Verizon vs Salesforce, Blue Origin's Test, Robot Marathon | Signüll, Ethan Ding, Matt McKinney, Errik Anderson, Pippa Lamb & James Wise

You're watching TVBN.

Today is Monday, April 20th, 2026. We are live from the TVBN Altar Dome, the temple of technology, the fortress of finance, the capital of capital. We have a great show for you today, folks. Absolutely incredible edition of The Wall Street Journal over the weekend. I saw these hit the app, but they're in print today, Monday, April 20th. There's two articles. From two tech CEOs that run companies that are almost identically valued. Salesforce is, I think, $150 billion, Verizon's $190 billion. They're right in that sweet spot.

And I'm assuming they have the exact same point of view, kind of same market outlook.

They have exactly opposite points of view. And I thought it was interesting seeing what does a victim of the SaaS-pocalypse have to say about AI? And then what does someone who is the most most resilient to the SaaSpocalypse, I have to say, and potentially is like, let me in. I want some of the drama. I want the smoke. And so Marc Benioff is who I'm talking about at Salesforce. So he says the software bears are all wrong about Salesforce. We had a bear, but it was too terrifying.

It was too scary.

So we skip it. But he says, people think we have our back against the wall, but customers aren't replacing offerings with AI. They aren't replacing Salesforce with AI. It out. Of course, everyone will say, well, yet. But let's dig into Marc Benioff's argument and how he's processing the SaaS-pocalypse and what he's going to do about it.

Crazy opening. Marc Benioff has some problems. His enterprise software company Salesforce is the biggest name in a category that Wall Street thinks may get decimated by artificial intelligence. Its business model is centered on selling software to large companies on a per-employee basis. But many of those firms are expected to downsize as AI agents become increasingly proficient at performing real-world tasks. It's a daunting double bind, and it isn't even the worst-case scenario. Salesforce stock is down a mere 28% year to date. The hardest-hit software-as-a-service companies are down about twice that much on similar fears. But Benioff thinks the bears have it wrong about the SaaS-pocalypse thesis generally, and especially about Salesforce. AI, he says, is making Salesforce more valuable to its customers than ever. The leading AI labs couldn't replace what Salesforce offers even if they wanted to. And they would rather partner with him for now. Anyway, nor could— he's like, I dare you.

I dare you.

I dare you. He's staring Benioff off. I wish we had Benioff on right now. It's a joy to speak with him. But continuing, nor could customers easily vibe code their own sales management software that could compete with Salesforce on security, compliance, and other vital features. "People think we have our back against the wall when in fact the opportunity has never been greater," Benioff said in an interview. An early investor in Anthropic, Salesforce has been developing and pushing its own AI tools for years. By the end of this year, it plans to unveil a new AI platform that automatically studies its users and takes actions on their behalf. Codename Agent Albert.

Hmm.

It's kind of cool to put agent in the name instead of just Albert. Agent Albert.

Well, they had Einstein, right? So Albert Einstein is their whole shtick here.

Agent Albert is the culmination of an effort that began 3 years ago when Benioff galvanized— when Benioff, galvanized by the debut of ChatGPT, instituted a standing Saturday meeting to accelerate Salesforce AI efforts. So let's lock in Saturday standup, everyone.

An earlier flagship product of that push, Agentforce, has been somewhat slow to gain traction. Launched in 2024, it is used by 23,000 customers of a —total base of 150,000. So what is that? A little under 20% of the customer base is using Agents.

They're building a product.

I remember Jason Lemkin came on the show, and he was like, I use AgentForce. It's solid. Yeah. Like, he's like, it won us back a customer that, like, we weren't reaching out to. Yeah, yeah, yeah. And it, like, it made me money. Yeah. So he was an AgentForce enthusiast.

And so Salesforce's yearly revenue growth at 10% is down somewhat from recent years. But interestingly, the decrease— deceleration in Salesforce's revenue growth started back in 2022, maybe even a little bit earlier. If we scroll down on this, on this image, you can see back in 2012 they were growing 36%, then as recently as 2020 they were growing 28.7%. There was acceleration in 2022 to 24%, but then 2023, 18%, 2024, 11%. 2025, 8%, and then we're actually seeing some reacceleration this year from 8.7 to 9.6 to 10.8. And so it feels like definitely too soon for the SaaS-pocalypse narrative to show up in the actual revenue growth rates of these SaaS companies. I pulled together the recent earnings from a variety of, public cloud software names, and they're all still growing, which you would expect. I mean, of course, like, you know, you wanna be growing fast and there's a lot of things and there's like the long-term durability of the value that, of the stock that actually informs enterprise value today. But we're certainly not seeing like, okay, there's so much churn at GitLab, for example, because people are forking it 'cause it's open source and vibe coding on top of it and they don't need to pay GitLab that you would expect GitLab's revenue to actually be shrinking. It's not, GitLab is growing at 23%, Adobe's at 12%, PagerDuty 2.7%, little low, but UiPath at 14%, Box is at 9%, Asana is at 9%. Asana feels like textbook, like you could just vibe code this. It's a Kanban board, a task list, but it's still growing. Zoom at 5.3%, Snowflake's growing 30%, Workday's at 14.5%, HubSpot's at 20%, Datadog's at 29%, Cloudflare, 34%. Monday.com, 25%. There is really, really strong revenue growth across the SaaS category. Of course, the expectations have been very high. So when there's a readjustment in expectations, you see a sell-off in the market. But this idea that the companies aren't growing anymore because they're being replaced so rapidly, that certainly hasn't taken hold just yet. So partner at Salesforce, investor Chicago Capital has been impressed with some of its recent moves. But what Salesforce really needs, he says, is positive word of mouth from clients talking up the value they derive from its AI products like Jason Lemkin. They need to show revolutionary jumps. Benioff has said Salesforce was destined to be an AI-first company as far back as 2014 when it launched its AI research unit. I didn't realize that an AI research unit all the way back then, but it makes sense just for like tagging and classifying different records inside the CRM. Still, it was caught flat-footed by the arrival of high-functioning chatbots in the form of ChatGPT. Customer superintelligence. Customer superintelligence. I mean, this is, it is funny because a lot of this, the agentic enterprise, it is a little like boring and it's not as like sexy as some of the more like crazy sci-fi scenarios. But in terms of just incrementally improving the value that is delivered to the customers, it seems like Things are doing okay. Early reviews were tepid, though. Customers complained of having to spend half their time preparing data so the AI could understand it, limiting the platform's effectiveness. To help fix the problems, Salesforce built a layer into its tech stack that automatically pulls in customer data from external sources and purchased a string of companies that include firms specializing in data management and AI-powered sales. At education company Pearson, agents now autonomously handle queries about order statuses, refunds, and lost access codes for its customers. This has increased the percentage of customer questions that don't involve human interaction by 40%. While AgentForce— where AgentForce has been lacking in its addressing complex customer problems and those require human touch, David Wemsley, Chief Digital Officer at Pandora Jewelry, said AgentForce hadn't been able to reliably recommend products on its own based on the vague context that customers share through its website, like, "My wife likes dogs. What should I buy her?" That's such a funny question, but I mean, I guess the natural language interface should be able to find you jewelry based on a love of dogs, although I think that a lot of people who are dog lovers might have different tastes that appeal to their jewelry.

Why not just get her a dog or multiple dogs? Yeah, I don't know. So how does Salesforce's view compare to the Verizon CEO?

So Verizon has a new CEO, His name is Dan Schulman. I don't think he's related to John Schulman, the co-founder of OpenAI and Thinking Machines, but he is stepping into the AI debate. He says for— The Wall Street Journal says for a big company CEO with big ambitions, Verizon Communications' Dan Schulman doesn't pull punches about the pain the technology could unleash on America's workforce. Just months into the job, he has predicted 20 to 30% unemployment within the next 2 to 5 years, which is staggering, staggering. And I can put that into some context. He says he warns that advancements in humanoid robots could upend the manual labor jobs still seen as safe today. And he has pushed for more education and reskilling to help workers adapt to intensifying tech disruption.

Okay, put this into context for me.

So he's— so this is an insane—

And yeah, and so this is like potentially the most aggressive stance. I mean, there are rumors, even, even, even Dario. So Dario had the clip going viral this weekend. It was a clip from last year talking about risks to entry-level white-collar job. Yeah. And Dario, which is like, you know, I would say like on the frontier around being concerned around job loss, was not calling for 30%. Well, no.

So, so the headline number, he says 50%, but —to entry-level white-collar work. —entry-level white-collar work. And that sounds really bad until you actually dig in and you realize that America only has 5 to 7 million entry-level white-collar workers. And the US labor force is 170 million people. And so if the Dario prediction came true, the overall employment— unemployment rate would sit somewhere between 6% and 9%, which is not great. And it's obviously deserving of intervention, but it's far from the 20 to 30% outlined by Dan Schulman. And it wouldn't even be— even if the Dario scenario, 50% of early stage white collar work unemployment, that happens, you're still below COVID, below the Great Depression. And I think that there's a whole bunch of government interventions that could offset that pretty quickly. Like 20 to 30% is truly like the do nothing. The government never engages. There's never any sort of, you know, incentive to keep hiring people. And there's this like fast takeoff in AGI and also humanoids, which I think people are, you know, worried about. But if you look at the deployment rate of self-driving cars, people have been saying that all the truck drivers were gonna go out of a job and, you know, this is a very, very slow takeoff. The number of self-driving cars on the road is well below 1% of overall cars on the road still. And this is just the case across the board. I was listening to a podcast with someone who was very worried about unemployment internationally, and he was saying that AI could upend the Philippines because the Philippines does a lot of customer service.

Which we haven't seen at all yet.

No, none of this is showing up in the unemployment data. It's not showing up in US unemployment data yet. I mean, That doesn't mean that you shouldn't be aware of this stuff, but it's certainly not sure. Just like the SaaSpocalypse is not showing up in revenue yet, you know, AI unemployment is not showing up in any of the employment statistics yet. So, but this individual was on a podcast that was saying that something along the lines of, you know, AI could be devastating to the Philippines economy because the Philippines is very dependent on customer service. And the stat he quoted was that 90% of the Philippines economy is based around customer service and customer support and handling people on the phone. And I was like, "90%? That seems really, really high." Like, they have to do other things in the Philippines because, sure, you go to work at your customer service job, but then you go home and you buy food and you live in an apartment. There must be real estate agents, there must be home builders, there must be people who work on roads, there must be doctors. There must be lawyers, like, an economy requires more than just like a single— like, it's— 90% feels like incredibly concentrated. So I looked it up and I was like, how big of an issue is— how big of an industry is customer service in the Philippines? Like, is it 90%? That feels high. Maybe it's 50, maybe it's 40. It was like 6%, 7%. It's really, really small. Now it's like, huge in terms of like, you know, much— You don't want it to go away. You don't want it to go away, of course. And you don't— and you do want, you know, you want a number of industries that are flourishing. And that is probably more concentrated than many other countries. There's probably many— it might be, it might be the country with the highest percentage of customer service intensivity in the economy, but it's not 90% of their economy. And so there's a little bit of this like, people don't seem to go back to the raw numbers because if you go back and you try and understand like, okay, what would 30% unemployment really look like? Well, that's like, you know, Great Depression level with no intervention, no, no action from the Fed, no action from the government. And it seems a little bit, a little bit crazy. So I can't tell if this is just like saying the biggest number, but there, there's a little bit of that going on here where If you want to grab headlines and you go out and you say, okay, well, someone predicted, you know, 5% unemployment, I'm going to predict 10% and then I'll jump to the forefront of the discussion. That seems to be a way to get earned media. But I mean, let's unpack a little bit more of his discussion because we will see if there's a way to steelman his take around 20 to 30% unemployment. Said, coached in the blunt AI talk is a warning for other CEOs. Be candid about the coming disruption or risk a public backlash. It's a very difficult time and everyone knows this, Schulman said in an interview with the Wall Street Journal. So I think being authentic—

I don't even agree with that stance. If you— everyone that's being candid or even talking about it is immediately getting backlash anyway.

