ROTH MKM: State Of AI Venture Landscape with Untapped Ventures Webinar May 16, 2024

Hello everyone. Thank you for joining. My name is Rohit Kulkarni. I'm the Internet analyst here at Roth. And today we have a interesting different type of a topic as in we have George who's a venture capital investor. He and focuses more on AI early stage startups. So we have the way I always, when I talk to people like George who are so far early on any theme, it always helps us frame what is the shape of things to come as far as two years out is concerned. So public companies, what could public companies be looking at from a thematic standpoint from. From a MA standpoint maybe 12, 24 months out as far as AI is concerned? And of course AI is something that is moving so fast. Just earlier this week, I think it's been a very Busy week for AI. OpenAI had their big event, which kind of was a big event, but didn't again, would love to get your take, George on that. And then Google had its IO conference. Also a lot of AI buzzword bingo. But also I think Google is playing it smart in my opinion. So yeah, it remains something that it's hard to keep your tabs on. So glad to have George just cut through a lot of the themes and companies that he may be keeping tabs on. He does have a nice slide deck. So we will go through the slide deck but not completely, but again we'll keep it informal and we'll try to pick his brains as much as possible. So George, again, thank you for doing this. Kick off with a quick intro and then we can run through your slides as well again.

Okay, sounds great, Rohit. Yeah, it was a busy week in AI, but it feels like every week is. But I do have some highlights from, from OpenAI and Google's announcements that I'll cover for sure. So. All right. You could see my slides. Okay.

Yep, looks good.

Perfect. All right. So that.

How did you. How did you go there? Like, what is going on?

It's my AI. It's my.

This is smart. I Amanda like that Tik Tok view.

Right. It's pretty cool. Yeah. It's actually available in the latest version of Zoom. Anybody could use it. It's just when you're sharing your screen, there's like these four options now and it's like the second one. So there you go. My first value out of the day. I'm gonna try to deliver massive, massive value to, to all of your audience. Let's have some fun. So yeah, yeah, I think what we're going to cover, it's a little bit different than my usual talks, but I think it should be pretty exciting. We're going to talk about what's happening with AI, the latest and greatest. And really what Rohit asked me to do is, you know, how is what's happening in the early stage AI sector relevant to what's happening in the public markets where, you know, your expertise and your, your audience's expertise is so, so, so I'll make some attempts to make some bridges. Well, we'll do some Q A. And Rohit, I know we want to make a conversational, so feel free to interrupt at any time, but otherwise I have a pretty good flow to get through a good chunk of the presentation. Quick intro on me. So I'm here in la. I was born and raised here, went to school at USC. My passion with AI began in my 20s when I read that book on the top right there, Ray Kurzweil's Age of Spiritual Machines. And I was just fascinated by the idea that we were going to build machines that would feel like they're spiritual. Spiritual as in not religious, but more so having emotion and sentience, seemingly. I went on to build a company called AMI that was in the AI and automation space, helping companies automate business processes and workflows using AI and computer vision and OCR and RPA robotic process automation. We grew that business, we sold it in 2018, and I started angel and fund investing and then I launched On Top Ventures in 2020. And so our MTP, our massively transformative purpose is to foster innovation and breakthroughs in AI and neurotech, to help us as a human species reach our untapped potential and become superhuman by today's standards, safely and ethically. That's kind of my why. And we have a couple moonshots in terms of raising a dozen funds in the coming decades and billions of dollars investing that in the leading AI founders and helping them uplift humanity and in the process building one of the leading global AI funds. So we're at the early stages of that, but so far so good. But that's, that's a little bit about me and untapped ventures. So with that, let's talk a little bit about the AI era that we're living in. So it's, it's important to pause here and just really take in the fact that AI is not just a new tech. It's a platform shift, right? Just like the platform shifts that came before it. This is what's going to drive the next five to 10 years, certainly the next one to three years. As we've already seen in the last year. And it's going to completely change and impact almost every aspect of our business and personal lives. And we see that more and more. Like you said, even this week with the announcements from OpenAI and Google, Google's putting it into search. So every time you search there'll be a generative AI response and I'll show some of the OpenAI demos. So, so, and it's important not to confuse it with some of the fads or hype cycles that came before it where there was the technologists that were hoping that it could reach mass adoption, but there was fundamental issues that we could discuss on why they didn't. AI is completely different. You go into chat GPT, you know, you could be, you know, trying to write an email and it'll help you write it. You could be a, you know, a new mom or parent trying to figure out something of what to do with your baby because it's your first time parent. And it could give you the right answer. So the immediate value that, that users can get in a business or personal context in a way that just didn't exist with some of these other technologies. So much so that Goldman Sachs, right, and others, you know, predict that it's gonna take a significant percentage of u. S. GDP, right? So it could be as much as 2% of u. S GDP and, and the analogy that's being drawn and you could see the numbers in terms of the total investment projections going from 200 billion to 422 billion in the coming years. But, but more than even the last slide I had, the comparisons are being drawn to the impact AI's impact on humanity to that of electricity, which is really fascinating if we take a moment and just kind of look at this picture and imagine what life was like at this time. And it wasn't just that electricity gave us light or cars, but it gave us everything. It gave us all electric appliances and it completely disrupted old industries. And for example, like the horse and carriage industry where within a few years there was just none of those on the roads and you had cars. And then, but not only that, you add in the kitchen, right? You add new categories. You had the kitchen appliances category and within there you had the refrigerators and microwaves and, and stoves and, and within there you had multiple winners in, right? So you had new industries under each industry you had new categories. Under each category you had subcategories and under each of those like subcategories you had like new winners. Whether it's Like Frigidaire or Whirlpool or. And then you have Viking and. Right. So my point is, so imagine that happening now where all these traditional players, some of them that might be some of the Big Fortune 500 and Fortune 5000, are going to get disrupted and you're going to have new players that are providing a new way, a better way, using this new tool set that's available to us. So I know that's a big claim, but just to kind of like take that in, Rohit, and go back to what you started with. So OpenAI had some big announcements, announcement, so basically they. They kind of snuck in their. Their update on Monday right before Google I o got started, which I thought was very clever of them. And. And so what they announced was the GPT4O O being the Omni model, improving on their audio, video vision and kind of emotion. So let me play a video for you guys. This is a clip out of that 20 minute video. I recommend you watch the whole thing. But. But here is a clip from that that I think really shows how much better it's gotten and that it's really now moving out of, you know, on your computer, chatting with ChatGPT to on your phone and having a totally different experience with it. You end the story, but do it in a singing.

