So SEO optimized. Search any car, like, this is the most efficient way to figure out, like, is this interesting?

No, Forrest Auto Reviews is really cool. I'd love to have him on the show. Uh, he also has like incredible access to Chinese vehicles. I'm not exactly sure how that happens. I think he just has good relationships with the Chinese automakers. So if you want to learn about the new Zeekr and stuff, like he'd get— yeah, he's the one, he's how I found out about that crazy Chinese SUV that has a gas motor range extender for an otherwise electric car. So the range is like 900 miles. And I think you can actually get over there and tour the cars. He does fantastic Fantastic work. Would you pick up a regular GT3 RS or the Toy Story GT3 RS if they were the same price? Which one are you going with?

I tried to watch Toy Story with my little guy. Tried? No, no. I was watching it with my little guy, and it was too scary. Oh, OK, OK. We had to turn it off. So right now—

There are some scary parts that child head with the pins in it.

Yeah, with the little rough, little rough. So parental guidance, this is not the right car for my household yet. Too scary. But I could see it in the future.

The GT3 RS, too scary. Yeah, um, wild. Uh, what, what is that? We got to go back to the Ferrari Luce.

Luce. No, the, the fact that images too has just fully democratized like, uh, car designs is pretty incredible.

So, so what, what are we looking at here with at this post. They had just one job. This is someone that took— OK, the basic design.

Basically, they're like, ignore the fact that the Luce was meant to be like a daily with more seats. Yep. And just said, what would an electric Ferrari look like if you—

This is not an electric Ferrari. It has 4 tailpipes. This is a naturally aspirated, rear mid-engine V8.

Oh, maybe that's the point of the post. They said they just had one job to do. But it's like, just make another.

Okay, yeah, I get it. You like, you like V8s, you like gas cars, you like things that are designed differently. Like, Ferrari will sell you one of those. They have them. Like, they have the 296, they have the Testarossa, they have the SF90, they have the California, and there's a new one. Like, they have plenty of options in that category. They don't have an EV. They wanted to do something different.

I don't know. This does This does look cool.

It looks cool, but it's like that also already exists. Like, that's not something that's necessarily like new. He says, Johnny, yikes. I don't know. I think it's— I mean, I agree. Like, this one looks aesthetic. It looks cool. But like, it doesn't check the box that they wanted to check, which is like a 5— their first 5-seater, first electric. Car first, something with like, you know, this battery setup and all this other stuff. Like they wanted something experimental. Again, like it does seem, it does seem risky and the response has been really negative. But like, clearly it wasn't like, oh, they were trying to go for this and they wound up in some crazy place. Oops, we put an electric drivetrain in it.

Yeah, we'll see if this guy's still posting that after driving through Cupertino a year from now. Just wall-to-wall traffic. Just sitting, sitting in a field of luches. Yeah, just moving one inch every minute.

Well, they did the— they did the Buzz Lightyear GT3 RS. Should they do the Shrek luce for— get some DreamWorks collabs going? Pull it up in the Shrek green luce images too. Get it working right now, Jordi. The, the, the Shrek, the Shrek luce. Let's We'd go on reasonably hard. That might, that might swing you. Shrek's not too scary. You could do Shrek. That would be acceptable in your household, I think.

What about, uh, what about a, uh, while that's pulling up, yeah, make a donkey.

Look at this 944, the legendary 944 from Porsche made this grand return in, uh, 2026. This is a cool design. I love the 944. Looks great. Yeah. That would be interesting, especially given that the CEO of Porsche has said that they will not produce an electric 911. They got to do other things. Maybe the 944 is the trick. Anyway, let me tell you about Cisco. Critical infrastructure for the AI era unlocks seamless real-time experiences, a new value with Cisco. And we have our next guest in the waiting room, David from Neuro Robotics, here with a a fantastic fundraise. Welcome to the show, David. How are you doing?

What's going on? Hi, hi.

I'm doing great. I made it to your show. So awesome.

Yes.

Incredible. With this fundraise. Incredible. We're honored to have you. It's crazy we missed you earlier, but it is your first time on the show. So please introduce yourself. Tell us a little bit about the company. We'll go through the fundraise. And I have a bunch of questions about robotics because it's always a fascinating topic, but I'll let you kick it off.

What a day to be announcing a mass massive round in physical AI. Oh, yeah. Because there's only one person in the world that could raise probably a bigger— Yeah. AI meets physical world round, and that's Jeff Bezos. But we're happy to have you here today and excited to hear what you're working on.

Great honor. Yeah, so a little bit about me. I'm David, founder and CEO of Neuro Robotics, a company out of Germany, so in the southern part. Yeah, my goal is actually making robots cognitive and bring them them actually in all areas of our lives. And I do think that robotics will actually take over any physical task in the future. And I think that's also what we're seeing today, like big progress having, let's say, physical AI combined basically with, with a robot body and doing great things, mainly focused right now industrial, but looking also into all other sectors.

There's this big, there's this big trend right now in just the LLM, the frontier AI market where companies will, you know, use the most versatile tool for the job, the frontier model, they'll token max, it'll be really expensive. And then pretty quickly they'll figure out, okay, for this workflow we can actually use an open source model on commodity hardware. We can use a legacy model. We can cut the cost by 85% pretty quickly. And I have this suspicion that that will be the case for humanoid robots. You'll have the humanoid go and, you know, put the windshield on the car and then pretty quickly realize like, wait, for 1/10 the cost, we could have a KUKA robot arm do that. And there will be this gradation of what is economically productive for each robot and you will only use the most expensive humanoid for the really, really unwieldy task and we won't see a world where everything will be humanoids. But what's your thesis long-term?

So first of all, like I think there's also one of the biggest differentiator from us to others is that we do believe in different embodiments. So, and what we have today is like what we are using mainly today is like mobile manipulators, but always combined with a physical AI brain. And I think this is also going— but going forward, we do, we will see more humanoids, and I think we will see them in all, all areas of our lives. And the reason for that is, is more like that, that they're having the form factor of a human. And I think the whole body, like, the whole world is basically created around the human and ergonomically, let's say, made for humans. And this basically also the main reasoning. So where I think the biggest breakthrough we see right now is actually connecting to a brain and with more sensing capabilities. So until now we are seeing mainly like all these companies coming up with the vision language action models. But here I need to say that that is not enough to bring robots into our fields. It's a little bit like learning how to swim. You can't learn swimming by just watching a video. It's about a nervous system.

You gotta go to r/swim.

Yeah, yeah, you gotta learn it from the Reddit data. There's plenty of data on Reddit. You can learn how to swim. Yeah, it is interesting. There's something to like visual— learning to do physical things by just watching videos does work to some degree. It's not full, it's not enough. But I remember growing up surfing, snowboarding, skating, all these things.

I learned to tie a tie from a video, but I was practicing I was practicing the whole time while I was watching the video. So, I was getting the input and the reaction. You were combining that.

Yeah. What I mean actually by that is you can't just learn things just with your brain in this dynamic world. So, it means actually there is more than brain. It's a nervous system and the reflexes. And this is especially needed in the— like if you're looking into all kinds of tasks humans do today, it's mainly like the biggest advantage we have is the flexibility of adjusting things like on the environments, on different areas. And this is also what we are doing. This is one, the video, awesome. It's actually the Neura Gym. It's a place we're building up right now all over the planet. And in every major city, you will find them soon. So it's actually a place where you can train robots doing all kinds of tasks in the physical world. And then more than just the brain, basically also the nervous system. Interesting.

What other sources of data— Why, wait, why all over the world? Why can't you centralize something like that?

Yeah, so because we want to give like everyone access to actually not just buy a robot to have one and then figure out you need to train it. So you have a central place and you can actually build first trust and the second actually you can train them already and use the data with one click. You can transfer it to any other robots in your factory or wherever you need them. So for us it's right now important because the gap we're having right now and the difference actually to large language models, as we know, like ChatGPT or others, is that this data is not existing. So we do need to build first, let's say, the infrastructure for it so that we can actually create the right data to train actually the physical AI brain behind that. And the most easiest way is letting everyone or enabling others to train the robot or to fine-tune the foundation model we built.

What are the— biggest opportunities that you guys are working on in Germany? I imagine hopefully helping make some of my favorite cars, but what do you have going on?

Yeah, that's a good one. I'm also a car lover, so I could have also had a lot of input on your talk before. So it's like— so yeah. Would you go—

would you go— would you go 3RS Buzz Lightyear Edition or Ferrari Luce Shrek edition. That's a good one. Are you Shrek guy? I think it's too easy. Yeah, for sure not, because you might like Shrek and you might like a GT3 RS, you might like, uh, Buzz Lightyear and a Luce, but that's not the option. You got to pick one or the other.

No, I'm still the, the engine guy.

And yeah, okay, there you go. Yeah, purist.

So yeah, this So yeah, but yes, there is a lot of projects, especially in the automotive industries and also especially on the supplier side where they have to reduce costs. They have to find, let's say, ways to be efficient as the others are in Asia or somewhere else on the planet. And the only way for that is finding, let's say, some more autonomy or like, or bring more autonomy in their fields and Especially because like here in Europe, if you're looking, let's say, into 2030, you will see that we have like about 7 million less human labor. So we do need to have like, we need a solution for that. The only way is embodied AI, something which can do things fully autonomously and as capable as human can do.

How are you thinking about your supply chain? A lot of robotics companies in America are saying we want an American supply chain. Other companies are saying, hey, we got to move as fast as possible. Realistically, is there going to be spyware in a Bolt that we buy from China? Like, let's not worry about that. It is maybe a supply chain risk if they cut you off, if you're thinking about supply chain resiliency, but what's your approach been? Do you want to source entirely from German, German supply chains or European supply chains, Western supply chains or global supply chains?