Oh, yeah. So I think being authentic, being realistic, telling the truth as best you can is key. That belief, he said, is why Verizon created a $20 million career transition and retraining fund for the age of AI. He's like, well, again, laying off—

we are going to have greater unemployment than in the Great Depression. And I'm going to— you know, he's taking the piece of duct tape. Yeah. You know, there's water pouring out.

I know exactly what you're talking about.

And to be clear, to be clear, this is not our view. Like, this is not our view at all. No.

But I mean, at the same time, like, they did layoffs. They probably hired a lot during COVID and beyond. And they— and a $20 million career transition and retraining fund is great. Good start. That is good. Like, this is a good thing. But he says the warning— this is from the Journal— the warnings are a departure from the messaging of other public company CEOs, many of whom have been bullish about AI's potential to unlock new levels of growth but demure on or even reject the idea of job losses. A lot of people are saying AI is coming, we're going to run out of jobs. It's exactly the opposite, Nvidia CEO Jensen Huang said last month, pointing out that every other technological advancement has brought more productivity and more prosperity. And there is some new data showing that productivity might be climbing, which is very exciting if all the economists wind up developing a consensus around that. Amazon CEO Andy Jassy is similarly sanguine about potential job losses to AI. Though some roles will be replaced, there will be other jobs created. In the short term, though, a cavalcade of companies from Snap to Amazon have invoked AI or a desire to find a efficiencies as they slash large portions of their workforces. Block, which cut nearly half of its staff, predicted other companies would soon follow suit. A new Boston Consulting Group report predicts that AI will shape roughly half of US jobs in 2 to 3 years, and that up to 15% of jobs could eventually be eliminated outright. Again, that's, you know, reshaped and eliminated would be offset by the creation of new jobs. So even in this BCG report, you're probably looking at like, again, maybe a transition of like 6, 7, 8, 9, 10% unemployment while there's an adjustment period. Many Americans fear that will happen too. In a Quinnipiac University survey of 1,400 adults, 55% said they felt AI would bring more harm than good, up from 44% in a poll last year. And so the average American is dooming for sure. CEOs are not thinking about this the right way, said Bill George, the former CEO of Medtronic, who is now an executive fellow at Harvard Business School. Too many, he said, are focusing on productivity instead of laying out a strategy for how companies can find new business models to grow or on how workers can best use AI. They should be very candid with them and pay paint a big picture. Shulman's big picture also included sweeping job cuts. The 13,000 layoffs he announced shortly after his appointment as CEO in October were Verizon's largest ever, but not— but necessary to make Verizon more efficient, he said. All together, he's seeking to cut $9 billion of costs. Verizon said its layoffs were not related to AI. The carrier was too hierarchical, too bureaucratic, way too process-oriented. As opposed to outcomes-oriented.

And the CEO saying this is already happening in a big way, but the layoffs we're doing are just that we're bloated. Yeah.

So Verizon has 90,000 employees and they laid off 13,000. So maybe like a little over 10%. Riff. Verizon's a weird one because the stock is basically completely flat over the last— it's actually up 15% year to date and It is an incredibly stable stock, and it's also just like, should not be a victim of the SaaS-pocalypse because they own spectrum allocation.

Very hard to FIB, cell towers.

They own cell towers, exactly.

Or they have like very, very long-term lease agreements.

Yeah. And I mean, there is the case that like, you know, if Starlink direct-to-cell gets really good, even, I mean, even Elon's saying like, it's not gonna be that good inside buildings. Like it's pretty difficult to get to a point where all of a sudden there's a new technology that's just like, wildly disruptive. Like maybe if you're like somehow transitioning from Starlink when you're outside to Wi-Fi when you're inside, like you could cut the cord with Verizon, but that just feels so, so far away for something that is like a pretty key utility in most people's lives, like their water bill or electricity bill, like their phone internet bill is like pretty, it's like one of the, one of the probably the one of the least elastic things. Like, they will just keep paying and stick around. Now, they might move to Sprint or AT&T, and there's gonna be some competitive dynamic. It is an oligopoly after all. But it doesn't seem like there's going to be some massive disruption moment where people are vibe coding their own cell carriers necessarily. So, in meetings, he has repeatedly told Verizon staff they must embrace AI, describing it as core to the company's future. He used it himself to comb through some 8,000 responses after asking— This is the future.

Trust me, I used it once. Schulman's embrace of AI goes deeper than cost-cutting.

He envisions a company wholly reshaped by the technology, from improved customer service to more personalized options for consumers. And he has encouraged staffers to talk to their children about AI at the dinner table. In one all-hands, Schulman recommended that staff ask AI to write their obituaries to see how the technology works and how it frames their lives. Just dig your own grave. He's literally—

the title of this article, he's saying AI is coming for your job and everyone knows it. So they're ready. In tech, in tech, we're starting to, you know, every company is adopting this technology at a rapid rate. Yeah. But everyone's like, hey, this, you know, basically like jobs aren't tasks, right? This sort of narrative is building. Like, we need to figure out how to— it's— we've built a bunch of very useful tools, and they are going to have powerful effects in our economy. But we have to kind of change the narrative around this because like a fear-based approach is not working. And meanwhile, Bill Gates comes in, he's like— AI is coming for your job. Everyone knows it. Write your obituary.

Write your obituary.

So not, so not helpful. And, and, and they're not even doing AI-related job cuts. No, I love— so I just don't understand. I don't understand this whole press cycle.

I love, I love, I love that he just comes in and just immediately starts blackpilling. It's so wild to just get this job and immediately start blackpilling.

Oh, so do you think this is a setup for, for a much like deeper layoff than they've done historically? I don't know.

I mean, it's possible, but I think that they would need to do some serious like AI tooling and implementation and actually figure out like, I mean, I would be surprised about those 90,000. Like, I want to know more about the breakdown of those 90,000 employees or 80,000 employees. Like, what are they actually all doing? Which ones are actually just sitting there being like, I just do tasks all day long. Like form comes in, I type it into an Excel sheet and I email it to somebody. Like that type of job, yeah, that's probably gonna be automated in some way and that person will have to find a different way to make a play inside the organization. But for a lot of the folks at Verizon, I imagine that they're, working on bigger projects than just throwing tons and tons of people at a single problem. At the same time, who knows? Maybe it's 50, maybe half the company is customer support reps. I don't know. He has invited staffers to experiment with AI by writing poems to their loved ones. Some employees responded by using AI to write poems for Schulman, and they weren't half bad, he said. Like it or not, we live in the age of AI. I happen to like it. I agree with that. It's like we all wanted to live in the Renaissance or like when fire was first invented. How cool would that be? He continued, we're in that stage. We're not just appreciating it for what it could be. That's a very optimistic take. I like that sentence. That's good. Some prominent CEOs are starting to join Shulman in acknowledging AI's potential for disruption. Others have also recently sounded warnings. There's real risk of artificial intelligence could widen wealth and BlackRock CEO Larry Fink wrote in his annual letter to shareholders last month. Jamie Dimon recently told investors that AI's productivity gains could lead to other derivative effects. It may happen faster than we can adjust to it. Shulman said AI may reach human-level capability, known in the industry as AGI, by the end of next year, on the early side of most, most industry predictions. So end of 2027 for AGI. Isn't that crazy? I mean, Sequoia has an event right now, AI.SN, and there's the whole keynote slide is AGI is here. And so, people are going back and forth on that. But I think it's not that crazy to imagine very, very human-level AI by the end of 2027. That doesn't seem impossible whatsoever. The question is just like, how much will this be additive? How will the government respond? What will the actual effect on the labor market be? And I think people are digging into that a lot right now. There's a good podcast on this on Odd Lots with Alex Imas. He's a professor at University of Chicago focusing on economics and applied AI. Highly recommend you go check that out if you're interested in going deeper into the labor market. Well, moving on. Blue Origin rocket stumbles on first commercial mission. AST SpaceMobile, who we've talked about a few times here, Jeff Bezos' rocket company, said the satellite from ASTS was deployed into an incorrect orbit. And so a little bit of a setback for them. I think the stock traded down on the news a bit. The launch of the company's New Glenn rocket started smoothly with the vehicle shooting into the sky from the Blue Blue Origin launch site in Cape Canaveral, Florida. During the flight, New Glenn's third ever, the vehicle, the vehicle's huge booster returned safely to Earth, only a feat only Blue Origin and Elon Musk's SpaceX have ever achieved with orbital rockets. But the mission later suffered a mishap. A satellite the rocket was carrying into, into orbit for AST SpaceMobile, a company building cellular broadband network in space, wasn't deployed correctly. In a post on X, Blue Origin said its rocket delivered AST's satellite into an incorrect location in space. The payload was placed into an off-nominal orbit, adding that teams were assessing what happened. AST said the satellite's altitude was too low to sustain operations and that it will be taken out of orbit. The cost is expected to be covered under its insurance policy. Stumble comes as Blue Origin works to ramp up flights with New Glenn.

Yeah, I saw some of the, the AST retail army saying like, don't worry, keep holding, keep holding, it's covered under insurance. It seems like very, very, very, uh, very unfortunate, but it seems like it's somewhat rebounded a little bit. Somewhat to be expected, right, as Blue Origin like figures out their commercial business.

Yeah, it traded down like 16% overnight, but it's down, uh, just 6% today. So a little bit of a rebound.

And we got a fantastic video of the booster landing that we can pull up here. Yeah. Uh, courtesy of Jeff.

Tyler, you dug into why doesn't ASTS just, uh, launch on SpaceX?

Well, so there are two things. I think one thing I thought was interesting is like there's this whole press cycle about it, like, oh, it was like a failure of a launch or something, but this has like happened a number of times before. Like SpaceX in, uh, in 2024 There was a Falcon 9 that kind of, the upper stage, like failed, and then a few like Starlink satellites.

Just went into the wrong orbit.