And so Byte found styles and it.

Really has a wide dynamic range. So I'm going to fire up another demo for that. Yeah. Hey, ChatGPT, how are you doing?

I'm doing fantastic, thanks for asking. How about you? Pretty good, what's up?

So my friend Barrett here, he's been having trouble sleeping lately and I want you to tell him a bedtime story about robots and love.

Oh, a bedtime story about robots and love. I got you covered. Gather round, Barrett. Once upon a time, in a world not too different from ours, there was a robot named Byte Bite. Was a curious robot, always exploring.

I started this story, but I want a little bit more emotion in your voice, a little bit more drama.

Got it? Let's add some drama. Once upon a time, in a world not too different from ours, there was a robot named.

No, no, chatgpt. I really want maximal emotion, like maximal expressiveness. Much more than you were doing before.

Understood. Let's amplify the drama. Once upon a time, in a world not too different from ours, there was a robot named Byte Bite.

Can you do this in a robotic voice?

Now?

Initiating dramatic robotic voice. Once upon a time, in a world not too different from ours, there was a robot named Byte. Byte was A curious robot always exploring new circuits.

Barrett here, calm down a little bit. Can you end the story but do it in a singing voice?

And so Bite found another robot friend and they lived circuitly ever.

Okay, so you get the point, right? Right. What did you think of that?

I don't know what happens to Siri now. Like what happens like I've never like on my iPhone. I don't know what happens to these assistants. And like the, and also the velocity with which that AI was adjusting to all the prompts. I don't know whether this is just a demo related velocity or.

No, that's part of what they're showing off. So in the past I've done the demo and you have to wait until, until it's ready and all that. Right. Because I've had it on my phone. I do this when I do my talks for five months. That was part of what they did was it's now live. So there's you know, it's. They've cut out that delay so you could interrupt it at any time. And. And then. Yeah. The range of emotions that it. It was able to show and just understanding tonality and singing and all that was really fascinating. I mean mind blowing.

Yeah, we shall see. It is, it is the, the rate at which are changing. Not only that interesting. So yeah, anyway. Cool. Cool. Yeah.

All right, awesome. So hopefully I have, you know, I have your attention by now. I always try to make these both educational and entertaining because you know, I think we could talk about all the exciting development. But seeing it is, is what really helps drive the point home for a lot of people. So now let's, let's drill down into the venture landscape where in VC land where I play every day. And so just kind of a quick recap. Right. So basically what we know is last year's venture market was in General tough. Global VC funding fell 42%. Deal count fell 30%. Q4 was kind of the worst quarter for VC in six years. Amongst all this bad news, AI was, was one of, if not the bright spot, the largest sector to show an increase. Basically one in four U.S. venture dollars went into AI and, and this year as Venture bounced back in Q1 of this year, 11% quarter over quarter growth, AI captured 11 billion of the 58 billion or 20% of global funding. So one in five dollars globally in Q1. So AI is doing really good. If we take a look at some of the largest deals here, you could see and it's clear that generative AI and some of the foundational models like OpenAI and anthropic inflection raise some monster rounds at considerable valuations. And the interesting observation here for your audience on the right is that a number of these were, the notable investors were the big tech, right? So they're not, they're not just the, the big VCs. And we'll talk a little bit more about that.