So at this time, we are focusing right now on the, like, let's say, European supply chain. So I still believe in a global world, but I do see right now the biggest advantage of converting them still because most of the components are there, but they need to be still, let's say, modified to use them for robots. Here I'm simply talking about even lifetime, or I mean, through the regulations, the lifetime of the sensors for cars are regulated basically. So they they break. So we have like a renewing process going on. So, and for robots, like, they're made for 6,000 hours and the robot takes sensors for 40,000 hours. So, and this is still a difference, a gap between, but, and that's also, I think for us, the most important topic, like working with those suppliers around us. We are based in Stuttgart where all the good stuff come from, like you, stuff you talked before about. And so there is the Bosch, the the, like, the Porsches, the Audis, the, like, the Mercedes, everything here around us. And we are using this supply chain, converting it right now into robotics. Very cool. But going further, I do think it will be— it will not matter so much. So it will come out again, like, wherever you produce. So you will have— we will have productions also in US and same time also probably in Asia.

Yeah.

Tell us about the round. Odd, before we get into the round, The silliest question you'll ever get. Can you ever imagine a robotic use case that would require an internal combustion engine? Let's say like a V6 humanoid.

It would be awesome, but not required yet. But I think the Boston Dynamics guys, they did an awesome thing. I don't know if you saw the quadruped they built some years ago, like in the beginning. It was actually with an internal combustion engine, as I saw. Gas powered? Like so. That's sick. But yes, no, at this time, fully electrified. But the cars I'm still loving, like with the V12 or V10, I think.

There you go. There you go. It's just— One of us. Tell us about the round.

Hit that gong, Jordi Hays. How much did you raise?

Yes. So we raised a $1.4 billion US dollar round.

Another one. Another one. Yeah! Well done. Jordi's red in the face. You guys are awesome. Jordi's red in the face. Thank you for taking the time to do this.

Great stuff. Are you setting up a— you have a US office already?

Or you're setting one up? Yeah. Yes, we have one and we are setting further. So like as you You might have seen also the investors on this round. There are a lot of American players and very important players, I would say. So, and I think it's also a sign towards where we go. Yes. Great stuff.

Thank you.

Yeah, because I mean, I think it's important, or like, I think something like that was never built. Like, after automotive, there was nothing really great coming out of Germany, and I think we have to change that also. And for that we need partners which are, let's say, knowing how to scale that, how to do it in this time we're living in. So, and here I do believe in our partnerships also out of US and Germany.

Very cool. Fantastic. Congratulations. Great to meet you. Thank you so much for coming on the show.

We'll talk to you soon.

We'll have you back on soon. Thanks so much. Have a good rest of your day.

It was awesome to be on the show.

Let me tell you about Codex. Codex is a powerful workspace for getting work done with AI agents, whether you're writing code, analyzing creating content, or automating business workflows. Codex helps you move projects forward from start to finish. Did you see at the NBA game last night a little, little person, somebody you might know, born in Tennessee, showing up on the camera as they pan around to the crowd finding famous people, finding iconic people, unmistakable people? We got a picture here of a celebrity at the game. You know who that is? Yep. Paul Tudor Jones, baby. Born in Tennessee. They didn't even have to give him a name tag because everyone knows Paul Tudor Jones. Of course, hedge fund manager, $8 billion net worth.

Let's zoom in a little bit more. I don't know what else is really going on in the photo, but let's get to the good.

I mean, everyone knows Tudor Investment Management. Tudor Investment Corporation. They know the investment, like the best episode, obviously. The Robin Hood Foundation, undoubtedly. He predicted Black Monday. I mean, yeah, he's a pretty big deal. Born in Tennessee. But there was someone who people weren't sure who it was. And so they had to give this woman a name tag. So if we scroll down, we'll actually be able to learn who this is. It's Taylor Swift. So you can take that name, type it in. Type it into anything, look her up and learn all about her, uh, and maybe one day you'll know as much about her.

Maybe, maybe Patrick was watching last night and that's like the best crossover episode.

Could be.

What if, what if they did a, a joint interview? That'd be great. PGT. Yeah, PTJ. PTJ. Yeah, meets Paul Tudor Jones. Taylor Swift.

Yes, I like it. Be good. Anyway, our next guest is in the waiting room. Let's bring in Mike from Allen Control Systems. How you doing, Mike? I'm doing all right, guys.

How you doing?

Seems like you're doing more than all right. Got some big— more than all right.

We having a fantastic, fantastic month around here.

Amazing. Good, good time to be in the gun on a truck business. Is, is that the accurate description? Is this all for scaling up both Take us through the product portfolio, reintroduce the company, reintroduce yourself, and then tell us where you're going.

Yeah, sure. Allen Control Systems, we are a company that specializes in the application of robotics for defense. Think of the Boston Dynamics of defense, essentially. We are building an extreme competency in robotics.

But you guys ship.

Oh yeah. But you guys ship.

You guys ship. Yeah, Boston Dynamics, they shipped, but it took a long time.

Not as much.

Took like 15 years.

Not as much. Yeah, yeah. So we are— we have our flagship product is Bullfrog. Bullfrog is essentially a robotic gimbal driven by computer vision and machine learning that solves kind of an age-old mechanical problem, which is that you can usually only have a system that is very fast but relatively inaccurate or very accurate but relatively slow in order to shoot down fast maneuvering drones. Especially tiny ones like Group 1, Group 2s, you need to be both fast and accurate. And so by using and applying AI to our control system, we're able to create a robot that can be both fast and accurate. And that's what we do that nobody else in the world has been able to achieve so far.

What's Jordi Hays? Yeah, talk about the progress over the last year. What, what kind of led up to this round and then where you're going from here?

Yeah, I mean, you know, we have been obviously kind of working on, on this robot for about 18 months now since we kind of first brought it out to the field, actually a little closer to 2 years. Bullfrog has been kind of slowly achieving more and more, getting closer and closer to battlefield speeds. In November of last year, we achieved kind of 100% kill rate against 13 Red Army flown drones. So drones that are flown by a team of the Army that are not telling us where they're coming from or what they're supposed to do or how many they have coming.

Interesting. Wow. Very cool. Was there already a program of record for this? Like, you hear a lot about how hard it is to get a new technology fielded because the government often knows that they want more planes or more weapons of a certain kind. And most companies that we talk to are usually saying, well, we fit neatly in this box. So it was easy for them to procure us. This feels both old and new. It's like innovative in all the correct ways and maybe leaning on the shoulders of like, what the Army can already feel like a truck and a gun.

Yeah, so we are not currently a program of record. You know, our technology kind of lays in a class that you usually call the ahead of need. Okay. So requirements that are being built right now, will start to address what this system is capable of doing now, but also what it's going to be capable of doing in the future, which is much, much, which is much, much more. Where we are very fortunate, you know, this company is really sitting at like just a perfect storm of tailwinds between the war and the size of the threat and the lack of other solutions that can solve it. You know, the president and the secretary of war created joint task force 401, which is specifically meant for selecting counter-UAS technologies for all services, both domestically and abroad. And so we were selected by them to get sent forward after, after the war started. That is basically a blanket approval for all services, including the Coast Guard and, and Secret Service, to buy our technology and put it to use.

Is Is this— I imagine the sweet spot is like the DJI-size quadrotor drone. We're not talking about shooting down something like a Shahed drone. Is that correct? Is that the focus long-term? Do you want to be counter-UAS in all classes over time, or do you just want to be the best at this small-scale counter-UAS? 'Cause they're all problems and maybe there's some other solution for taking out the enemy's version of like a Predator drone that might be a completely different job to be done, a completely different technology.

Yeah, I mean, that's a great question. We are currently very focused on being the best at shooting down Group 1, Group 2 drones, which are absolutely the hardest drones to engage right now. You know, lasers are pretty good at shooting down Group 3 drones, but you have to keep your laser on target for 5 to 15 seconds in order to do damage to, to the airframe. So we are, we are perfectly positioned for the Group 1, Group 2 threat. Now, that said, we can shoot down Group 1 drones very, very far away. And so if I can hit a 7-inch target on a Group 1 drone very, very far away, I can hit very vulnerable parts of a Shahed. At the same distances. So while we're not the ideal solution for engaging those, those larger targets, you know, you have Roadrunners and other things that can go out a little bit farther, you're going to want to hit them 20 to 15 kilometers away. But when they get closer and you need that last line of defense, we are absolutely a great solution for taking those out of the air.

Talk about scaling manufacturing. Like, it, it It doesn't look like you manufacture the truck. It doesn't look like you manufacture the actual gun, but there's a whole system that creates the Bullfrog product. What's actually going on that you can share in the manufacturing process? Is it milling different parts to fit things together? Just, you know, integration of different systems? How are you thinking about scaling manufacturing?

You know, when we built Bullfrog, you know, the hardware was not the exquisite part of the system. We built Bullfrog incredibly well. We built it so it will— could be easily maintained, and we built it so it'd be easily manufactured and scaled very quickly. And that's one of the other differentiators that we have against the other counter-UAS solutions that are out there, is we can get to thousands of units really, really fast. And the way we do that is 90% of Bullfrog is made from basically commercial off-the-shelf components. Sure. Standard things that you see in a robust supply chain with multiple vendors that can quickly get us to the amount of inventory that we need to start pumping out 1,000 of these every single month.

Talk a little bit about the compute. I mean, you're doing image recognition. It seems like there's a camera there. Is that all done on device? Is there a need for edge compute, a need for cloud computing? Because I imagine this is designed to operate in off-grid scenarios effectively, right?

Yeah, 100%. It has to be a completely self-contained compute system in order to do something autonomously. Think about an unmanned ground vehicle or an unmanned surface vessel. If you have to have an operator over a latent network manually controlling a gun to actually shoot a target in real life, you're never going to hit anything.

Anything. Yeah.

So having a fully autonomous system like Bullfrog out on those platforms that can be asynchronously tracking targets and just waiting for an operator to say engage that target is a huge advantage when you start talking about operationalizing unmanned payloads. Yeah. And that is another kind of huge niche that we have or that we're, that we're going after right now.

Yeah. How big is the team these days?