Yeah, basically. And then they're too low and then they just get burned up. Yeah, companies don't typically like to talk about when they send something into space and lose it, basically. But that happens too. Yeah. I remember at the beginning of last year, there was a launch for a venture-backed space company and, you know, they basically put a satellite up and almost immediately lost contact with it. So it's not unusual.

Unfortunate, but they'll be back. In 2016, SpaceX blew up a Facebook rocket, which was crazy. Mark Zuckerberg laments the loss of Internet.org satellite. The Facebook CEO said he was deeply disappointed in the explosion of Falcon 9 rocket carrying satellite intended to provide internet coverage to parts of Africa. So writing on his Facebook page, Zuckerberg said, "As I'm here in Africa, I'm deeply disappointed to hear that SpaceX's launch failure destroyed our satellite that would have provided connectivity to so many entrepreneurs and everyone else across the continent. The accidental explosion of the Falcon 9 rocket early Thursday morning," this is back in 2016, "referred to as an anomaly by SpaceX engineer, destroyed both the rocket and its cargo, the Amos-6 satellite, which Facebook had planned to deploy to provide internet coverage to parts of Africa. Fortunately, we have developed other technologies like Aquila that will connect people as well. We remain committed to our mission of connecting everyone, and we will keep working until everyone has the opportunities this satellite would have provided. Uh, contrary to Zuckerberg's description, the satellite did not belonged to Facebook. In October 2015, Facebook partnered with Eutelsat— I can't pronounce that— a French satellite company to lease the broadband capability of the Amos-6, which was built by Israeli company Spacecom. According to SpaceNews, which reviewed Spacecom filings with Tel Aviv Stock Exchange, the joint lease cost $95 million over 5 years. So this like $100 million satellite just blew up on the pad. So lots of different setbacks, but there is one technology that is not having setbacks, which is robotic marathoners. And there's new— This is crazy. There's a crazy video. I don't know if the video relates to this, but a Chinese robot beat a human best time in a half marathon after a stumble. Tech companies are making progress to of fixing humanoid runners' malfunctions. A year ago in Beijing, humanoid, a humanoid robot half marathon race, the first runner to cross the finish line took more than 2.5 hours. In this year's event, the champion beat the fastest human ever. Sunday's race demonstrated China's rapid progress in humanoid robotics, a field American tech leaders, including NVIDIA's Jensen Huang and Tesla's Elon Musk, say is the next big thing. Beijing. This video is pretty wild.

Is this the actual car? This is one that's failing. Okay. Didn't make the half marathon, but you can see there's dry ice spilling out of the back. This is to help with cooling. Oh, they're cooling it off? Interesting. Which is like very cyberpunk.

Whoa, that's crazy that you have to load it up with dry ice as well. I like that. About 220 yards from the finish line, the 5-foot-5 Lightning slammed into a barricade and collapsed. Collapsed. The red and black robot managed to get back on its feet with help from humans and ran across the finish line in 50 minutes and 26 seconds according to state media. I feel like if you fall down and you're a human in a half marathon and humans help you get up, that's still fair game. You're still good as long as you finish. This probably still counts. Last month, the humanoid— the human, the actual human world record holder from Uganda finished in 57 minutes and 26 seconds in Lisbon, Portugal. Lightning and two siblings from Honor swept the podium. All three navigated the course without human control, excluding the one-time help the champion got. The race penalized the completion times of those relying on constant human remote control, including one that finished the race in under 50 minutes, but it was teleoperated. The Tiankung Ultra, which was developed by Beijing-based Lab X humanoid and won last year's race more than halved its finish time this year, clocking in at 1 hour and 15 minutes without any human intervention. That is pretty impressive. Like, they cut the time in half in just one year. Wow. The 13-mile course included more complex terrain than last year, such as slopes, narrow passages, and sharp turns, testing robots' abilities.

If you assume an acceleration in progress, eventually one of these humanoids is like running a marathon and like an hour or like 30 minutes, just like truly insane, insane speeds. Yeah. I mean, as fast as a cheetah.

Tyler could still take it out, though. I think so. China is moving to quickly dominate the humanoid robotics industry and cement its place in the global supply chain. While the US controls the best chips and other technology for robot brains, China leads in the manufacturing ecosystem for humanoid robot bodies. That has been reported many, many times. Well, without further ado, we have Signal. Here in the TVPM UltraDome, let's bring in Signal, who is launching Sky, an agentic AI home screen replacing the iPhone app.

There he is with content. Handsome fella. Driven. Dude, I can't believe it's been too long.

This long. We've reacted to your posts like so many times. So good to have you.

I— oh my God, it's incredible to be on here, guys. Welcome. It's been a long time coming, but I just want to start off with saying congratulations. Thank you. This has been ridiculous. Yeah, I've been watching you guys since day zero.

Really? You watched the first episode?

No, because I think we were covering— I think we did. Your post would have been making it into like the first episode for sure. You're probably like, why are these two guys in suits?

Yeah, reading, printing out my tweets. This is pretty wild.

It is, uh, I tweeted this out a little bit, but I was like, man, when I read that and you guys were reacting to what I posted, because I didn't really think about what I posted and you guys analyzed it, and I was like, oh Oh my God, this is ridiculous. Holy crap. What did I even write? I don't remember.

You have a ton of bangers. They're all good posts. We had a great time. That was the lifeblood of the show. It was so much fun. Anyway, we're not here to talk about the first episode. We're here to talk about your first episode in this new journey. Talk to us about what you're launching. Very smooth, Jon. I try. I try. You guys have come a long way, for sure. Anyway, introduce yourself a little bit, introduce the app, introduce the product, where you want this to go, and I have a ton of questions.

No, absolutely. You know, I've always been in consumer software, and I think there's just not that much other than the sort of main players. I noticed, and there's a lot to be done, and what a time to be alive. We're experimenting at the very basic layer of how to make this stuff really easy to use, really easy to access for normal people that have not really come into this agentic AI world in full speed other than chatting with chatbots and whatnot. And I think it's an early experimentation of what we're up to is how AI will speak to you as well as how sort of ambient AI will be in various surfaces, starting with your phone, right? Maybe your phone may not look exactly the same way even in the next couple of years. So we're kind of operating at the very earliest stages of experimenting of how AI will communicate with you. And we're trying to think of creative surfaces. And one of the most interesting places, you always— people take out their phones and they glance at it. And it turns out, This stuff hasn't really changed in such a long time. Yeah. The iPhone home screen is 20 years old. That's 2 decades. And it's just static icons. Yeah. Roughly speaking, it has not really evolved in any meaningful way.

They one-shot it. They one-shot it.

Yeah, that's the steel man. The steel man is like, it's good. It's the final form. It's the final form.

Some things don't change.

We're all going to— We could always do a 3-wheel car, 5-wheel car, 6-wheel car.

I mean, the thing that I've been pressing on is like, you would think, you know, if I could rewind 2, 3 years, I would not expect, and knowing how much progress there would be in AI, I would not expect to look at the top 25 apps in the App Store and only see LLMs. Like chat apps. I would expect to see like a variety, given just like how many magical experiences people have had, a variety of like new products coming out.

Like a new Uber, new Instagram. And that's like, There was a previous app boom that was like very diffuse. You got, you got Candy Crush and what's the one with the pigs, the Flappy Bird and Angry Birds. And like you got all these different apps, Runkeeper and diet products. And it feels like this has really collapsed down into just chat apps.

Yeah. And, you know, I actually recently got a new phone and I was installing apps. You know, I always like to set up my phone bare. Like, I don't like to transition my phone.

Yeah, interesting. Maybe Really?

It gives me a little bit of a reset on not only what I need and what I don't need, but also how to think about software as it exists today. Like, I don't want to be tied to what I was before. I want to kind of be anew, if you will. And, you know, it turns out I installed, you know, GPT and Claude and whatnot. And I was like, man, I don't think I really need that much stuff. You know, these LLMs are kind of collapsing how and what you do into an interesting dynamic. And I think generally they're incredibly powerful. And the fact that those top 5 apps are all LLMs roughly speaks volumes to the zeitgeist and speaks volumes to the impact of the actual technology. I don't think— previous, I think technology has always been a little bit more evolutionary than not. Obviously, hindsight is 20/20. But But this feels so, so different. You know, as a technologist, this, this world feels very different. It feels just, just the things are going to rapidly change from here on out in terms of how people experience their lives, how people interact with each other, and how, you know, we're going to facilitate brand new interactions potentially or, you know, completely reinvent old ones. So I'm very excited for that. And I think our company and the way that we think about it from a consumer perspective is just to make things, you know, easily accessible for these individuals. And I think we're gonna attempt to do that. Very basic stuff, you know?

Yeah, what is your, like, more as like a CEO, it sounds like you guys have raised a little bit of money and like are just like in an experimental phase. Is that like generally the right read?

I think, yeah, I think generally, I would say two things. Number one is, look, we built a really fun product that we're going to give to lots and lots and lots of people. Turns out, you know, this era is non-zero marginal cost. We have raised a little bit of capital. But, you know, inference is non-trivially expensive. And especially if you operate at like agentic inference or just background inference, that stuff is just consistently going. Unlike, you know, Claude or GPT where people actually make requests or go on there and type something, we're doing it on behalf of you, right? Like we're doing those things in the background where we anticipate, we listen to contacts, we turn, you know, every time, for example, every time you get an email, we process it with one of our agents. It sort of turns out, it buckets that item, it tries to figure out what to do with it, it tries to see if there's, it deserves some higher order, like ranking, and then it tries to figure out, oh, can I complete this task? Can I draft a reply? Oh, maybe it does deserve a reply. Let me go back to John and say, okay, hey, here's a reply that I've crafted based on everything I know. And it's a very— look, replying to emails has been, you know, 30, 40 years, but this is a new world in terms of how you think about communication and how agents kind of mediate this world. And apply that to any, you know, any aspect of your life, whether it's health or finance or where— even where you are, right? Like, one of the underlying things that we do is location. And, you know, imagine learning about the world around you through an LLM, whether it's through text or voice, by simply it being on your home screen wherever you are. If you're at a museum, if you're in a new neighborhood, and that's one tap away, like our goal is to make intelligence either zero or one tap away, not one prompt away.

So for learning question, how do you, how do you, how do you make this product free so that everyone can use it? Will we see ads on the home screen of the iPhone? Ad-supported? I think ad-supported could make sense, right? I'm walking by a coffee shop and I get, I get an offer.

There's ads in Apple Maps now.

So you're seeing my notification stream. And so you can really, you know what I like. Yeah, I am. I think the interesting thing here is like, like you guys can operate, like you can kind of wake up and ask yourself, like, what would Apple do? If they were like truly excited about AI versus like seemingly scared of it?