Interesting. Facebook doesn't show up in any of these, in that slide. So which is interesting, but just a, that they, they don't seem to invest a lot in AI startups, do they?

They haven't yet. And, and I think part of that was, you know, I think, I think at some point last year and I'm going to talk about it in a little bit when I get to the Magnificent Seven. But you know, it, it took a while to get Zuck to give up on the AR VR and focus on AI, but I think they finally got him there and so, and then they released Llama and you know, really came back onto the scene strong, which I'm glad they did because that, that was a smart move. But basically what we see here is that we have AI startups are commanding a higher valuation, right? So kind of 20% higher at seed, 40% higher at series A, 60% higher at series B. And again, a large part of this is based on that corporate attention, the big tech attention that they're getting as big tech in Magnificent Seven and others are trying to make investments that put them in a position where they could be close to the innovation that's happening. So with that, let's keep going and let's talk about one of the, the, the stars of the public markets, right, that Nvidia has been really riding the AI wave. And, and let's kind of break down a few things. So you know, there's something with, the way we look at it is there's the four levels of value creation and generative AI. There's the apps on top. You know, these are like all the different apps that do the functions. There's the infrastructure layer, which is the tooling. There's the models themselves, like the GPT4O and then the compute, which is the, the chips. And that's where Nvidia lies. And so yeah, so obviously, as you guys know better than me and Nvidia stock has done great. They've had a 239 run in 2023 and up over 80% so far this year. First chip maker to become a trillion dollar company. Sales are soaring. Projected to exceed a hundred billion by next year, you know, but I think there's some issues, right? The monopoly is potentially unsustainable because of the rise in competition and the rise in the, it's the lack of energy efficiency and there's low power alternatives. So specifically I think the stat is that basically if you look at that tweet, it says that next year Nvidia's 3.5 million units will effectively consume what is the equivalent of the, of the entire power consumption of countries like Guatemala and Lithuania. So there's a lot of people working on better versions that are, you know, because now it's still the gpu, right? It's not an AI chip from the ground up. And other companies like Google who just announced Trillium last week, or sorry, this week, they're better, faster version of their AI chip and many others are going after the chip landscape. You and I were talking about Grok before we hit record. So Grok is on there and there's others on there that are going after it. You know, you have small earlier stage startups like a couple that we've invested in. Like one is Lemurian Labs that you know is creating something called an spu. So you know, they're, they're taking, you have all these founders that are taking concepts from physics and biology and taking entirely new first principles, ways of, of attacking the problem. Because just the Moore's Law, we can't keep going the way we're going as it relates to the compute needs. So a great example of that is also Extropic. I was with the founder Gil Gill, Gilliam Verdum of Xtropic. What's interesting is, you know, he was one of the leaders in the quantum computing field, as a leader, saw the problems with it and said we got to figure out a way to do it differently. And now is building a startup AI chip startup using concepts of biology and physics, specifically in thermodynamics. And so in quantum computing it's all about trying to isolate the noise so you can do what you need to do. And in thermodynamics, it's actually inviting the noise and leveraging the noise to be able to do things better and faster. I can go more technical, but I'll leave it at that for now. So that's kind of like a little.

Bit about at a high level. Are you seeing kind of startups or private companies like Xtropic solve a specific kind of a problem that people have realized in the last 18, 24 months, or are they rebuilding the entire kind of chip architecture itself? Like, is there a specific pain point that they're trying to solve or is this something completely redoing what Nvidia has been doing or some of the other AI chips are doing?

No, I, I think that like these generally, like these examples, like they're, they're trying to be the next Nvidia. Right? So, so, so there, there's a lot of demand. The AI chips on the market are backlogged, they're high priced and they're inefficient for AI compute. So there is great opportunity and there is a bunch of well funded, really, really smart founders that are working on getting to market, getting to scale, and at different levels of proof of concept of proving it out, you know, in actual chips and then getting that to scale.

Okay, okay, very interesting because I thought, like in this, Lemurian Labs, you had this, it was a curious little bullet that data centers currently consume 1.3% of world's total energy and is expected to quadruple by 2030 if GPUs are not updated. So that could, in the foreseeable future is going to be a pretty big problem in terms of energy consumption. So I guess hopefully companies like Lemurian Labs or other guys are looking to solve that one specific problem that is kind of ballooning quite quickly as such.