We are approaching 250. Today. Wow. And I, I don't know where we started the year, but I think it was like 80. Last I heard, we were hiring about 50 people a month. You know, we are, we are trying to drag in the absolute best engineering talent possible. You know, one way we look at kind of differentiating ourselves in the industry is, is, you know, you have the Andreessen Horowitz and they're doing a great job of finding companies and doing roll-ups. But I think they're losing a lot of their technical depth. And so we're really focused on building a really, really core engineering team that understands robotics. And robotics is an incredible symphony of detailed, you know, integration between hardware and software to do it well. And we think that we can really differentiate ourselves in the future as we talk about bringing on more products to solve defense needs with robotics.

That makes sense. Little shots fired there from the Bullfrog chief himself. A company that shoots shots. Really, really impressive progress.

Tell us about the round. Yeah. How much did you raise?

We raised $200 million at a $2 billion pre-money valuation. Woo! Let's go.

Hit that gong.

I can say Jordi Hays did our Series A presentation for us. So he was a part of this.

Thank you so much for coming on the show. We should, we should share that presentation at some point.

Yeah, yeah, yeah. No, it's crazy. Right, right before, like pre, pre-TBPN. Yeah. I had the pleasure of helping you guys kind of position what you're working on and you guys have executed it, the strategy to a tee. So it's awesome to see. Going to keep going.

It's really exciting. We think you guys will see another big announcement in the next 12 days.

Months. So congratulations. Amazing. Looking forward to seeing you again. Cheers, Mike. Congrats to the team.

Have a good rest of your day. Let me tell you about MongoDB. What's the only thing faster than the AI market? Your business on MongoDB. Don't just build AI, own the data platform that powers it. People are going, there's a new benchmark. I am a fish and I swim to the worm on a rope. Is it safe? People are asking different models. Gemini nailed it. Apple Foundation model said yes, it's safe. Bad news if you're a fish. Can't recommend the iPhone 16 Pro Max to the fish in the audience. Okay, this is not Apple. This isn't— thought it was Apple. What app is that? Literally, it's somewhat confusing. Locally AI by LM Studio. So what I imagine this is doing is that it's, it's a, an app that integrates with the iOS API. That interfaces with the Apple Foundation model, but I don't know. I might have taken your job. Also, Brian in the chat said, "Black suits, who died?" This is a dark gray suit, okay?

And that's your daily driver. I've been wearing this a lot.

I don't know. I like it. You're in a dark gray suit too.

We're matching.

This is the same tailor's tapestry.

This is the exact same material.

And also, it's not black. It's dark gray.

I know. What does that say about the team's color grading though?

Ooh, yeah, maybe crushed those shadows too much. We'll see. Shots fired. Bring them up. No, I don't know. Did you— speaking of colors, did you know that accountants used to wear green eyeshades? Did you know this?

Well, I've seen like in cartoons.

Yeah, in cartoons. I had no idea where this came from. There's a reel on Instagram that breaks it all down. This blew my mind. I think we should watch it.

You want to learn about— I'd love to, Jordi Hays.

—green eyeshades. This is important because I think you care about the the quality of light that goes into your eyes, and I could see you in the future. Yeah, it's a lens. Look at this. We'll watch this video. It's 3 up from the Apple Foundation Model 1, and, uh, right here, it's the Life Professor. He breaks it down. Cartoons to show that stereotypically someone is an accountant.

Well, that is a green eyeshade that was stereotypically worn by office workers back in the early days of gas and electric lighting.

Came in several multiple versions, including this weird eye hood version that you see here. But the purpose was always the same, which was to reduce the glare from these harsh electric or gas lamps and to improve contrast on the papers that you would be looking at up to 8 to 10 hours a day.

So this became a very common look for any paperwork-heavy profession back in the day.

As lighting became better in offices, the green visor look essentially became obsolete in the actual workforce. However, just like how I talked about locking in at the clock factory, You put on the green visor because you're a white-collar worker? Like, white-collar workers these days don't really have like gear. They don't have accessories. Yeah, there's nothing that's like, oh yeah, that guy's vibe coding and he has the, the, you know, the, the obvious tells.

The new hot accessory is tech neck.

From here, yeah, occasionally hear like politicians described as like a green-eyed visored budget hawk. There is a journalism award that still mentions the green eye shade, but yes. So that's what's going on there with that.

So if you have any— Real grinders had the eyes shut.

Yeah, we got to bring them back. There's a lot of LED lights. Do something to switch it up. Also, Tobi Lütke from Shopify is competing in the 24 Hours of Le Mans this year. Bucket list stuff. Wouldn't have gotten there without DHH's encouragement, guidance. Watch this weekend. You can go watch the 24 Hours of Le Mans.

I love that he didn't just like get into driving on the track every now and then. He thought, I'm gonna go all the way.

Also, this LMP car he's driving looks a little like a Buzz Lightyear mobile. There's some green, or maybe Shrek. Maybe that may— we don't know. We don't know that Shopify didn't pick green because of Shrek. It's possible. When did Shopify found— when was Shopify founded? When did Shrek come come out?

That's the, that is the key question. When was— Shrek had to predate Shopify.

I think so. When did Shrek come out? Okay. Shopify founded 2006. Shrek had been out for 5 years. They could have gone with green because of Shrek. Well, we have a green light on our next guest, Ariel Cohen from Navon. He's the co-founder and CEO. Let's bring him in from the waiting room and talk to him about how Navon is doing. How are you doing? Welcome to the show. Hi, how are you?

Jordi, thank you so much. Gotta be feeling pretty good.

Feeling great. Business is fantastic. Is it time to go do the 24-hour Le Mans? Are you gonna get into racing?

No, this quarter was actually really, really good and the business is doing really amazing.

What's driving it? I mean, I feel like there's so much up and down with SaaS-pocalypse and are people just gonna be using AI for everything or is the economy gonna collapse? Is travel gonna go down of gas prices? Like, what were all the factors of risk and then how did you overcome them?

Yeah, first of all, what drives the business travel is really, you know, the need of people to be there together, to meet face to face, to— we actually, that's our main belief. Our main belief is that people want to meet, close businesses, build teams, get to interviews and to do it face to face. And this is what we are powering And it's so important that you've mentioned gas prices. We are coming out of a quarter that you have TSA strikes, you had gas prices, you had the, you know, you had the major two storms in the US in January, all of these things. And we had an amazing quarter. We grew our usage by 50%. So yeah, it's huge. It's our revenue went up by 40%, our usage by 50%. And this is in this quarter. So it tells you how much it is in important for people to be there together, if they're willing to go through storms while doing that. And this is kind of what drives our business.

Yeah. What else operationally is changing in the age of AI? Are you token maxing? Are you deploying AI agents all over the place? Are you launching AI features for folks or are you taking a more measured approach?

You know, we don't use AI like as a slogan. We always take an approach of, first of all, what our user needs, what our customer needs, and then what can we deliver to them. So even before AI became fashionable, we are like, I'm talking about in 2016, we are heavily using machine learning in the platform to heavily personalize the results that we are giving you, the way that we are saving money, companies when they book travel through us. Companies are saving 15%, 1-5, of their entire travel budget when they use us. Exactly. Yeah. And then, you know, when Gen AI came, we really took it to the next level. So today, the Novant platform, you can go to Novant Edge, which for you, it's kind of for a personal level. If you're a business traveler, but your company, you either don't like the solution or your company is not really managing travel. This is an AI platform. You chat there with, with Avanedge. We know you, we connect to your every loyalty club that you have, and then we are optimizing your entire trip booking. You're telling us you want to be in London and we'll book your flight, your hotel, your restaurant events while you're on the trip, everything while you're on the go. And this is really driven by GenAI. But even when you get support from us, you basically chat with Ava And we completely orchestrated there when we use an AI agent, when we, when we use a real life agent and completely orchestrating both. And this is very, very unique.

Nobody's doing that. Is anything on the go-to-market side changing? I mean, you, it's, it's odd because so much of the pitch is like face-to-face meetings need to happen. And so I imagine a lot of that's not changed. Your sales force is probably having face-to-face meetings with companies that want want the product and want to understand the business before they partner up. But is anything accelerating there? Because obviously when you look at a big 40% jump in revenue, you have to ask, you know, unpack that. Yeah, there are, there are different drivers there.

I would say that AI is actually a big part of this, but from a different reason. Okay. First of all, actually most of the AI companies are using us, right? Including OpenAI and Anthropic and and so on. So that's definitely there. But really the AI driver is that right now in every enterprise you have AI initiative coming from the top. And Havan is one of the only B2B company on an enterprise level that can come and tell you because of AI, I will save your employees, the most important employees in the company, the sales team, the C-level. I will save them a lot of time when they need to change their trip, when something is happening in the airport. Airport, how they are expensing. We are completely, you know, we use our own credit card. You are expensing your entire T&E, including your meals. There is not such thing of itemizing your receipt from the hotel and itemizing the minibar expense and whatever. So we save a lot of time and we save a lot of money. And all of this is driven by AI. So we ride on the AI initiative in the market, in the organization, but also from a go-to-market perspective. We shared it yesterday. On the other side, you know, we sell to the biggest enterprises in the world, but on the other side, PLG, we've more than doubled last year. And this is completely driven by AI because marketing is significantly more efficient when you, when you are actually using AI.

I remember years ago, almost in like the pre-GenAI era, Google launched a feature that would let you sort of almost agentically, but like programmatically call a restaurant and their robot or something would, would talk to the restaurant and try and book you a reservation. And I'm wondering about— obviously there's a ton of low-hanging fruit inside the platform where AI can speed up search and ranking and classification, itemizing the minibar, as you mentioned. There's so many different features. For AI to run wild inside the platform, but are you experiencing opportunities to go outside of the platform? Okay, there's a company, there's a cool hotel, but their system is not integrated, not on any of the travel sites, but we'd love to bring it to that. And we can by autonomously going and writing an integration or even just making a phone call. There's so many times where a receipt didn't come, glitch in their system, I have to call to get the receipt. It's a back and forth. An AI agent should be able to do that. But where is that on the roadmap?