You know, I've posted about this a little bit with respect to Apple. Apple relies on a deterministic world, right? Like a world where lots and lots of elements from design to the experience to the underlying context, that doesn't change as much. And it's very deterministic. So, for example, iPhone is very much kind of software is very broadcasty, right? It's a one-to-many. They write it once and it runs for everybody and roughly speaking it runs exactly the same way. Now with a non-deterministic world, that could change drastically. Like for example, for when we, you know, give this out to everybody, everybody in there, everybody has a very different experience with our app because it's completely non-deterministic. And Apple lives in a deterministic world and having to transition into non-determinism and non-deterministic software, it's actually a really non-trivial transition, especially for a company like Apple where every single corner or every single thing needs to be tightly controlled. So I think for us, we're kind of paving a path where we can marry some level of expectations and determinism with the beauty of non-deterministic elements of AI and create a really great user experience around that that lives on your home screen and works for everybody. That's how we think about it. So we're kind of maybe in some sense moving ahead of Apple, who has to deal with a few billion users, you know, just, just a minor amount of users. So I think, I think that, you know, this, this world deserves more experimentation like, like what we're doing in terms of both with user interfaces and experiences and feeds and ranking. And Jordi, going back to your point on monetization, like, I think that's the number one thing that I think about quite drastically. Look, we're trying to capture real estate on your device that is the home screen that is available at a glance. And, you know, I think I've posted about—

It's actually the greatest billboard of all time. Exactly.

I mean, it is insanely powerful if we can activate that. Now, that's a tall order. But at the same time, you know, like I've posted about ads and advertising. Look, I've been in technology for such a long time from consumer, you know, whether it's like feeds at Facebook or, you know, even Google and whatnot. But, you know, you start to think about advertising has done a lot of good for the world. You know, it has actually made things accessible. And I think generally if advertising is done well, it is actually one of the most useful things useful economic paradigms in terms of delivering equality in services to people who would otherwise not have been afforded. So I believe in advertising. I believe in good advertising. I believe in advertising that is actually like complementary to the user experience, not necessarily taking away from it. It's tricky to do.

But, you know, the worst ad experience I've ever had as an ad enjoyer was like in college. I was buying Kindle, and they were like, do you want the ad-supported Kindle for like $100 or the ad-free Kindle for like $120? And as a college student, I was like, I don't think I'll mind ads. Why not? And it's on the home screen when the device is locked, and they're just showing me books that I would never read. So it's terrible targeting, and it's a device that's just like sitting around my house all all the time, and it kind of looks like, wait, you're reading that? Yeah, yeah, yeah. So if you can avoid that.

Yeah, definitely. There's a question from the chat. Why are you anonymous? Do you plan on continuing to be anonymous? I mean, I imagine as you grow the business, like, it would be interesting to stay anonymous forever, but, you know, at some point a journalist will want to know, and I imagine that unless your OPSEC is super tight, it'll come out. Not that it's like bad, I'm just wondering like, how have you processed the anon thing?

That's a great question. You know, my entire online existence as it exists today with respect to this account is a giant accident because I was trying to find my old username and password for, 'cause I just wanted to post a little while ago, maybe, you know, whenever I started posting. But I think when you guys actually started started, uh, TBPN as well.

And I think, yeah, your account must have gone from like 1,000 followers to like 80K, over 100 within the first few months of us.

It was kind of nuts. Okay, so the whole story is, you know, I, um, I was kind of, uh, I was like, you know what, I just want to read. I maybe I'll reply or something. And I had a lot of fun. And, you know, whenever I'm having fun, I like to double down on things, you know, maybe not change, change it around too much. Look, I think certainly this, this world is going to change at some point. Um, but in the meantime, you know, I love leaning into fun new things. You know, if you want to do anything new in the world, you got to do it a little bit differently. You got to be a little bit more unique. You got to be a little bit more mysterious, a little bit more— um, yeah, I don't know. All of this stuff is just so wild to me, right? Like the anonymity aspect of it and the X culture and the anons around it. And you guys have had, you know—

yeah, you can never Dox. I'm sorry, you can never dox. I mean, like Banksy, Banksy getting unmasked, is that good for the, is that good for the brand? Like, like, you could be, you could be like taller than John, gigachad, everyone just imagines, and it won't be enough, right? Like, it was, it, we all picture you as a philosopher king.

Yeah, swords. Oh, it's beautiful, isn't it? I mean, my, my bio and my philo, like, the actual content are like sometimes, sometimes the same and sometimes disjoint, but That's, I think, the beauty of it. So in that realm, you know, like I was actually talking to some people around this and I was like, I sometimes, I basically kind of do what I feel like and right now it's like just been so much fun to kind of lean in on this world and, you know, but I think certainly I'm more like I'll marry myself to the zeitgeist, if you will, and if that calls that, then I'll continue and, We'll see what happens.

Not married to the game, married to the side guys.

Married to the side guys. Okay, take me through iOS development, like current status, because this seems like a really good idea that people would want. I do think people are bored of the grid and having something that's more dynamic and agentic makes a ton of sense. And I think you could deliver a ton of value. But my fear is that Apple's just like, no, that's our real estate. So is there a clear path right now through like widgets and the Shortcuts API to actually do interesting things? Or am I going to have to like sideload a different OS? Like how, how like consumer-friendly and like easy will this be? Or will you be bumping up against the walled garden of Apple pretty quickly?

That's a really great question. Look, every single thing we do is within the confines of the Apple ecosystem and experience in the way that they've designed and made it work. And we've designed our product to fit almost like a glove into that ecosystem. You know, when you onboard into our experience, you just connect the things in your life that you think you would want more.

Yeah, the server-side connections, I totally get. Like, all of that makes sense. You can like OAuth with the email and use APIs or MCPs. Like, there's a million things you can do on the server. What I'm interested in is like, is like, how big can the widget be, or can you actually take over the whole home screen?

We can, we can take care of the— that we basically ask you to install two widgets, a medium widget and a large widget that encapsulates the entire home screen, and those work in dynamic together. So the, the medium widget will basically what we call a wildcard widget internally, which shows you precisely what you might need to know at this point in time. And then the rest, the big widget is what we call a For You widget, which is just a feed. And, uh, you know what we're doing, John, I think, is we're building kind of the new iteration, the, the Facebook News Feed 2.0, that's entirely AI-generated about your life, highly personal, and that lives directly on your home screen. And you can browse it as easily— the feed paradigm is so familiar with individuals. And the beautiful The beautiful part is that it's all like AI-generated and AI-mediated. Every— we have 22 agents that work continuously to generate content for that feed.

It's funny because like everything, everything that's fancy, I'm like, yeah, that's easy. And then like getting people to install 2 widgets, I'm like, that's the hard part. And I don't know if I'm right, but like it feels like, like, like that, like I'm still in like prosumer territory, but that's a good place to start. And then you can open it up. Hopefully make it easier. And then maybe Apple opens it up to a point where like you downloaded an app and just by clicking yes, it just installs the widgets by default or something. It does feel like— the biggest thing is like, I just love when like these new surface areas are explored. I mean, you mentioned Nikita, but like he's done a great job of really understanding all the different hooks, what you can do within iOS, you know, pushing those to the limit, creating like new UX, new experiences on top of like what Apple gives you. And it feels like that's really underexplored.

So have you thought of, have you thought about trying to build any products within any of the LLM ecosystems? Just given that they already have an exist, you know, massive existing user base? Yes.

That's a great question. And you know, I've definitely explored like every single thing with respect to an API or a plugin platform that comes out, you know, I, you know, for example, the WWDC APIs that come out every year, I, I read them like, like the Bible, you know, like I would, I, I, I, I, I go to the church of Apple. You know, that's my, that's my, you know, religious holiday or whatnot.

And every time— Tim Cook is the Pope. He's your Pope. Yeah, exactly. We get it. Yeah.

And, you know, so I scrounge these, I think. Generally these API, the sort of apps or ChatGPT apps are— it's unclear to me what the incentive structures are for the app developer just yet. And it's unclear to me why applications deserve to exist inside of an LLM just yet besides just adding more context, because theoretically an LLM is powerful enough to even generate an app to be able to do something, in which case I'm not sure exactly unless you bring some highly proprietary data like Zillow or whatnot, maybe some of these experiences work. But otherwise, I don't know if the small-time developer— you guys remember the flashlight apps and the lighter apps on the iPhone? Like, you know, you're not really seeing those types of experiences.

All I remember as a probably 12-year-old was the beer app.

The beer app was good.

Just thinking that, thinking that that was like, that was peak humor.

What about the I'm Rich app that was like $10,000? It was just, it showed a picture of a diamond. And it was just like you could just open a picture on your phone, but if you bought that app, you could show people that you just wasted $10K on an iPhone app or whatever. I think it literally maxed out. It was the most expensive app you could possibly, like, type into the App Store.

I think the guy sold, like, over 500 copies before Apple removed it.

Yeah.

For no— not offering real person. So I think he made $500 times $1,000, and it was incredible. See, these are the kinds of these experiments and creative elements that, you know, I think deserve to exist in the world today. Yeah. And I think, you know, we're a few hand— a handful of individuals that are kind of trying to make that happen. And we're trying to be really creative, really fun, really interesting, engaging. And AI is a super powerful tool to be able to do that. More powerful than the original iPhone APIs. Like, holy crap. So I think we're early, But we're going to try to—

It seems like a dream. Like, there's so much—

You're doing this the right way, because there's another scenario where you could raise a Series A right now, or like, you know— You have a team.

You have the technology. The team's small. And the waitlist. You'll be able to test things, and experiment, and learn. There's so much opportunity. Well, what an exciting time. Thank you so much for joining the show. This is great.

Hey, thanks a lot, guys.

Yeah, great to finally have you on, and congrats again on the fund. Yeah, 1,000%.

And maybe next time we'll talk about hot takes for sure.

Oh yeah. What do you think about this AI-powered cannabis vape with blockchain rewards? I'll have to, I'll have to try it out before. So John was, John showed me like a screenshot. He was showing you the website for this, this morning. And I didn't realize today is 4/20. So that's like, this is like a 4/20 joke.

Is it a joke though? I think it's a real website. Like I think it's, yeah.

Yeah. John, you can make a joke website. Did you know that?

I know, I know, but I think they're—

John Claude Design, man. Claude Design. You can do anything.

Okay, so you're telling me if I put this in the Wayback Machine and it shows up as of yesterday, it's not a joke? 'Cause I bet you this existed yesterday. Let's see. Let's see.

I bet you if they were planning to make a—

April 6th. It's in the Wayback Machine. What does it say? It says, let's see. It's loading. It's got blockchain still. I think these are hustlers who have been just tacking on every single possible trend to go as viral as possible. I don't know. We'll have to dig into it. Well, thank you so much for joining the show, Signal. Yep. Fantastic news. Great to hang. And congratulations on the progress.

Thanks, thanks. Have fun.

And we'll talk to you soon. Cheers. Have a good one. Up next, we have Ethan Ding from TextQL. He is the co-founder and CEO.

Here to announce a fundraise, to talk about enterprise analytics, enterprise workflows.

I think we had him. Let's check in with the team and make sure that he is here. I believe he is. So let's bring Ethan Ding in from TexQL into the TVP and UltraDome. Ethan, hello! How are you doing?

What's going on, guys?

Oh, how you doing? Uh, how's it going? It's good.

We have two of you, so please introduce both And it looks like it's being filmed on a phone camera from 2005, but we're excited to have you on.

Hi, I'm Ethan. I'm the CEO and co-founder. This is my co-founder, Mark, our CTO. We're actually at a customer office on-site, and this is actually a company that does not— I think like IP bands like Zoom, from their, their internal network, which I found out like really last second. Interesting. Phone, phone. Yeah.

Okay. No, no, it's great. It's, it is, I'm, I'm super, we're super excited to have you guys on. Walk, walk us through history of the company since it's, it's your first time.

Yeah, I think we started this at like late 2022, right before ChatGPT came out. We had this, this idea that we, if we, if we spent a lot of money and time on like try to make analytics work, it'd be worth something. We really didn't know what we were doing when we first got started. Since then, what we've really realized is after 2 years of rebuilding the product 10 times over, we landed on something where I think the typical enterprise has 150 databases, 10 different dashboards, BI tools. They have 20,000 different charts, another 400,000 tables, and every single vendor wants you to migrate to into their thing and like drop another $20 million on like systems integrator for another 10 years to like do the migrations. Or you're like, let's like build a thing that can like connect to everything and hopefully be the— do for analytics what like high-frequency trading firms do for like stock markets.

What was the first like analytics stack? Because if you're pre-GPTs, are you just doing like word clouds or clustering based on keywords and tagging different phrases as they flow through a system? What was the initial, like, okay, there's a stream of data, there's a bunch of text in a variety of databases, like how are you giving the customers value from that?

Well, it was pre-ChatGPT, but still post-GPT. Okay. And so we had GPT-3, but I think as we all know at this point, The IQ of that was way overstated, at least for business stuff. And so, yeah, text box on the right, where you type in your question, like how do I get revenue? Then a text box on the left with the SQL.

And then you run it, and you copy it, and then paste it into your whatever tool.

And then that was it. We've come a long way from that, fortunately.

So walk us through— I mean, you don't have to tell us what customer you're with today, but what does actually, as a group of co-founders going to a customer site, like what are you doing? How deep in the weeds are you? Is it just sort of like a high-level pitch or are you rolling up your sleeves and working on actual integration?

Yeah, well, I think like today's session was basically like walk us through like the, I mean, every single Fortune 500 company like basically spends like 9 figures on AWS, GCP, and Azure. They spend another 10 figures on labor that manages all these systems. And it's kind of this constant war of you go from spending $20 million to a Teradata on-prem to spending $30 million with a Databricks in the cloud or something. And so what we're trying to map out with them is Anthropic is telegraphing that they want to be the size of AWS in 3 years, right? And they're an enterprise-oriented company. They expect that basically come out of companies like the that we're working with, it's like budgets. And so they basically expect them to take their entire IT budget of, let's say, $200, $300 million, double it, and materialize this money out of thin air and spend it on inference. They're walking us through the parts of their roadmap where what kind of SaaS can they sunset? What kind of labor costs are they thinking about the trade-offs around on what time horizons to free it up? And also what kind of workloads that are going to be extremely expensive that they can kind of start basically triaging off the large models. Because at the end of the day, these are kind of the people, like when you see the token charts, everyone token maxing, these are kind of the people who pay the bills on those tokens and they notice the size of that bill blowing up like 10x year over year. Yeah, interesting. It's interesting side of the equation, I guess, like people don't spend a ton of time talking about. Yeah.

How have you been processing the SaaSpocalypse narrative? Benioff was given some pushback. Other folks were very bullish on it. There's a whole bunch of different data points. I've been just shocked that we haven't seen revenue declines from really any software company that's been targeted. Although, you know, you could talk about the long term, but slowing growth, the growth, like these companies are still growing even if they are in like the direct path of the AI companies.

Yeah, for the longest time, I assumed that everyone likes to buy good enough. I think really in the past 3 months, I've heard 6 different CIOs or CEOs of Fortune 100s talk about how they're ready to— Salesforce has increased their headless tax and they're ready to do a 2-year migration off into Postgres or something. It's not like they're going to another vendor. They're just like, I need to free up money to set on GPU. Set GPTs on fire. Sure. And I'll pay like anybody to like, like move me into like, like a Postgres instance and like, like Google Cloud or like AWS.

Interesting. Okay, so that's a little bit of your opportunity. You help with migration?

That one's kind of adjacent to us. We look at a lot more like low-level infrastructure, like Databricks, Snowflake, like Tableau and otherwise. But it's interesting, like a CRM is basically the second most important system of record at an enterprise next to the ERP. They're willing to entertain moving entirely off CRMs within a 2-year time horizon. Maybe they'll completely fail, but the willingness to forge into that is then an order of magnitude more higher than I expected. And surprising based on the profile of company. That's something you'd normally expect from Google or Facebook, the most mature engineering companies in the world. But you see this effort starting to come from 100-year-old companies as well.

Interesting. And is that driven more by they think that over the long term, there will be a net cost savings to having their own system or more that they want something that's completely bespoke and more custom to their business?

It's mostly not. There's an interesting thing Larry Ellison talks about when it comes to enterprise sales, which is that enterprise sales is kind of like buying like Gucci bags. It's much more like any given year you have a CIO come in, right? The half-life of a CIO is like 5 years before they retire. New one comes in, they have to start a set of new initiatives. Often they're finger testing the market and trying to figure out who has the best vibes. Sure. Basically, whoever has the best vibes, I'm going to throw 8 figures against, and I'm going to move that 8 figures away from someone like a company that has worse vibes. Sure, there's this like weird, um, self-fulfilling prophecy where they're vibe procuring. Yeah, yeah, like you walk into a room, you can tell when that like company stock is up. You can tell when the CEO is like extremely cocky. You can tell like when the, when the FDEs and the salespeople are like, listen, like you don't even have to move this today, I'm gonna get your business next year no matter what because I know we're on the up and up. And you can also like tell when like a vendor is like desperate right? And they're like, like, moving for discounts, right? When, when some— when, when— as long as the perception that a company, like, the market is short a given, like, SaaS vendor, um, customers start asking for discounts. The most aggressive ones go first, but now, like, all the sellers on that team are, like, like, bad. Like, you know, they're traumatized between, like, the next set of renewals, so they're, they're progressively going to give more and more ground. Um, it's kind of a rough, uh, yeah, wouldn't want to be one of those right now.

Interesting, interesting. Well, your business is growing, you're raising money. Tell us about the latest round.

Yeah, it was actually like a two-part round. The first led by like Hoff Capital and the latest part like Blackstone. How much did you raise? I think it's $17 million total.

Well, this was awesome, guys. I really, I really, Enjoyed speaking with you both and thank you for making time on while you're hanging out with your customers. And we appreciate the perspective. Yeah, let's, let's do it again soon.

And good luck out there. We'll talk to you soon. Thank you.

Have a good one.

Have a good rest of your day. Up next, we have Matt McKinney from Loop. He's the co-founder and CEO raising a big Series C in the waiting room. Let's bring him into the TDP and Ultradome. Matt, how are you doing? What's going on, guys? Not too bad. Good to have you here. Uh, first time on the show, why don't you introduce yourself and the company?

I'm Matt McKinney. I'm the co-founder and CEO of Loop, and our mission is to unlock value that's trapped in the operations that power the physical economy. And we started specifically in the back office where no one else wanted to go, and back office services and automating things like accounting and general ledger coding and in payment services. All we do to exist is to make our customers more efficient so that they can better serve their customers. We work with some of the most important companies in the world, 20% of the Fortune 100. Wow. Excited to be on with you guys.

Yeah. Well, how— I mean, there's a different world when you say procurement or accounting that you could have been like, "We're an agentic accounting firm," or something. Feels like this is much more cross-functional. How are you actually positioning like the integration? Who's the buyer? How does the business like instantiate itself inside of your customer?

Yeah, we started with a very acute problem, which is if you look at the supply chain industry, roughly 30% of invoices are wrong or they have an error. So clearing of them is really painful. —people. Yep. And because of that, a bunch of services firms popped up, specialized services firms popped up. Sure. And all they do is address, uh, address that specific problem, which is you've got to adjudicate this invoice, you got to ensure that it's accurate, you got to remit payment to the truck drivers and the carriers across the world. And that's a big industry. It happens to be a $5 billion industry alone, and it's all done with human labor. And obviously LLMs presented a perfect opportunity for that. But I think what got my co-founder and I most excited is really the data. If you can go organize the data, then you can obviously automate things like we just mentioned, the accounting, but you can use that as a wedge to expand into adjacent use cases, whether it be in compliance or planning or procurement, and continue to unlock value for your customers.

Yeah. So what does it look like for a customer to actually onboard? I imagine you need access to their emails, like the actual PDFs of the invoices, whether things have gotten paid. There's often like multiple versions of a particular invoice and then you need to plug into bank data to see what's actually moved around or accounting systems. Like how long does it take to integrate and how key is that? It's wild.

I mean, what's so wild about supply chain, supply chain is just a network of networks. Yeah. And you've got, you know, a transportation carrier, you've got a supplier, you could have multiple versions within your own enterprise. Some— a lot of these big companies, they do a bunch of M&A, and so you've got all these different versions of the truth. So just take, for example, the weight and dimensions of a package or shipment. There could be 7 different versions of the weight and dimensions for that package. And so you need almost like a, a data auditor itself to prove that that's the correct data. And we get it from a bunch of different sources. We get it from you know, email, we get it from EDI, we get it from API connection, we get it from Excel spreadsheet, PDF, you name it. It's just really messy, unstructured data, data that no one's ever organized. Quite frankly, no one was able to organize until the power of LLMs came out.

Does, when you talk to customers, do they care about the AI buzzword at all?

Yeah, are they trying to buy AI?

Or do they just want a solution? Like, how are you thinking about how, how front and center AI is in your value prop and pitch?

All our customers want is just outcomes. Yeah. And they don't really care how you deliver it. Now, sometimes they might have a top-down AI mandate. Sure. That they're looking to check the box, but at the end of the day, they don't care how the sausage is made. They're buying an outcome. That's all that they want. And if you can demonstrate that you can deliver a superior outcome It's better, it's faster, for example, you're finding more errors or whatnot, and you have faster resolution times so you can close your books faster. That's what they're buying. They're not buying agentic buzzword fit into their checkbook. They really want to buy the value.

Yeah. What is the value of getting this business to scale? Is there something where you can optimize the supply chain, introduce— different clients that need different resources and help them buy the best product at the right time, or reviews, or even just like panel data of like how the economy is moving. We've seen that from some financial, some fintech companies. Like, how are you thinking about the value that's unlocked as you become a platform? Yeah, there's more.

You mentioned like procurement, for example, there's more obviously when you have a bunch of data that you can use and say, this is good, this is bad, you should be working with the supplier and you're not, because we see it over here. But I think that, you know, really building context across the network is probably where we see the most gains, where you can take learnings from a carrier that's working with, you know, suppliers working with multiple parties in the network, and then say, oh well, this carrier always likes to build a specific way, and so propagate that learning through the LLM's context across all the, all the companies that work with that carrier. And I think that's really I'd say the unlock in automation that you get at scale. And then obviously, you mentioned the intelligence one, the time to value. You're able to get someone on much faster because you're working with 95% of their suppliers instead of 2% of their suppliers. That's a huge win as well.