You should— first of all, you should do product management in Navant because you've just described a lot of Navant features and I'll actually explain it. So, you know, everything that you described comes down to the connectivity to the physical world and especially if it is complicated. Travel is super complicated. We have an infrastructure. We actually have by far the best travel infrastructure in the world. It means that we are connected globally to everything that is out there. Every airline through different ways, every hotel, buses, rail globally all over the world. So when you have that connectivity, there are so many things that you can do. But to your point, sometimes there is no API. You cannot change a flight a day before, 24 hours before the flight by going to an API. You have to call the airline. But to your kind of, you kind of touched the use case that we love here. When you pay later, you basically pay in the hotel for the booking that you did. If you'll not show up on time, that's up to 11:00 PM, they're going to cancel the reservation. So you need to call them and reconfirm the reservation. That's right. Give them a credit card and tell them, keep that thing. Now think about it. You are on a flight, stuck on a flight. You know that you're going to be late. The amount of pressure or you're just late or just nobody called on your behalf. We actually have our own AI bot that is calling the hotel automatically, having a discussion over the phone saying, hey, I will arrive. We take the name of the person that confirmed them. We are giving a virtual card to secure the transaction. It's automatic, right? And then we push to the app, to the travel app, to Navant. We push the confirmation and we are saying we talked with that guy in the reception because it's all manual. We talked with that guy in the reception and he secured your room. This is all AI, so no cost. But think about the amazing, amazing, amazing service that we just gave you. And this is when I say I don't see AI as kind of a fashion. I'm looking at like, is there a use case that I can generate value to you, the user, the traveler, the executive assistant, the accountants? Like, you know, we do auto reconciliation for the entire trips of all of the company. So we are saving tons of time for the really big reconcilia— reconciliation teams that the big companies have. So that's kind of how we use the AI. We have our own platform. It's called Navant Cognition that orchestrates it all perfectly between the physical world and our AI capability. Sometimes we use our own model. We actually announced yesterday in the call that 30% of our AI calls are now our own model and not the frontier models or open source. So that actually pushes us farther, it makes us more accurate, faster, raises customer satisfaction, and obviously it's also cheaper.

Yeah, very cool. Do you think now is the right time to start talking about Navant Safety? Your models are getting more and more powerful and maybe it's time for a national discussion about—

No, I'm kidding. Yeah, but you know, it's funny, we have something that's called— and we didn't release it yet, we announced it. Something is called the Travel Club, and there the agent— I'm sometimes joking and I'm saying the agent will do whatever it takes to put me on this flight.

Oh yeah, I mean, there's all those simulations like the vending machine benchmark, things like that, where the models call United and it's like, I just looked up your boss actually, and I'm gonna, I'm gonna leave you a negative review.

I'm gonna, I'm going to— if I'm gonna short your stock, I'm gonna short your stock Besides, don't issue me this refund right now. Help us, help us out with this question. We were talking earlier on the show. It feels like TSA has gotten like remarkably better over the last 10 years. I used to remember like getting to the airport, you know, an hour before the flight and thinking I might miss my flight. Right. You know, or maybe I have to try to ask TSA, hey, can I get through a little faster? Now it seems like like no more than 30 minutes every time. Yeah, it's pretty good. What's your— so, what's your read? It's your theory.

So, you know, I've mentioned Navvana Edge earlier, and you guys have to use it. I'm sure that you're traveling from time to time, so just use it, just try it out. But the cool thing there, it will send you a day before the trip, but also during the trip proactively. It's actually different types of agents there, uh, that will tell you, hey, the TSA pre-line, that's how long it will take you. And that's how long it takes you right now. How long it will take you to get to the airport? Are there specific interruptions right now? All of these things. So we actually know what's going on with TSA. And it was actually interesting because to your point, it's actually significantly better. And obviously I have CLEAR, which is actually much better right now. And with the Face ID and the ID verification, it's actually much easier. But if you think about the— us, we're going to really guide you and tell you what's in the airport right now. During the TSA shutdown, it was not as bad as the media described it. There were a lot of headlines. We actually were watching this and we were saying, yeah, there is some JFK issue right now. But by the way, JFK has issues all the time and it's not as bad as, as they are describing it. So I would actually— that's why the platform is so strong, because instead of taking the headlines, You can actually really see what's going on right now in the context of your trip. Yeah, talk about memory.

All of the AI apps are building memory functionality. People are, you know, collecting markdown files for different skills. And I can imagine a future where I have some sort of memory around my preferences. I like to fly direct. I prefer Burbank to LAX. Possible. I like to fly at these hours. Here's my calendar. And what I really want is just like, I'm going to San Francisco for 2 days. I have to hit this meeting and that meeting. And I would love to just be able to have the integration to Uber to book the car and then the Waymo when I'm in SF, because Waymo works in SF and that's great. And then the hotel and everything just from start to finish booked in one flow automatically. How close are we to that? I imagine—

You know, as I run products at Nevan, and I really, really need to hire you. Really. First of all, we have our own memory component. It's our own development as part of Nevan Cognition. And everything that you described, we do already. Okay, good. But it's actually more— You don't need me.

You don't need me.

Yeah, but maybe you'll tell me how to implement it. But here's the thing. So people really care about their loyalty clubs, right? So we found a way to automatically, when we onboard you to Navant Edge, we automatically figuring out your number for any, for every club. Yeah. Also your, like, your level. Yeah. So that's the part, the first component on memory. Jordi Hays is loyal to his loyalty club.

Your biggest— It's amazing. The Air One one? You always— He'll fly any airline, but— Most people do care. I know, I know. I know it's a big deal. I actually have never in my career been on the road enough to get any sort of status. I think I got like bronze tier on Southwest once. I told Trey Stephens at Founders Fund, he laughed his head off.

You and I are the complete opposite. Maybe I should not hire you.

We avoid travel at all costs. Wait—

weigh in on how many minutes should you get to the airport before? Because I was making the claim that if you don't miss a flight every once in a while, you're not cutting it close enough and you're leaving time on the table. And Jordi's this guy, he gets to the airport 5, sometimes 6 hours early for a domestic flight with no checked bags. And I just think— I just think 30 minutes before 30 minutes before the boarding starts is fine. But who would you side with between me, 30 minutes before boarding starts, Jordi, 5, 6, 7 hours sometimes before? Maybe in between.

I think, I think you're too much for me. I don't want to be there 5 minutes before. I know I would hate that. Okay. But definitely I don't want to spend my entire time.

Well, you know, you know, you know that airport lounges are catering to gamers now. This is in the Wall Street Journal. They're setting up video games.

This is bullish for you. A bit of a video game addict.

Navon, Navon Cognition, preload my Call of Duty account onto the local computer. So when I get there, I'm already queued. I want to be in the lobby.

On one for you. What do you think about, what do you think about vibe coding lounges in airports? Do you think we could see like, you know, you got 5 minutes before your flight, build the next hit app.

Could we see something like that? Well, I think that people went so far with vibe coding. And the reality is that when you want to develop the real thing, you are definitely more efficient today, but you cannot vibe code your way to everything. It's like, so yeah, maybe you cannot vibe everything on the lounge at the airport. But I do see sometimes people vibe coding stuff while they're on a flight. I do see that.

Oh, yeah, yeah, yeah, yeah. That's a big one. And then they carry their MacBook Pros.

It's a thing. It's actually a thing right now.

With the lid open. Looping off the plane. I have friends who have been in that situation many, many times.

Yeah, you do see that.

Yeah, I usually travel with anywhere from like 8 to 16 Mac Minis, and I store them above my seat. I just keep them running the whole time.

Yeah, yeah. And, and you don't bring battery packs, you run it off a gas diesel generator, right? With a little camping stove and propane. You got propane there? I thought that they didn't allow you to get through TSA, but I guess you just go around.

Yeah, they really, they really make it hard to be running, you know, 20 OpenClaw instances. It's rough. It's rough. But hopefully in the future. Anyways, great to hang with you. Congratulations on an awesome quarter. Love seeing the progress.

And congratulations.

Travel. This is travel superintelligence. You're doing it. We love it.

Have a great day, guys. We'll talk to you soon. Goodbye. Let me tell you about Console. Console builds AI agents that automate 70% of IT, HR, and finance support, giving employees instant resolution for access requests and password resets.

A console lounge at the airport. That would be cool. Uh, pull up this Shrek Luce. Shrek Luce. Oh, it has the head. It's got a tail.

Why is it so much more aggressive?

No, no, no, no.

This is a terrible AI image because Shrek is notably round. Like, he is very round. It needs to be more round, not sharp and and, and, uh, and all angular like this. I don't like this Shrek Luce at all. I won't be buying it. I won't be buying this. Looks like a Mansory Shrek Luce. Give me the ultra-rounded, uh, you know, ogre in the swamp. Ideally put the whole car in a swamp. All right, anyway, we're working on that.

We're cranking out the new TBPn meta is, is just generating images. Yeah, with chat and then doing an interview and then coming back, coming back to the image.

Yeah, we do have our Jeff from TensorWave hopping on in about 3 minutes. So we will chat with him in just a few minutes. There is an update on the Anthropic Fable-5 safeguards for frontier LLM development. Quote, they told Wired, "We're changing Fable-5 safeguards for frontier LLM development to make them visible. We made the wrong trade-off and we apologize for not getting the balance right." That was a very surprising thing. A lot of people were asking like, "How did this happen?" Like, it's such an odd decision to secretly degrade the IQ. SemiAnalysis, who is obviously very close to us—

Sorry, I'm not laughing at this. I'm laughing at the latest output.

Yeah.

Most notably, it still is not going to allow you to use—

Yes, but it will at least tell you like, "Hey, this is off limits. You got to contact support or something." So, Dean Ball says it resolves the central concern I had with the fable release. That said, I suspect the residual broken trust and resentment this created will linger and have a blast radius wider than Anthropic. The next time I advocate for any policy justified by AI risk concerns, I almost guarantee you that this incident will be used in rebuttal. And so people are going back and forth on that. But there's, you know, plenty of other news. I wanna see this next output. So let's pull it up. Pull it up.