What were you doing before this? Did you always have a love for back office supply chain optimization?

I can tell you, I could tell you, since he was a boy, I came out of the womb and I just said, I want to automate the back office.

Yeah, complex logistics.

Yeah, I've always just been obsessed with systems and making them more efficient. And I think, you know, when you look at the physical world and its supply chain, you know, the trucks and the train, and that stuff's just so exciting. You feel like it's the last frontier where you can truly unlock that. I mean, look, supply chain alone just in the US is $11 trillion in spend. Wow. You know, one of the largest—

trillion— That's huge. Yeah.

Yeah, if you want to have an impact on GDP— Oh, too high.

Sorry, I was giving you that for the $11 trillion. You said the biggest number.

So that was huge. If you want to have an impact though on GDP, at the end of the day, GDP growth is just productivity growth. And technology is the leading advantage in that. You're saying that $11 trillion in spend goes to supply chain and no one's really touching that. We got to go attack that problem.

Yeah, I have so many more questions, but—

But before, yeah, before that, my co-founder and I were at Uber. Okay, yeah. Early on the freight team. There you go. And we saw a lot of the, you know, the real world there where you had nasty PDFs, you had bill of ladings, proof of deliveries, just a total data mess, to be honest with you. And we knew that there had to be a better way.

Yeah. Dude, perfect VC pattern match. Uber freight team.

Yep. That's a great team, great company. We know so many entrepreneurs that came out of Uber. Tell us about the round. What's going on? How much did you raise? We raised $95 million.

Couldn't pull together that last $5 million.

That last $5 million, what was going on, buddy? It evaded you, it evaded you.

Sandbagging, sandbagging.

No, you had to set up, Sada had to set up the next round.

Yeah, you gotta leave a little bit for the next one. Yeah, exactly. We're super excited.

Really, really stacked line. Where are you guys based?

Valor, Atreides, 8VC, Founders Fund, Index, JP Morgan. Wow, you got everybody. You got everybody. Congratulations. Wait, where are you guys based?

Sorry, I missed it.

We're based in San Francisco. We got offices in Chicago as well as in New York.

Is that because Chicago's a major logistics hub with the trains and trucks and stuff?

Yeah, there's a lot of reasons.

No, there is a major, I'm not making this up, right? 'Cause the boats come through the lakes and like, it was like a major hub of transactions.

Is that 'cause Chicago has trains and trucks? They do, they do.

This is real, right?

Am I wrong? You're not wrong. I'll give you stats. So 30% of goods that enter America go through Chicago.

Thank you, thank you. 30%, not bad. John was right. I was right. You're like, oh, Chicago, why is such a weird pick? It's not. It makes a ton of sense. Well, great to meet you, Matt.

Congrats to the whole team. And hopefully, hopefully you're back on the show this year with some more news.

Really appreciate it, guys. Thank you.

Love the show, by the way. Hey, thanks, man. Have a good one. Up next, we have Erik Andersen from Alloy Therapeutics. I love messing with you. You love messing with me?

Never gets old.

Well, without further ado, Alloy Therapeutics has raised a Series E. To build full-stack AI biotech infrastructure. Erik Andersen is here live. What's going on? Erik, how are you doing? I'm doing great today. Yeah, good to have you. Thank you so much for hopping on. First time on the show, please introduce yourself and the company a bit.

I'm Erik Andersen. I'm the CEO and the founder of Alloy Therapeutics. We're a biotech infrastructure company that works with a bunch of companies all over the world, helping people discover and develop drugs.

Okay, infrastructure. What exactly is going on in the biotech stack? I imagine everything from centrifuges down to writing PDFs for the FDA, to trials. There's so much that goes into that. What do you focus on specifically?

You got that. So the technology we have is really around more drug discovery and then into drug development. So the folks that do regulatory, the folks that do clinical, we work with them to discover drugs with the companies that we help to support. So we work with big pharma companies, we work with small biotech companies. Companies. What's going on in the industry today, the big trend right now is when we talk about infrastructure, who's going to actually do the wet lab work today in the lab? In addition, we've got everything going on with TechBio of all of this in silico work that's happening that's really exciting and sort of bringing those things together.

Yeah. And the idea is like you can have like, you know, a million times more ideas for different drugs, but then you still have the constraint of the physical world needing to kind of be able to actually run and experiments until ideally we could simulate a lot more in the future, but maybe we're not there.

Exactly right, and you can simulate all these things, but it's a bit of the design, build, test, and then you learn. It's that loop we all know. And right now in the biotech space, we've made some incredible progress with GenAI and machine learning capabilities to come up with a lot of those ideas. And you gotta close the loop then and actually be able to test them. And then there's a lot of skill that goes in, just the skill of drug development of what makes a good drug. And connecting those things together is what we do really well.

So how much of this will wind up looking like an AWS for drug development? I can provision a certain machine and you have it and I can interface with it all over the internet and you will do all the physical stuff for me. Yeah, that's part of it.

So AWS just launched a service, Amazon launched a great service that connects these generative companies back to a backend of how you manufacture the protein and you test it. That's one small piece to the process, I would say. A lot of scientists out there, they can design things in silico and then send them to a lab like that.

That will help. Is that not like a crazy side quest for them? Like it's—

AWS has like centrifuges now and like— No, AWS is actually just connecting them together.

No, it's pretty awesome.

Oh, okay, okay, got it, got it, got it.

I think what they're going for there is that they wanna be the cloud. If you're gonna do your compute to come up with these in silico things, they wanna make sure they see all the traffic and then just connect it to anyone else on the backend. That's I think what they're doing.

Got it, got it. So can you zoom out for me and just do a temperature check on biotech? We read one article about how Boston's going through a really tough time at the same time. Like every time I open up the Wall Street Journal, billion-dollar acquisitions.

Yeah, so last year felt like the dark ages for biotech. Yeah. Felt like everyone, every biotech investor that we were talking to was like, I don't even know why you'd keep investing in these companies. Like, the returns have been so bad. And then this year is like the biggest year of biotech M&A since 2010.

It feels like they're just rich from the GLP-1 boom, but I'd love to hear your sort of narrative setting.

Certainly the GLP-1 boom. Well, one of the trends that's going on here is that the industry has to restock the shelves for new drugs. As drugs go off patent, there's these revenue cliffs. And so Big Pharma is sitting on probably the largest pile of capital they've ever sat on, and they look to acquire a lot of their innovations from small companies. Those are the folks that we support to create new medicines. So there's definitely a huge M&A trend that's happening this year, and it will continue to happen for the foreseeable future. A big thing that's happening, as you read about that news in Boston, though, is We're doing drug discovery and development here in the US, but there's been a big shift to move it overseas and to offshore it, and for pharma to acquire assets from outside of the United States. And so that's been more of the big, I would say, the headwind for the domestic biotech space.

Okay. How are you feeling about just acceleration in drug development generally? Like, you know, in cybersecurity, Agentic Coding, like there's these very clear, scaling laws. And, you know, we have the meter chart of the amount of time an AI system can work on a single problem. It's doubling. It's on a very clear trend. I haven't seen a chart that's like that for biotech, but it feels like we're going through advancements and we're just getting more cures. So something's happening. But are you tracking it at a quantitative level yet, or just qualitatively, how do you feel about drug development?

Sure. We're living in an era right now where the trend of all of the drugs that we're discovering, we have an acceleration in the amount of innovation that's coming from our labs. So, so that is just an incredible trend that is supporting everything we're doing in the lab. So that the byproduct of that is, of course, generating massive amounts of more data, making sense of that data. And this is where a lot of the machine learning is coming in is how do we digest all of the data that's being created in the industry, making sense of it and then turning that into new cures as rapidly as possible. So that is an enormous trend right now in the industry. Yeah. And the advances that we have just from literally using Claude and OpenAI even in the lab and just day-to-day things. Yeah. You're seeing a handful of things come together. Huh. So first of all, you see this explosion of data, you see the wet lab capabilities largely looking the same. And what's happening then is we're trying to organize that data and turn it into new cures as rapidly as possible. And in the background, what we have is a number of new companies, new entrants in the tech biospace that I think are just natively better at understanding this massive data problem and being able to make sense of it. And then on the other side, you have the pharma and biotech, which are actually, you require all of those skills to actually make sense of what's gonna work in like a human biological system. So what's coming together this year, I would say, is that there's a lot more folks in tech bio that are getting better. But then we're just seeing the bio folks actually get better at tech. So basically you're at this place where like the tech bio folks need more bio and the biotech folks need more tech. And if you're bringing those things together, I do think we're gonna have an acceleration in a lot of the cures that we're making.

Interesting. Jordi? Give us a view into how big pharma executives are thinking right now. Are they confused why everyone and their grandma is injecting themselves with Chinese peptides, yet at the same time, like, more skeptical of vaccines than, than maybe ever. I feel like this kind of this interesting dichotomy. Yeah, it's—

well, in the industry today, I would say the executives in pharma are looking at it from the same place they always do, which is about efficacy and patient safety. And so we're looking for drugs that work and then proving that they work in animals and then ultimately in humans. So peptides being injected into humans, I think everyone's just saying, hey, what's safe? What's efficacious? The FDA has done some pretty amazing things in this administration, I think, to give flexibility for what's allowed. There's been some incredible changes that have happened that I think will accelerate the pace of innovation, what we do in the regulatory space, in the clinic as we're testing it in humans. And so pharma is apprehensive about the changes, but overall they're excited that things are moving along and that we're going to see a lot of new drugs coming on the market.

Can you help me understand the flow to get to like a data boom in biotech? Because I would assume that that's more driven by like dropping costs of DNA sequencing than anything from the GenAI world because— but maybe there's something where the GenAI world can process data that was previously locked up in PDFs or something? Like what is actually driving the data boom?

It's those things really coming together. So certainly the continued falling cost of DNA sequencing. Yeah. And then the falling cost of all the other types of data that you can generate along with when you're sequencing someone's DNA. Yeah. So if you go to a place like Function Health here, like what you can pull off your Oura Ring or your Apple Watch, there's actually a massive amount of data that is being generated passively And right now, the ability to connect that data to, I would call it real-world data, that we can connect to what's happening in patients is a level of complexity that we just haven't seen before on the data side. But it's gotta be rooted back in basically the wet lab. You're describing these things of taking your DNA or your tumor DNA and you're learning from what's actually happening in a patient. And bringing those together is actually a huge data problem.

I imagine you're talking I know you work with mostly like big pharma or real biotech companies, but we're also seeing this boom in a single person vibe coding a cancer cure for their dog. Yeah, it's pretty crazy. Do you have a policy for how small a team can be? Because I imagine it's not far from you getting an inbound from a single person who's like, I wanna do this by myself. I'm a solo founder.

They're like, I can be trusted.