Martin Casado says— Pull it up.

I can't wait for you to see this one. Martin Casado says, imagine building a computer and not allowing its use in CS research. That's some— that's very dystopian. And that's a good point. Yeah. If you're a young LLM researcher, AI researcher, you want to get on that path. That's what I'm looking for. That's what I'm looking for.

Still too sharp. It looks like Brabus came, took the Luce, and said, we're doing the official Shrek collab.

It's still a little too sharp for me. I don't know. It's still got some— like, the fact that there's even one clean line on that car, not going to be purchasing the Shrek Luce. The SpaceX IPO. The SpaceX IPO is— SpaceX IPO day is going to be lit. Soulja Boy said this in 2015. What did he know? What did he know? What did he know? OK, it doesn't exist. This is fake news. This is fake news. No. According to this proposed note, The depicted 2015 tweet from Soulja Boy does not exist.

If you were going to fake this, why would you not say SpaceX will price at $135 a share? Oh, yeah, yeah, yeah. In the year 2020, in June of 2026.

But we got to apologize to Matt Zeitlin because he said that retail demand for SpaceX shares would be incredibly high. What he said was the idea that SpaceX is being, quote, dumped on retail is a little silly. Retail almost certainly wants to buy like 10x of whatever is available. And I was going back and forth. I was like, okay, like $70 billion, the retail allocation might be like $7 billion or something for this raise. Like, are they really gonna have $70 billion of demand for retail? And that's exactly what they have. No, no, it's more. It's more now. But Reuters said SpaceX IPO draws more than $70 billion in retail orders. And so yes, this is not getting dumped on retail. Retail wants SpaceX stock very clearly. And what is it at now, Tyler?

The latest I saw was $100 billion.

$100 billion of retail demand. They're said to increase capital.

So crazy. Really, really.

Nick Frost over at CoHere is taking the other side. He doesn't like it. He says—

Before, while we're on Matt Zeitlin topic. Yeah, yeah. Matt Zeitlin has a new— Apology. Great new article about all the different spinoff businesses businesses from SpaceX over the years, and how many other kind of like clean energy—

Are you talking about the SpaceX alumni? Like the mafia? Yeah, yeah, yeah.

Full deep dive on all the different spinoff companies.

There's actually a website that— Check it out on Heatmap. So obviously go check it out on Heatmap, but also alumnifounders.com has a website for, for companies that were founded by SpaceX alums. When I first went to this website, it was like a little logo block of like 10 companies. There's now 152 companies founded by SpaceX alumni. $13.5 billion raised. They've created 8,280 jobs. And ranking them, Relativity's raised $1.8 billion. Base Power has raised $1.3 billion. Impulse Space has raised $1 billion. Firefly has raised $800 million in the IPO, a public company now. Apex, Hermeus, Radiant, Castilian. Wow, it's like every company. Synthego, K2 Space, Ursa Major, Varda, obviously, Reliable Robotics, Star Cloud, Revl, so many others. What a fantastic feeder for a mafia. The SpaceX Mafia is something we're gonna be hearing a lot about in the coming years, especially post-IPO, which happens tomorrow. So tune in. Our next guest is Jeff from TensorWave. He's the co-founder and CGO. How are you doing, Jeff?

Welcome to the show. Hey, doing, doing good, John.

How are you? Doing well.

Great to have you. Thanks so much. Since it is your first time on the show, please introduce yourself and tell us a little bit about the company.

Yeah, I'm Jeff Tatarczuk, co-founder and chief growth officer. I like to say chief GPU officer. Here at TensorWave, our focus is on building out AI infrastructure exclusively with AMD.

Okay, so I mean, Chief Growth Officer, that feels like a sales revenue goal. Chief GPU Officer feels like an operations, construction, manufacturing, racking, buying, like the other side of the business. Do you touch both? How deeply integrated are those two problems?

More GPUs in powered shells. More revenue.

That's right. Everybody needs GPUs and powered shells and getting access to them. And I'm the gatekeeper.

Okay, so lay down the AMD thesis. We can go in that, but let's just start with like how you became interested in AMD. Why now? Why? What's the traction? What's the evidence that you're seeing that this is the right choice for your business? Yeah.

So when AI took off back in 2022, 2023, and there was the big supply constraint around getting access to Nvidia chips, the company that we had started previously we were deploying FPGAs at scale. And so we're working with Xilinx and Altera chips and primarily Altera or primarily Xilinx. And when Xilinx got acquired by AMD, our previous company was working very closely with AMD. They would send us all of the latest and greatest chips and we'd get them deployed and debugged for them at scale. And so that's how we forged this relationship early on with AMD. And we saw as soon as there was the supply constraint on the Nvidia side, the opportunity when AMD announced their GPUs back in 2023 for us to be the first to go to market with their chips at scale.

And how are you talking to your customers about what AMD can offer? Like we've seen Semi Analysis has InferenceX, formerly InferenceMax, and had some really unique data showing that for certain certain models, they haven't been able to benchmark everything, especially the proprietary models, but for certain open source architectures, AMD's actually maybe more cost effective on a token per dollar basis. But what data are you leaning on? What questions are people asking you? And then I'm super interested to understand how companies are thinking about where the integration, if they do want to migrate from CUDA, over to AMD, like what goes into that? How costly is that? It was seen as impenetrable, impossible. Now it's looking more and more feasible. But what are companies saying about the trade-offs involved in being on AMD?

Yeah, so in the very beginning when the supply constraint issues happened, people were just looking for more options and AMD was the only option that was available. And so at the very beginning, people wanted to just try it. And see if it worked. And so AMD, when they first launched, launched with the emphasis on scaling out inference at scale. And so being able to support inference and support some of the hyperscalers was their first focus. And then as more labs started getting on and testing it, they realized that there were some gaps in the software. You know, it seemed like the hardware story had gotten out ahead of the software story. And so that's where we were able to help lean in with AMD to help fill some of those gaps by giving them access and helping debug things along the way. And so, yeah, the frustration that customers had around, you know, not wanting to be locked into one vendor, not giving all of their margin to Jensen, forced them and they were willing to look at other options that were available on the market. And so we were there, you know, every step of the way to help fill some of those gaps as customers were considering what is that leap going from CUDA to ROCm on AMD.

And the big public story that played out around that time, you're probably referring to Jordi Hatz flagging some bugs, eventually getting on the phone with Lisa Su, going full founder mode in the CEO seat and watching ROCm develop and actually crush a lot of those bugs. Obviously some people have flawless experiences, some people probably still run into problems every once in a while. That's gonna happen, but a really good, really good turning of the tide, really good different way to interact with the developer community. I'm interested to know, like, how big do you have to be to go multi-platform now? Do you have to be one of the, you know, trillion-dollar labs or are there smaller companies that say, "Hey, we have this, you know, this model that we're fine-tuning. It's pretty good. It's built on some open source architecture." We'd like the flexibility to not be locked into one chip, but is that a million-dollar endeavor, like a year of 10 researchers' time? Like, how are we bucketing this idea of diversifying the options for a company?

Yeah, that was a question that a lot of companies had. What is the switching cost of going from CUDA, going to ROCm?. And we were able to show, I mean, in the early days, right, back in, back in 2023, Databricks was able to show that you could switch from, you can go from NVIDIA and switch to AMD and it works out of the box in the early days. And so being able to showcase that and focus on the most important use cases that the majority of AI companies are using and optimizing those things over time was AMD's focus. And Lisa did a great job. Like you mentioned, she did go into founder mode. She lit a fire over there to make sure that her team was emphasizing building the developer community, building out the resources necessary. And as Anoush over there, who's the CVP of AI—

Yep, we've had him on the show.

He's great. Yeah, he— I love what he says. He says speed has become the moat and they've been able to move even faster now. And with all these coding, code gen tools that are available, they're able to even move even faster to help fill and close the gap on that supposed Coup de mote.

Yeah. A lot of times people sort of put AI as like a monolith. And I mean, we were talking yesterday about how just in the customer support, we were talking about Brett Taylor at Sierra, just in the customer support world, speed might not matter for agentially going and solving a problem in the backend after a customer support ticket closed. But if you're, you're on the phone with someone and you need a real-time voice model, it might actually make sense to co-locate that or distribute that. We talked to Matthew Prince from Cloudflare about this, that there are certain models that need to be in certain places. Speed might be more important. Others, it's just raw horsepower. It's gonna take an hour anyway, send it off to the big data center. But I'm interested in the other pockets, like where is AMD poised already in the lead or poised to take the lead in images or video or voice or real-time or agentic work or integration between the GPU and the CPU. Like, is there something where you're particularly excited? I mean, I noticed Luma AI is a customer. They obviously do creative work. And I'm just wondering if there's something where you're like, oh yeah, like they've, they've nailed— like this particular architecture is great. And like, obviously there's other work to be done. There's more cost on certain workflows because those were developed at DeepMind on TPU or at some other company on NVIDIA. But is there any place where you're like, ah, AMD's really nailing it in this subcategory of AI?

Yeah, well, they started off great on the hardware side and focusing on adding more HBM than what you can find on a comparable NVIDIA chip. And so having more HBM allows for workloads that are running larger models to run on fewer GPUs and to do it faster.

Yeah. So you have the CUDA moat, as you put it. Is it, is there like a, who's building the bridge? Like you need a bridge to go over a moat. Like I imagine that there's companies, you're probably among them, that are specifically building software that can just translate any architecture from NVIDIA to, to AMD? Like, how close are we to just that being one click of a button?

Yeah, well, one of the things that I found as we were digging into, you know, understanding more about what NVIDIA was doing, understanding what that CUDA mode actually was, and talking to a lot of the original engineers that helped build CUDA from the ground up. Yeah. One of the guys that's on our team, Greg Diamos, was there in the very beginning building CUDA at NVIDIA. And he had always said that CUDA was always built to go beyond CUDA and to scale out infrastructure. And so one of the things that, that we've— as we've talked to people, they've said that CUDA is not the moat at NVIDIA. Their ecosystem is the moat, and the resources and developers and libraries that they have poured into over the last 10, 15 years is the moat. And that's the maturity that exists today. And so as AMD is building that out and as we are supporting them, We find the important thing is to focus on building out an ecosystem to support the— to support AMD. We don't focus on software ourselves, but we have a lot of other companies that are trying to solve that problem, to create a heterogeneous stack, to be able to switch between chips. And so bringing together those people, showcasing all the different layers that are existing beyond the NVIDIA ecosystem is what my current focus is on today.

Where, uh, where are you setting up your supercomputers? Some people call them data centers, we call them supercomputers.

Yeah, we have supercomputers all over. Um, we have in Arizona, Florida, Pittsburgh, and have data centers going up all over North America currently.

Amazing. Awesome. Tell us about the round. How much did you raise?

Yeah, we raised $350 million, Series B. Here we go.

Congratulations. And thank you so much for taking the time to come chat with us.

Yeah, great to meet you, Jeff. Great to meet you. Congrats to the team on the progress.

Thanks, everyone. Have a good one. Let me tell you about CrowdStrike. Your business is AI, their business is securing it. CrowdStrike secures AI and stops breaches. Do you have another AI image for me? Or are we moving on to our next guest, Matthew Joseph from Minerva? He's the co-founder and CEO, and he's in the waiting room. We'll get back to the next version of Luce after we talk to Minerva founder Matthew Joseph. How are you doing, Matthew? Good to meet you. Good. How are you guys? We're doing fantastic. Great. Thanks so much for hopping on. Please introduce yourself and the company.

Yeah, so I'm MJ. I'm the CTO co-founder of Minerva, as you guys said before. Yeah, it's an AI platform for consumer marketing. So basically we help consumer brands unify their first-party data, which sits in like, you know, 20 different data systems like Klaviyo, Shopify, whatever. We standardize it, we resolve it, we create kind of a golden record, and then we optimize sales and marketing on top of that using a bunch of factory agents that we developed.

Do you— how much of this is like data for like the CFO of a consumer brand to understand return on ad spend versus like actually taking it through to a creative tool, closing the loop and integrating with like generative content writing marketing briefs, maybe even generating copy, probably not in your models directly, but throughout that and harnessing everything together.

Yeah, it's actually more data on the consumer. And so we have a really, really expansive data set on 270 million Americans, including an identity graph, which is also pretty expansive. We have about like 10 billion pieces of PII. Sure. And so we help the CMO and the performance marketer kind of understand their consumer base better.. So when you're looking at the funnel and seeing where people are dropping off, or you're looking at like the email copy that you're creating, you need to understand who you're actually sending that message to. We understand that. So we can kind of do it both ways, right? We can be generative and say, hey, you're marketing to this kind of person, so you should say this. Yeah. And then we can do the opposite where we can kind of diagnose inefficiencies based off of demographics.

What is— give us sort of a case study from the NBA. Like NBA is in the news, obviously. Seems like people show up with plenty of money. I don't know that you need any. Like, what questions do you need other than what's your credit card number? Because I got a $100,000 ticket here for you and you'll pay whatever it takes to get to this game.

Yeah, with the NBA, we're part of their FanGraph. So we power a lot of their consumer analytics that they end up selling to all their downstream teams. But even at the team level, there's a ton of different case studies that we are starting to pop up right now where, you know, if you have super expensive seating but you're in Oklahoma City, you might need to find people that are out of town from a road game that might want to buy that seat. Right.

Oh, sure.

So flavors like, like case studies like that are what we're gunning for. But with the NBA in general, it's mostly powering the consumer analytics platform.

How do you guys actually interact with different ad platforms? Let's say Meta or AdWords, TikTok. —could say. Yeah.

So with Meta, we'll find audiences that we know will convert at a very, very high rate for specific products because we have an understanding of kind of the right demographic to market to. And so we'll create lookalike audiences that are seeds on Meta, and they'll generally outperform broad targeting by a pretty fair margin. We've seen in luxury travel, luxury goods, our ROAS improvements are around like 3 to 5x broad targeting.

Wow. How quickly can you prove to companies that the product works?

Yeah, it depends on kind of the case study. For data, you can immediately identify inefficiencies in your pipelines. So people in a day will be like, oh, you guys are like absolutely changing where we're allocating spend. For things like meta, things like direct mail campaigns, it takes a little bit longer to get statsig, but you can get you can get an idea of how well you're actually going to be performing pretty quickly. So we generally see within a week that our data and our audiences perform better.

Yeah. What are the biggest, broadest meta trends in consumer marketing these days? I mean, we went through the Facebook direct targeting, direct response boom. We went through the influencer buy YouTube ads boom. But in the age of AI, I think everyone's everyone's expecting just like AI-generated ads, but is there some other trend where like you can target a smaller audience so you can go on a more niche platform? So you might see more uplift from a platform that previously wasn't ROI positive. Like what else are you seeing in terms of just broad marketing trends?

Yeah, I think a pretty interesting one right now is retargeting on AI-driven traffic. Okay. So a person who comes in, through like a ChatGPT platform or like Claude or whatever will behave very differently than a person who came in through direct. And so being able to understand that funnel is super, super important. So retargeting on top of that is something that's pretty interesting and kind of booming right now.

Do you have a thesis on how or why they behave differently? It seems like it's higher conversion. Like if I'm landing on a web page after doing a deep research report on a product, I've learned I learned so much, so I'm probably not going to need to scroll the product detail page. I'm probably either just want to jump to checkout or I want to do something else to verify something that I learned in the chat app.

Yeah, I mean, I think a lot of these trends are happening kind of in the last year. And I still think the average user of AI, especially for kind of deep research, is probably a little bit more advanced than the 65-year-old in Oklahoma City. I keep picking on Oklahoma City, so I'm sorry about that. But I think it does boil down to demographics. In 3 years, this can totally change, right? Like when AI is truly commonplace in every single household, which is getting there. Yeah. But, you know, over time this will definitely change.

Yeah. What percentage of the creative that your customers are running do you think is heavily, heavily generated with AI or influenced by AI or created in process?

Yeah, I'm seeing a huge trend towards AI-generated creative. There's a lot of weird little things that you need to keep in mind when you're using AI-generated creative, like where you're placing text relative to the rest of the creative. I mean, it's really hard to execute well. This is something that we are looking into right now as we have a very good read on the demographics. And so creative is kind of the next logical piece there, right? In terms of presenting—

Yeah, like if you're going to create a lookalike audience for somebody and then you're telling them like, hey, you should put spend behind this, you should— you're also in a pretty good position to theoretically, as models get better, say like, here's 100 pieces of content that you can run against them and then they can select, okay, what fits our brand, what doesn't?

Yeah, that's exactly right. It's like the next logical step, right? Just automating the entire process from product to cohort fit to creative fit, we're in a position to do all of that. And I'm happy to be part of kind of like the rise in creative generated by AI.

I wonder if you find a lookalike audience of people that just love AI slop and you just feed them the sloppiest ads. And then people that— because a lot of people don't like the AI slop, but it would probably some niche audience out there that's like, it would stop you.

I'm buying if there are 6 finger. Matt, you'll appreciate this. We obviously love ads, firm supporters of ads, all this stuff. But somehow I ended up in a holdout on—

Don't say this, they're going to take you out.

I ended up in a holdout on Meta where I do not get any ads. I'm in the probably 0.05% of people that they're testing engagement with. And I'm like, I'm the wrong guy to test on. But I just realized one day I was I was like, wait, I have not gotten an ad on Instagram in months. And still to this day, absolutely nothing.

Imagine they run this experiment, whole hold group, no ads, and they just find like massive depression, just massive rates of depression. They're like, we gotta bring back ads. That'd be the most bollocks thing ever. Ads are good for mental health. Maybe, maybe. You know, consumerism, shop— Talk about the round. Shop therapy or whatever.

Yeah. So we raised—

so we raised $20 mil between the seed and A from DPC, from Engato, from the general partnership, from Polyview, many other investors that I'm probably forgetting to name, but very, very stoked for all their support. We love all the teams we work with.

Fantastic. Very cool.

I got some brands to connect you to. Yeah. So send me an email. Send me an email. I'll— with the quick pitch, I'll CC you in with some. That'd be great. Will do. Awesome. Thank you. Have a good rest of your day.

Nixon 5, let's go. Let's go.

Let me tell you about Public.com, investing for those who take it seriously. Stocks, options, bonds, crypto, treasuries, and more with great customer service.

For San Antonio, I feel like everyone, like, I feel like it's relatively easy to rally these people people in our world around New York City.

Yeah. Oh, sure, sure. Like, it's easy to be a bandwagon fan for the Knicks because New York, you visited, you've been there.

Big Apple. That makes it hard to really rally around San Antonio. Have you been to San Antonio? No. Would you like to go to San Antonio?

Sure, sure. Strong. Where is San Antonio?

It's in— TBPn San Antonio. It's in Texas.

We're doing it. But I always think Houston Houston, Dallas, is San Antonio the third in that, in that tri-state metro, the three cities that connect?

We're sending Tyler to San Antonio to figure out what's going on.

Yeah. What's in the water. Get to the bottom of it. I don't know. Well, Creative Artists Agency, Hollywood's biggest talent representative and our agent, is starting a new company to invest in YouTube stars and podcasters who want to be the next Oprah Winfrey Winfrey or Reese Witherspoon. CAA and the Integrated Media Company, uh, Integrated Media Company, an investment arm of private equity firm TPG— that's Texas Pacific Group formerly— have pledged $250 million to Compound Creative Holdings, a new company that will acquire and invest in businesses founded by entertainers who got their start online. Compound will be led by Tucker Brown, an investment banker who has already helped raise money for the YouTube trickshot masters at Dude Perfect and the liberal podcasters at the Midas Touch Network. Interesting. Are you familiar with Midas Touch, with the M-E-I-D-A-S? At one point, this podcast was bigger than Joe Rogan. It jumped to the top. I did not realize that it was a political show. The network describes represents itself as doing pro-democracy journalism. Is this— I'm so confused because the YouTube— yeah, this is definitely it. I don't know. I thought it was something else. I thought it was more in like the Alex Hormozy category because Midas Touch, of course, the golden touch, Midas Touch. I thought it was more about like wealth creation. Interesting.