Interested with advanced AI bioinformatics. And maybe they should, maybe they shouldn't be. I don't know. I just want to know how you've like grappled with it. Maybe it's just not a business concern, so you don't need to worry about it, but I'd love to know like how are you processing that idea?

No, I mean, I think everyone can be trusted with this generally. So that's actually not the problem. Don't worry about kind of these fearmongering that's happening on this front. Okay. At our company, we service anyone who is interested in discovering new drugs. And so that's a lot of what we try to do with our company is how do we democratize access to many of these tools. Yeah, yeah. We started in a particular place nearly 10 years ago, and that was before we had any of this GenAI work going on. Sure, sure. And it is lower, that single-person biotech company, or maybe just more of the, the virtual biotech company where you've got 5 or 10 folks and they can work with a different contract research organization to help do their drug discovery and development. Yeah, that's been a trend for a while. What's happening is we're just getting more efficient at it today. Yeah. So I Do you think the future biotech company is gonna be much, much smaller and be able to plug into these different resources and be so much more efficient at coming up with ideas and testing them?

Yeah, yeah, I mean, it is interesting. Tech loves to sing the praises of like the 5 or 10 person team, but there are a lot of biotech companies where it's like one genius researcher who discovered something and a couple support staff and then a lot of outsourced stuff all the way to, to, okay, we're ready to sell it, and it's mostly just a research org. This is not entirely new there. You mentioned that you're not worried about doom related to bio, but how are you grappling with the idea of somebody making a super bubonic plague or some new flu? Do you feel like there's safeguards, or how have you processed that? Us that question.

There, there's lots of groups of folks that are worried about this, and I think the United States military and some of the, the consortiums that we've been involved with are giving that thought. It really comes down to, though, is you can make a lot of things that might be dangerous. There's always a bad actor problem. Yeah. The way that we think about that is through our biosecurity division. We try to make sure that the capabilities that we need to have in the United States are always available and accessible to us. To call it medical access capabilities. We've learned a lot of lessons in COVID. Yeah, everything from could we actually just test for, for viruses in the community and do we have the reagents available to do that? And then of course we had masks and everything else. So from our perspective, we think about biosecurity as a way of making sure that you do all of those components from being able to surveil what's happening in the theater all the way out to can we discover and develop a drug very rapidly. And there's some great different initiatives that are going on of these ideas. Can you go from a threat to 100 days later actually have a drug ready to go? And our company and others participate in making sure that there's just a really robust response available with incredible supply chain and sort of make sure we can do everything here in the United States at all times.

Yeah, yeah. I mean, it certainly seems like the advancement in the positive side, like we're seeing a major, major swing to the upside right now. Every time I hear about one of these new medicines or new treatments, it seems really positive. There was a huge story about pancreatic cancer recently, which was a complete wipeout.

Right now we've got a conference that's going on and we saw two different studies that were showing just for some of the patients that were on drug, that it was 6 years later, they still had no disease.

And that's—

we saw another study.

And that's what we know about pancreatic cancer. It is really, really tough. Oh, it's really incredible.

It's one of the worst cancers we have. But I think this is the overall trend. We are living in a world where we will cure everything that is curable, I think, in the next 3 decades. It's really just a question of— and maybe even faster than that. It's a question of how fast can we accelerate and really give as many people as possible access to these incredible tools. And just like in tech, I mean, where you saw these very small companies be able to do incredible things in the last 5, 10, 15 years, I was investing back during the first internet bubble, back in venture capital 25 years ago. And in those times, the world just looked a different place.

We actually, middleware was like— Were you 10 years old?

I'm like 48 right now. So yeah, I was still a child.

You look late, you look like late 30s. So that's good for you.

I was a child at the time. But it was incredible because there was a lot of the same criticisms. That we're hearing in the tech space right now and as it relates to biotech, it was like we were wrong about everything back then except for what we were right about. Yeah, we did see the flow of investing coming in and many, many things went bankrupt, but the things that didn't go bankrupt changed the world. Yeah. And that exact same thing is happening today that you're going to see an incredible number of winners. And I do think biotech is going to be an important part of the big market that everyone's going after in the AI and the ML space as well. The tech space is going to go conquer biology as much as it has in other places. But I would say also that pharma and biotech very much have a seat at the table. Our academic researchers have a seat at the table there, and it's really about bringing the two domains together. They're going to be critical. One of them is not going to go it alone without the other.

Again, how much did you raise? How much did you raise?

Raise $40 million this time. Yeah, we hadn't raised any money.

There we go. Thanks, man. Congratulations.

Appreciate that. We've been working about It's hard to build a real business.

So yeah, this was the first round.

It's been a long haul for us.

This is our Series E. I'm a little old school. Back when I was a child, when I was a baby investing 25 years ago, I learned that it's okay to just letter your rounds just like normal. Our last round was in 2022. Yeah, I'm kind of old school in that way. So in 2022, we did a round. And between then and now, we really laid down the infrastructure so that we had operations across 17 time zones now. We're operating in the United States, Japan, the UK. We have an incredible group that's actually working in the GCC right now in the Middle East. Obviously, there's a lot going on there that's, that's creating a bit of a headwind, but we're very bullish on the work we're doing in Saudi Arabia, UAE, Riyadh and Doha and, and over in Abu Dhabi. So, yeah, bringing all this stuff together has been really important for us. And just our view is you stitch together the supply chain of innovation, you link it to great world, real-world data. And you bring a lot of AI and ML in there and we're going to really accelerate the pace of drug discovery and development. Very cool.

Well, thanks so much for joining the show. Great to meet you.

Enjoyed the conversation. Have a great rest of your week. Thanks for having me on, guys. Have a great day. Appreciate it.

Goodbye. Up next, we have Pippa Lam returning to the show from Sweet Capital alongside James Wise from Balderton Capital Partners. I believe they are in the waiting room calling in from across the pond. Across the pond. Are you both over in the UK today? They better be. Yeah, we have.

Given the news. Well, welcome to the show.

Why don't, Pippa, why don't you reintroduce yourself? And James, since it's the first time on the show, you can introduce yourself as well.

Yeah, big day. James's first time on TVPN, guys. Yes, we're very excited.

Fantastic to have you.

Yeah, so those of you who haven't met before, I'm Pippa Lam. I'm a partner here at Sweet Capital. I'm also an active angel investor here. And also scout with A16Z, but generally sort of sit between the ecosystems of the US and the UK tech.

How do you— I didn't know you were, you were the triple threat there. How do you do—

how do you decide who gets, who gets the deal? Well, that's interesting.

No. Yeah. Secrets. Yeah. James. James.

Yeah. First time caller, but longtime viewer. Thanks, guys. Great to have you having me on. So I'm a general partner at Balderson Capital. We're a $7 billion venture firm. We're based here in London. We used to be called Benchmark Europe. There we go. We got the buzzer right. So back, back in 2001, we were Benchmark Europe. We, we spun out in 2010 when the European market really took off. And for the last 4 months, I've also been helping set up the UK Sovereign Fund. Okay. Yeah.

Take us through that. I think that's what we're here to talk about. How long have the talk's been going on. How big is the fund? What's the strategy? What makes it unique in that it's so linked to the UK specifically?

Yeah, well, about a year ago we kicked off this big AI opportunities plan. There was huge news about it at the time and we've been working through all the various parts of that. And Pippa can talk more about things we've been doing in data centers and compute. But the fund was set off to really accelerate some of the UK's big UK AI successes. But with a lot more than capital. So we've got about £500 million. It's a starting point. It's the entry ticket in this game. But more importantly, we're providing access to UK supercomputers here. So huge amount of compute. We're providing fast-track visas. We're providing unique datasets. We're providing government as a customer. So procurement through government, basically to give a lot of the companies here that could benefit from working with the government a massive boost.

Yeah, I have to say. That supercomputer is such an underrated term. Whoever thought to rebrand supercomputer to data center is an idiot. Terrible. Terrible. Because data center sounds so boring and terrible. Supercomputer is futuristic, cool. Supercomputer is awesome.

Not gonna take my job. I'm so into supercomputers. I've always wanted to use a supercomputer.

Yeah, so, but— It's your grandfather's data center, right?

It is.

It's the thing that supercomputers, like what IBM used to build.

Yeah, yeah. And they have them in like cool scientific locations, like CERN has one. And like you always hear about, oh, they're doing astrophysics on them. There's so many cool things. Yeah. But that does link to, you know, where will the UK get the most leverage out of investing? Because I imagine that there's not a lot of value in just trying to like clone Google Search or Instagram for Europe. Like those platforms, it's fine. But at the same time, sovereign AI, does have value. Where in the stack does that live in your mind?

Yeah, so we're building out a whole range of talent, right? So we're going to be doing everything from electron to token, but obviously working with American partners, doing that, and international partners as well. Sometimes at the chip level, sometimes at the model level. And then we're starting to focus on where the UK has huge advantages. So life sciences is a huge area for us, right? If you look at Isomorphic, which is actually a spin out of Google. Yeah. You know, they're based here in London. Yeah. The whole team basically is European and they may be the closest we get to an AI company getting a drug all the way through to approval. That sort of end-to-end AI design. Yeah. Outside of life sciences, loads of stuff in physics and material sciences. So a lot of the AI science end of the spectrum rather than the AI slop end of the spectrum.

Yeah. Pippa, are you seeing Are you seeing an effect where entrepreneurs from around Europe are moving to London in the same way that folks from Chicago move to San Francisco to do startups? Is there a proper movement to make London the destination for ambitious European founders these days?

Yeah, I think it's— I'm obviously fairly biased having spent a lot of time here, but I think that the UK has always batted above its its weight in terms of the deep R&D pockets we have here, you know, in terms of universities, the talent that are coming out of the local schools. That's already created a very fertile ecosystem. And I think that, you know, one of the reasons we wanted to jump on here today was because it feels like, you know, UK is having a bit of an AI moment. It's not to say that we weren't already at the forefront of many of the innovations happening here. You know, of course, we always talk about Demis and as Hannah Sabes from DeepMind that was founded in the UK. Obviously, we would have loved to have kept that here, but of course it also went with Google. But, you know, we have long been at the forefront.

I think it's super critical that DeepMind maintains such a big presence in, in the UK because you guys now with the Sovereign Fund, if someone wants to leave DeepMind, spin out like you guys can be there to provide capital.

And exactly. And that's what we're seeing now. I think both from the idea that we want to be an AI maker, not just an AI taker. We're seeing, you know, grassroots innovation come out of the schools here. But as you say, we're also having people either leave DeepMind or so in terms of our partners across in the US. I think, you know, James can correct me if I'm wrong. I feel like a lot of the, you know, the first destination headquarters within Europe is huge. Usually in London. I think that that is pretty much most of the large AI companies we've had. So I think that there is definitely, you know, it's always been a good hub for AI, but we are seeing a lot of additional tailwinds, of which sovereign AI is one of them at the moment as well.

We went through this period last year and it might still be going on. I don't really have a pulse check on it where every AI company seed round was in like the hundreds of millions of dollars. And I imagine that you're not going to spread this fund over just 5 companies. And so do you think that—

No, John, it's one bet. It's one bet.