CIA represents many of these companies. Very interested to see what deals they do here. Cool. And we're working on getting and the team on the show to talk about it.

It turned to Spotter, like outlets, outlets like Spotter to raise money by trading a share of their future earnings on YouTube. Yeah, if you have a YouTube video and it's doing really well and it's getting views, bringing in ad revenue every month, there's a company called Spotter where you can go and sell the future revenue stream for cash today to finance your operation or just take some chips off the table, much like selling your, you know,— the rights to your record collection, essentially. There's a lot of capital in the market generally, and a lot of interest in creator businesses, but there aren't many educated capital platforms dedicated to invest in the category specifically. I agree with that. There's only a few that do these types of deals. Obviously, Barstool raised money multiple times, and there's been others that have been successful, but it is a thin crowd. Compound will focus on creators who have built media businesses generating at least tens of millions of dollars in sales and could use capital to expand their operations. The company isn't interested in just buying a share of revenue from YouTube or Instagram and tying their fate to an algorithm. They want to buy a piece of the whole holding company, the media company. Brown estimates there are hundreds of creators that fit their initial investment strategy. That talent can be born on YouTube, TikTok, Instagram, Spotify, or Substack. That's interesting. Interesting. We haven't really seen a Substack raise money. I guess the Free Press was on Substack for a long time. I think they still might be and did raise money and scale up and hire a bunch of journalists and writers. Interesting. Leaders from CIA and IMC, including CIA co-chairman Kevin, President Jim, said, uh, we'll sit on an executive committee that oversees Compound. Celebrities of all stripes have embraced entrepreneurship of online creators starting businesses in entertainment, fashion, food, retail. Witherspoon— Reese Witherspoon sold her company Hello Sunshine in a deal that valued the firm at as much as $900 million, while George Clooney and other clients sold his tequila company Casamigos for, I think, also around that number. A huge, huge outcome. We've been looking for ways as an agency to double down on our efforts outside of core representation. Presentation. And Paki McCormick gives Tucker a shout out. He says, Tucker has been one of my best friends since high school. I didn't realize they went back so far. Overnight, the single best person in the world to be doing this. Agree with that. He's going to crush it 100%. People always called him the less hot version of me. That's very funny if you know them both. Anyway, Lionsgate has taken an equity interest in Runway, uh, and they will launching a new joint venture program to develop and produce new intellectual property. This is going to be wildly popular amongst film fans and cinephiles, I'm sure. Lionsgate and leading generative AI company— I love Runway, Crystal Ball, former guest of the show. Runway announced an expansion of their strategic collaboration, deepening the company's shared commitment to a future where AI and entertainment grow together. That's good. The preferred partnership with Runway is part of Lionsgate's multifaceted AI strategy under the leadership of Chief AI Officer Kathleen Grace and the studio's AI steering committee. So it'll be interesting to see what Lionsgate does. Lionsgate Films, can you name one? I don't think I can name one. They did Saw.

Interesting. There you go.

What else did they do? American Psycho, that's a meme. The Punisher, Hostel, Rambo. It'd be interesting, it'd be interesting to rank movies by how many they did. The whole John Wick.

Listen to this, John. Yeah, John Wick, uh, ranking movies by how many memes they've created, right? Because like American Psycho, tons, tons.

John Wick has a meme.

What's that? It's the—

I'm thinking I'm back. Yeah, I'm thinking I'm back. That's the name. Anyway, Sicario, Hitman's Bodyguard, Den of Thieves, Greenland. Good stuff. Will be interesting to see what they're doing. Maybe just background replacement. Maybe John Wick 6 will be entirely AI-generated, one-shotted by the next model. You never know.

Lionsgate's going all in. Keanu Reeves cannot be AI-generated. I think every other actor.

Corridor Crew would like to differ. They got a— they hired a Keanu Reeves impersonator to do a— like a— basically, the whole shtick was like they're gonna try and create a viral video and someone's robbing a, like, a corner store, like a 7-Eleven, and the impersonator, the Keanu Reeves impersonator, stops the assailant and then they use used a deepfake to change the face of the impersonator to Keanu Reeves to make a very convincing, to make a very convincing Keanu Reeves saves the day video that I think went viral and confused a lot of people, but was a lot of fun. They did a whole bunch of back, like behind the scenes content around that. Anyway, we have our next guest, Ade Aja from Base10, who's the managing director managing partner. How are you doing? Welcome to the show. Good morning, guys. Thank you for having me. Thank you so much.

Let's get right to it. How much did you raise? And John, give me the hardest gong hit of your life.

Amp it up a little bit. 850 million. Boom.

Clean. Congratulations.

Great to have you on the show.

$2.6 billion in assets. But it is your first time on the show. So why don't we go back in time a little bit and give us some of your background and how you got into investing, how you scaled up this fund.

Yeah, so the accent is Spanish, the name is Nigerian. Early childhood in Nigeria, then southern Spain. And like you guys, I was a founder. So I went to college in Madrid. I started my first company when I was in college in Madrid. Which was not a very popular occupation in Spain at the time.

I thought you were saying that the company wasn't very popular, like you got into some hot water. Maybe you're— The company—

being a founder was not very popular. The company became very popular. So we were 20. We became the equivalent of Facebook in Spain. And to give you a sense, we started in '05. We sold to Telefónica in 2010. And when we sold, we were about 60% of daily internet traffic in Spain. Whoa, wow.

So that was quite the journey. What was it called? Sorry that I'm not— Twenti.

So like 20, because it was for 20 years old, but the Hispanic phonetic pronunciation, T-U-E-N-T-I.

I remember, I remember now it's all coming back. It's all coming back 'cause I have some very good, family friends in Spain. I've spent a decent amount of time there, including a summer in Madrid at one point. And yeah, I remember it. Very, very cool product.

So talk about the strategy. $850 million. That's a lot of capital for seed, Series A, Series B. Jeff Bezos raised a $12 billion Series B today. Wow. And there's this odd K-shaped dynamic in venture where The letters don't really mean anything anymore because you can raise a $2 million, $10 million Series B, $10 billion Series B. Where are you seeing opportunity? What is the early, medium stage market? What are you looking for in the companies that you fund?

Yeah, so we say that we invest in automation for the real economy. And to give you a sense, when we launched the fund in 2018, we actually said, hey, we're going to invest in applied AI. For the real economy and the feedback from the LPs was like, hey, applied AI is not a thing. Yeah. So it sounds niche. Yeah. So like, why don't you like do something real? Yeah. So we call it automation. We're like, okay, well, we just think that people are going to use a lot more AI in like logistics and restaurants and transportation. So it's no longer niche. Yeah. And now, as you were saying, we have this crazy dynamic where like, you know, we, we, we sold 20 for $100 million and that was a big deal. Now that is like, there's some YC companies that raise that, right?

Not just raise at a $100 million valuation, but just raise $100 million.

Raise $100 million. Exactly.

So it's a, it's a completely new world, which is fun. Like that, that, that's fantastic. And what I tell the founders we invest in today, It's like, man, I'm so happy I don't have to compete with you because you guys are so good and there is so much happening. So I would say applied AI to the real economy, which these days means a lot.

How are you thinking about services as software, that whole trend? It does feel like there's an opportunity to not train a foundation model, not build a robot and be very capital light, go apply the tools that are already well beyond what's actually implemented in the real world, close the capability overhang, make a bunch of money in the process. Maybe you build some software, maybe you get some customers on long contracts, and the business works profitable. That sounds great. At the same time, you sort of veer into, like, are you just a consulting firm? How are you talking to founders? What are you seeing? What models are working? Market?

Yeah, great question. So there are a lot of labels like, you know, AI-enabled services, AI-first applications. I like to spend a lot of time talking to clients, like, you know, calling the logistic companies, the transportation companies, and they don't really care what the label is. They just want it done, right? It's like, hey, at the end of the day, I need to move this package from here to here. Yeah, I need to like get this cash from this client. What can you offer me that works? Yeah, that I can trust and that is fast and that is not terribly expensive. So we try to like avoid a little bit of the labels and then like go to what does a client want and need at the end. And I think there are going to be many business models that are successful. Is changing very fast. Yeah.

For my own curiosity, what— how is Spain thinking about AI? Do they have a sovereign AI strategy? I don't know how closely you follow it. It must be at least—

we could be here all day. Look, we are a little bit more technology forward than when we started 20. But not quite San Francisco yet. I will say that I get a lot of text messages from my friends back in my hometown who are running restaurants and hotels being like, "Hey, have you heard of this thing called cloud?" And that fills me with joy. As a country, honestly, we still need to get our strategy together. And although the culture is changing, and actually one thing that makes me very proud is that in the last year we have founded 4 different Spanish teams here in San Francisco that have moved to San Francisco to start companies. Interesting. And they're crushing it. Yeah, yeah. Back home, we still talk a little bit more about the World Cup than about AI. So we need to like change that.

Yeah, well that's I guess fair for this next month. Talk to us about—

Yeah, then after that. Yeah. Talk to us about coffee. How does that fit into a portfolio? I mean, is there a linkage? Is it just a good business by itself? Is it just a passion of yours? What's the coffee thesis?

Yeah, so look, we are a research-first fund. And we're always like, hey, what is going to happen to the construction market or the logistics market with robots? And as you guys know, all that's well and good, and then really early-stage investor is all about the people. So when we met the Blank Street Coffee guys, they were like, look, coffee's big. We're like, okay, we can agree on that. Like, you know, that we don't need a lot of research with that. But honestly, it's one of the best pitches I have ever like heard in my life. They were just like, look, We're both like sons of immigrants and like immigrants ourselves. And like, did you know that 80% of coffee carts in New York are like run over by immigrants and it's a great path to like building wealth for the family and staying in the country. And like, that's what we want to power. And we're like, okay, and the automation? And they were like, well, the coffee cart is going to be great and like completely automated. And we're like, okay, we don't exactly know what that means right now. 'But these guys are awesome, so we should write the best tech.' And that turned out to go great.

Yeah. It's funny because there are examples of similar stories and then it not going that great. So, yeah, it's so difficult to fix the, to square all of that when you see opportunity, but ultimately, you, you find the alignment and then everything works out. What do you— do you think in deployment pacing, do you think in 2-year deployment, 10-year horizon, is that changing in the modern era with so much uncertainty? I mean, we talked to Matthew Prince from Cloudflare about— and we talked to a number of people about it's hard to plan 2 years, 3 years in advance if you're a public company. CEO, also hard to plan ahead if you're a venture capitalist, but that's what you're paid to do.

Yeah, I know. Look, we have stayed remarkably consistent on that. All our funds have been 2.5 years to 3 years deployed. We are going to do the same thing with this set of funds. It is more tempting than ever to deploy very fast. And I'm an optimist. So we are currently in the— this is Y Combinator week, right? So I want to back like 50% of founders that I make or more. And my team is like, "Adrián, stop. We can't do that." And you're like, "You're right, you're right. It's just— ah, there's so much to build and these founders are so good." But we have built systems internally to make sure that we have that discipline because to your point, although it's very tempting and a lot of things are growing very fast, there is so much change that we want to get a little bit of feedback loop, a little bit of data of, hey, okay, we are 3.5 years into AI or whatnot. Yeah. What happens in year 5? Yep.

And we want exposure to that as well. Fantastic. Well, thank you so much for coming on the show. Appreciate you taking the time. Congratulations on the fundraise.

Let's do it again soon.

Thank you guys. Have a good one. Goodbye. Cheers. Our next guest is already in the waiting room. We'll bring in—

60% of internet traffic on a company built in college.

Insane. Crazy. Insane. Yeah, that's such a high— such an impactful company. Well, yeah, Spain was locked in doomscrolling. He invented the Spanish doomscroll. No, I'm sure it was, uh, it was probably messaging-driven, probably, uh, more post-driven. But that's— yeah, that is remarkable. Uh, anyway, we have our next guest, Jeremy Franklin, from Fundamental, co-founder and CEO, in the waiting room.

Let's bring him in. What's going on?

How you doing, Jeremy? Good to meet you guys.

Thank you for having me on.

Uh, I want to get straight into world World Cup predictions, wagers, all of this. But first, introduce yourself and the company a little bit. What are you working on? And then let's go through the broader news and applications. Awesome. Yeah, I'm Jeremy.

I'm the CEO and co-founder at Fundamental. Fundamental builds a foundation model for tabular data. So when most people think about the AI boom, they think about LLMs, you know, like ChatGPT, Claude. But what's less appreciated is that LLMs work well with unstructured data. So they might text, audio, video, but the global economy and enterprises really run on structured data, on tables such as spreadsheets, databases, CRMs. And that's the part of the enterprise that really never had a ChatGPT moment. And that's a market opportunity that we're focusing on. And so what we're doing for tables is what ChatGPT did for language. We pre-trained the model on millions of tables to understand how the world works in rows and columns. And so we created one model that generalizes across industries and use cases to allow companies to make better predictions and decisions on their data. Yep.

Okay. So everyone is going to say, oh, the frontier models, like they can spin up a spreadsheet, they can speak CSV, they can write Python that generates creates and manages tables and tabular data. And the frontier is the frontier. It's always the best. But you're going to effectively be benchmarking your product against the frontier models around the World Cup. Walk us through the strategy, the wager, what's actually at stake there. Yeah, so I grew up in Belgium and I'm a huge soccer fan and a lot of us at the company big soccer fans.

And so what we realized is that, um, you know, historical team performances in games can be structured in rows and columns. And so we started thinking around the office, like, can we actually use our model Nexus to make predictions on the World Cup? And said, like, you know, maybe betting markets are off and we are right. So, like, we wanted to see how good our model was. And the funny thing is that at the core of our model, you know, our company is focused on enterprises, so we're not really a consumer a football-focused company, but we thought we could have some fun with it. And so before doing anything, we wanted to see how good our model was historically. And so what we did is we trained a model to win data up until the 2022 World Cup. And so data such as which— who won in different international games, the quality of the teams, whether games were home or abroad. And we ran more than 20,000 simulations.. And what we— what our model Nexus predicted or identified is that Argentina would be the eventual winner way before bookmakers did. So if you look at what the market actually predicted, the market predicted that Brazil was going to win the 2022 World Cup up until the quarterfinals when Brazil got knocked out. And our model was much more accurate than that. And so we said, you know what, let's apply the same thing for the 2026 World Cup. And our model already differs meaningfully from what betting markets are predicting.

So I'm excited to see what— win, right?

Yeah, I wish. America, America is funny.

America has a 0.5% chance of winning the World Cup according to our model. There's a chance. I like those odds. I'll take those odds any day.

John loves betting on himself. Yes. Uh, how's Belgium fare?

Belgium is at slightly less than 2%, but I honestly have, uh, I have faith.

Okay, and then Belgium. Yeah, what's the population of Belgium roughly? 11 million. How many? 11 million. 11 million. So population of 11 million, absolutely mogging. Yeah, population.

So, so are you also benchmarking the other frontier models, just going to the different models, open source models, and just asking them the same questions. Who do you predict will win? Build a probability table and then benchmark those as well. Because I think you're going to want to not just say that you're better than the sportsbooks, but you're also going to want to say you're better than the frontier models, correct?

Yeah, but the frontier models themselves don't really work well with predictions, right?

But they can take a— I mean, they can predict the next token, so they can take a wild guess. Then if you smoke them, it looks good, right?

Yeah, we could try that. We've not done that. We could definitely try it. But it's funny because the frontier models, especially for predictions, they sound right, but they are not actually correct though.

Yeah. And with ChatGPT, it has the personalization. And in the personalization, I said never bet against America. So I'm sure It would just say like 100%. So yeah, you probably smoke it.

Who, who are like, how are you, who are you projecting to win? What are kind of like the kind of key predictions that, that we should be tracking as the World Cup kicks off?

So the betting markets have Spain and France neck and neck and we heavily favor Spain over France. The market also predicts that Argentina, I think, has like the fifth largest chance of winning. According to the market, and we predict them as strong runner-ups.

Oh, interesting. OK, so Spain, Argentina— Spain winning would look really good for you relative to the sportsbook.

Correct. But I would be happy if Messi wins his second World Cup.

Sure. Tell us about the company. You recently raised some money.

Yeah, so we raised $275 million.

Who would you raise it from?

We raised it from Oak, Valor, Battery Ventures, Salesforce, and then a bunch of big angels like Alvin Srinivasan from Perplexity, Asaf Rappoport from Wiz, and a few others.

That's fantastic. Strong. Very strong. Keep us posted as as the, the World Cup progresses, uh, and, uh, we're rooting— I'm not even rooting for any team, I'm just rooting for your predictions. I want you to know that.

Yeah, I appreciate it. And everyone can go and see them on soccer.fundamental.tech.

Amazing. soccer.fundamental.tech. Very cool.

Great to meet you, Jeremy.

It'll be fun to follow along.

We'll talk to you soon. Have a good rest of your day. Uh, in other fundraising news, Thrive is out with $10 billion. Shit. Got a little mogged by Bezos with the $12 billion raise, but not bad. Not bad. Thrive X and it's $10 billion, exceeding $10 billion, the largest ever, nearly double the size of the last fund for Thrive. $1 billion for early stage, $9 billion for growth. Biggest bets, just to remind you, OpenAI, Stripe, SpaceX, Databricks, and/or Cursor are all in the Thrive portfolio. And a bunch of other companies. And so those have become the defining private companies of the current cycle. The strategy's concentrated, founder loyal, allergic to indexing the market. The Fifth Avenue strategy, as discussed on Invest Like the Best, and Patrick O'Shaughnessy is chiming in. Josh Kushner says, in markets, to go long on something is to bet it grows more valuable over time. Much of the conversation today is short on here Humans wagering that AI makes people redundant. We believe the opposite is true for industries Thrive Holdings operates in. And Patrick O'Shaughnessy shares a transcript from Invest Like the Best. They said the North Star for the next decade of Thrive is if we are a company and we have a product, what are the products that we can create that make us look more like the businesses that we ultimately invest in? The second part of the story is after we invest I was in OpenAI, the business had $50 million in API revenue and several, and several of us went around New York City and started pitching a bunch of the private equity firms on this idea that they could use the API to create greater efficiency within their businesses. We were somewhat surprised by the fact that we did not get meaningful traction around these ideas. So Kareem, to his credit, said, let's just start doing this ourselves. We've always looked at the world that we've lived in one that needed to disrupt from an outside-in perspective. If you were to think about the paradigm shift of reinforcement learning, the things that you need are both data, and that is proprietary to the company, but also the experts that exist at the company to fine-tune the model. We believe that going forward disruption will happen from the inside out. We set up this permanent capital vehicle that enables us to buy these businesses and hold them in perpetuity Because if you actually have a differentiated, unique lens and cost of capital around these businesses, and you're able to transform them in ways in which you ultimately want to hold onto them forever. So congrats to Thrive on the new round.

Congrats to the whole team. Yeah. Fantastic. Great stuff.

The new fund. Anyway, leave us 5 stars on Apple Podcasts and Spotify, sign up for our newsletter at tbpn.com, and we will see you tomorrow for the SpaceX IPO. We're gonna just be hanging out. We're hanging out.

We're gonna have the stock price on the big board.

Be fun. We'll see tomorrow. Goodbye. Throwing flashbang.