But I mean, do you imagine being like, like investing alongside other funds and doing more structured rounds where a lot of capital comes together to take like big swings? Or is there actually some sort of structural shift in in the type of startups that are getting built today, where $5, $10, $20 million can actually put some points on the board early on and get in the game in a meaningful way.

Yeah, you can see that you guys can make a number of bets in the app layer. You talked about energy, maybe some more science-focused, Neo Labs, maybe Neo Cloud. Like, I feel like $500,000 will really, can kind of like seed a bunch of different players.

But yeah, how are you thinking about the dynamics of early-stage startup fundraising right now?

Yeah, I mean, in the UK, actually, we are seeing those $100 million and even billion-dollar seed rounds now. So there's a rumored billion-dollar seed round in a founder who's out of DeepMind, which is incredible.

Let's go.

And there's been a bunch of those in the last few months, which is great. The fund is actually doing it the other way around. So we're saying— We're going to give you tens, if not hundreds of millions of dollars possibly, of government procurement contracts. You can build it. If it's good enough, if you are world class, we're the customer. We're ready to buy. Sure, sure. Same thing with compute. We're able to give a significant amount of compute to early stage companies. The quid pro quo is, hey, the taxpayer is taking all this risk backing you. What's our upside? We want to be on the cap table. We want to do it on commercial terms. It's going to be on the same terms as an Index, as a Balderton, as an Accel, as an A16Z, whoever's there. We're going to be good partners. But the money isn't just about unlocking upside and getting, you know, getting on the cap table. It is also about helping the companies to some degree. And obviously, you know, still £5-10 million can make a difference. I don't want to be like too flippant about it. You know, those things at this stage do matter. Yeah, but, you know, we're here to help the taxpayer get a bit of benefit from all the risk we're taking supporting all the AI companies.

Yeah, I think that's also one of the things that jumped out at me as well as, you know, as a commercial VC, it's hard to compete purely on capital, right? Especially in AI right now is where the numbers become so, so huge. Yeah, I think that something I've been really excited to see from the Sovereign Fund is, as exactly as James said, is this kind of hands-on ops help, whether it's how to navigate large datasets that the government may have access to. That, you know, portfolio companies will have help navigating, whether it be early procurement opportunities, as James said. And then obviously we're going back to the 1 million GPU hours of compute from the supercomputer, which is, you know, a huge, huge help. So, you know, it's not purely trying to compete purely on a capital basis, which I think, you know, frankly, would not play to the strengths of what the fund what AI can do.

So AI is extremely popular in China and India, deeply unpopular in America. Where do you think the UK will land? Do you think there's a chance that the population broadly will be supportive of artificial intelligence?

They're like, "AI allows me to spend more time at the pub." I suppose.

That's the way to sell it. That's what I'm saying. No, look, the UK, we've been pretty strong and early adopters almost of every wave of new technology, right? Whether it's financial services, like, you know, paying with this thing, the Brits are way ahead, earliest adopters in the world. And when it comes to sort of first-generation AI tools, your Claudes and GPT, I think the UK is number 1 or 2 in Europe for adoption. So, so far so good. But of course, there's loads of issues. You know, I think that there's going to need to be some really strong political leadership to explain to people the tradeoffs between, you know, these things can be transformative, make, you know, your wealth and your health and your security of your nation better off. At the same time, you know, we're going to need more energy. There's a load of issues we need to navigate on online harms and copyright. So some of those battles have been fought publicly. Some of them are still to come.

Yeah. You say that there's an opportunity for the government to be a buyer. I think a lot of people jump to defense and military. I'm more interested in if you're seeing any opportunities in non-defense sector opportunities for efficiency. I feel like, at least in America, everyone laments like the DMV is a huge weight, and like if they just had, you know, a piece of software instead of a physical form, things would speed up. Are you seeing opportunities for startups to increase efficiency across non-military portions of the UK government? Oh yeah.

I mean, I mean everywhere. And I think, you know, there's a bunch of businesses here in Europe. I'm sure there are in the US as well who are using AI for you to complete procurement contracts, you know, these like 400-page things you have to do. So Private companies have been doing that in response. You know, the UK government at least is already smartly using AI to read them as well and prioritize them, right? They don't make any decisions, but they help you navigate some of these processes. Same thing in transcription and voice. Like, obviously, we're already seeing GPs and doctors benefit hugely from being able to use these tools to quickly take notes and actually look at the patient rather than spend all their time sitting in front of the— computer filling out the health records. So there's been some early wins. And I think, you know, part of the job at Sovereign AI is to work out which of the companies we should work with and the government to see where else those wins are.

Yeah, makes a lot of sense.

Jordi Hays, super smart. I hope that you can work out like on the cap table to just say the United Kingdom. Yeah, because like I feel like as a founder, if you just see your country on your cap table, you're like, well, I got to, I got to, I got to deliver. I got to deliver. I deliver. But very cool. I think together we're done. Yeah. The country's lucky to have you both, you know, leading this effort.

Well, thank you so much for taking the time to stop by. Have a good rest of your day and we'll talk to you soon. Congrats. Goodbye. Bye, guys. Bye, guys. There's some huge news in the world of robotics because they made a slot machine that can follow you across the casino floor.

We actually got a slot machine stalker before a humanoid demo. Yeah, this is before— I guess not demo.

Well, this is before it can actually fold your laundry. Like we've been seeing a lot of laundry folding demos.

We got the slot machine. Never ask a woman her age, a man his salary, or humanoid robot founder to let you be alone with the robot. For 30 seconds.

Okay, but what is actually happening here? Because obviously, like, the joke is that it will follow you around so that you never stop gambling, but that can't be why they actually built this. It must be because they want to be able to reconfigure the layout easier? It's probably that, but also the novelty.

Okay.

Oh, if you see something moving around. If you're walking through. You'll be like, I gotta chase that down, and throw a couple bucks in.

Oh, that's funny. It's a robot slot machine.

This also looks like— I don't know what Novomatic is, but this does not look like it's at an actual casino. This looks like it's at a trade show for casino equipment, which I think might be— this is a demo of something that this company is going to try and sell to casinos. I don't know that this is actually at a major casino just yet, although people are walking around, but it looks like It looks like trade show bags to me. I don't know. This says trade show all over it. I do wonder if this is being teleoperated or if this is end-to-end machine learning. I need to know. I need to know the tech stack. And I need to know, is this truly autonomous, or is this just being puppeteered by an Xbox controller? Because we've had— we've had— this is not— this could be just a remote, a remote-controlled car, which has existed since like, what, the '80s, maybe, maybe longer. But we're in this, we're in this phase where we want to layer on, oh, this is AGI, this is, this is truly AGI. What else is going on in the timeline, Jordi Hays? Anything else? The must-have item in Silicon Valley is a $178 sweater with a CEO's face. Leaders from companies from NVIDIA to Palantir are now driving fashion. Signaling a new era of the cult of the founder. We saw Nick on our team rocking the Jensen sweater from GTC. It looks great. This is a very beautiful sweater. It does look great. Very funny, very silly, and just a nice departure from just a normal t-shirt. You know, for what, 20, 30 years, the tech merch was just a t-shirt with a logo on it. That's fine, but why not mix it up? Why not go in a different direction and make a sweater with the CEO's full cartoon character on it.

Why not? Dolly Bolly says, is showing a screenshot from, I believe, Biz Buy Sell, of a laundromat, which is selling. It's got $421,000 of EBIT, asking just under $3 million. So getting a better multiple for your laundromat than most, than most public SaaS companies out there right now. And it was established in 2024. Wow. They just made this business in a couple of years, lifestyle, and they're like, I would like $3 million for it now.

Okay, couple more posts. John Fio, friend of the show, says the Sphere is probably the most important piece of architecture in the last 100 years. It's a hot take. It's what the VR trade was trying to be, but manifested in the real world with a real novel experience instead. It's what Apple and Meta were trying to go after, but failed because they tried to shove it into a scalable box instead of building for real life. A Sphere in every major city will be a proprietary technology for a new kind of stadium. It will suck in only the best acts, and they'll stop playing regular stadiums. It's the perfect example of mixing real novel tech with real novel life. This is how you capture value over the next cycle. Article. And I agree with this. I think this is a great take.

So, uh, yeah, rewind a year ago. You're like, companies are going to announce hundreds and hundreds and hundreds of billions of NVIDIA orders. Do you want to— meanwhile, we have the Sphere, which they built. The Sphere, it's one concert venue. They built a, they built a really cool concert venue in Vegas. They got a bunch of debt. But it's awesome. Which one do you, which one do you want to own? Nvidia or the Sphere? And of course, on Nvidia, you were, you were still up, you know, 100% over the last 12 months. But if you had bought the Sphere, you were up 442%. Yeah, did very well.

I made a whole YouTube video about the Sphere 2 years ago and was pretty bullish on it. Had dug into the founder and how it got built. And it was just a fascinating, fascinating story. But I think he's right that there will be a Sphere in every major city. There is technically a Sphere-like location in Los Angeles over by SoFi Stadium. It's not technically a full Sphere with LEDs on the outside, but it has a big screen. You can go and watch a football game there. And I think it's doing well, as well. I haven't dug into that one nearly as much, but I do think that these types of like immersive experiences. Uh, the Sphere is unique because it grabs, uh, attention from all over the world. You fly by on a plane, you just see it, and you see the emoji on it.

Kind of a miss that we've never done anything with the Sphere. No. Uh, last little white pill here. Yeah, what's up? Uh, Meta announced Level Up, a free 4-week training program that takes people with no prior experience and prepares them to work as fiber technicians on data center construction sites across the US. We built this program with CBRE because the fiber technician field and the broader construction industry is facing a nationwide shortage at a time when data center demand is higher than ever. And I'm sure people will come up with reasons why this is bad, actually, but I think this is, I think this is great. It's a great opportunity. And Tyler, we didn't get a chance to talk with you about, talk with you about this before the show, but you're actually gonna be going through the program, starting Tomorrow we got you a slot. Fantastic. And, uh, so get— this actually seems fun. I would be interested in doing this.

It's data center. They made data center simulator in real life. They made data center simulator in real life. That's amazing. Uh, Tyler, do you ever play Elden Ring? No. No? Oh, it's such a good game.

More like online games. More on—

Elden Ring can be online. Okay, I don't know. You can play with people. Uh, they're making a movie about it. The live-action adaptation of Elden Ring produced by A24 in partnership with Bandai Namco and filmed for IMAX, is slated for release March 3rd, 2028. Wow, that's a long ways away. Production will begin in spring. But if you haven't played Elden Ring, it's a fun time. It's really, really hard and sometimes it just gets like a little bit too much, but it's a good time anyway. Well, folks, thank you for tuning in. We will be—

it's been an honor and a privilege.

Leave us 5 stars on Apple Podcasts and Spotify. Sign up for our newsletter at tbpn.com. And Flashbang out. We'll see you later. There we go. Rolling Flashbang. Rolling Flashbang. We love you. Goodbye.