Yeah, exactly. And I think if you talk to any of these, if you talk to Jay or you talk to Gil, they'll say that there's going to be a place for different kinds of chips and you're, you know, you're going to use the right chip in the right place depending on your type of compute. So, you know, eventually. So we have CPUs, we have Nvidia's, GPUs, which were really the graphics and kind of gaming cards. Then you'll have like AI chips that are truly optimized for AI built from the ground up for AI. You'll have like quantum chips for quantum use cases. Like if we're trying to, you know, not quantum chips, but quantum computers for quantum use cases or when you're trying to, let's say, tackle a drug discovery problem and stuff like that. So it's not like. But there will be, there'll probably be winner take all or, or a few key winners in each category, but not necessarily like the whole market is going to move to one or the other. The problem with that, the reason you have what you're seeing with that stat with Lemurian and then the previous stat that I shared with, you know, is the inefficiency of the chip. Again, it's a GPU that's doing, it's a graphics. The compute for AI is different than the graphic compute for graphics. So basically that's the best thing on the market is, is these graphics cards, these GPUs from Nvidia that are being used in mass because it's the best thing on the market. But it's very compute inefficient and energy inefficient. Okay, all right, so, so let's just keep moving forward. Aside from Nvidia, you know, the other big tech companies are, are, are increasingly involved in AI. And I would argue basically most of them are AI companies. They've transformed themselves into AI companies like we were just talking about. Meta made that change. Tesla is AI on wheels. Amazon has been an AI company using AI for their recommendation and many other things. Microsoft started off as a, you know, software company and then a cloud company and, and now an AI company. Apple is the one that has kind of fallen a little bit behind. And you brought up Siri. I think they're going to be incorporating chat GPT in it. I could be wrong, but they are. They are going to be making some big announcements at their worldwide developer conference and obviously, you know, we'll see what they say.

And do you know. Yes, I have heard of some rumors like that. Is that something that. Do you know when is there wwdc? It's supposed to be sometime in like July. Okay, that's okay. I'll look it up. I was just curious because it would be a pretty big deal for somebody like OpenAI if they include ChatGPT into iPhones.

Yeah, yeah. I mean, I think it would be, it would be under Siri. So it would be. Yeah, so it'd just be a better Siri. But. But yeah, I mean Apple has the ability to do a lot right, because they control the computer, right. They make the Mac, they make the iPhone, they make the iPad so they could build in their own AI chips and their own AI stack at the edge. They could really be an interesting player in all this. And this is pretty typical for Apple. They're never the first entrant. They kind of wait, they see and then they, they come in usually with something that's pretty smart. But they're behind and everybody's waiting to see what they're going to do with AI. Yep, they had, they have started to make some investments and acquisitions and so has everybody else. I know this slide is busy, but if you look at kind of the key underneath each company you'll see like the different investments and again, big tech is pouring a lot of money into this, into AI and so because they see the opportunity for infrastructure and healthcare and AI agents and all the different use cases. But I just want to keep us going forward here. We've talked a lot about Nvidia and the Magnificent Seven, you know, but if we pan out, there's a lot of innovation happening in other spaces that we're excited about and following. You know we mentioned a little bit about using AI and drug discovery and design that's going to touch everybody in biotech and pharma, material sciences. Right. There's a lot of companies that are built on the industrial metals and all that. Like we're looking at decks all day that people are using AI to build, you know, new metal, new concrete, new wood, new foods, new new materials, which is a really exciting field. You have neurotech, you know, technologies that help us understand the brain better, use the brain, cure stuff in the brain, also model the brain to be able to then build things into the physical environment, synthetic data, robotics, quantum AI. So you know, as an example, this is another one of our investments, Climax Foods. They built an AI model with the 300,000 species of the plant kingdom and use that to create food that is award winning, winning like the best cheese competitions. It's, it tastes as good if not better, it's healthier, it's plant based, it does less harm to the environment. The Michelin star chefs love it, they're in the top restaurants in the U.S. users love it. And so there's just a lot of innovation. So you think about all the foods players that are on there and you know, interestingly their first two kind of global scale out partners are dairy companies that see the need for this.

And just dumb it down a little bit. George, as in what, what did they exactly figure out? Like they use some data to come up with new how to make new types of cheese or.

Yeah, so basically so, so, so that right. So Oliver is a Harvard astrophysicist. He led data science at Impossible Food, SpaceX and Google. They want to solve the problem of global kind of food shortages and hunger. The impact to the climate of. Right. The Fatulans from cows. We, we know that that's the biggest contributor to the ozone and, and issues. So okay, so here is a smart founder that wants to tackle a big global grand challenge. And so what I love about it for our portfolio right from our investment thesis is it's not a bunch of foods people and vegans that Said, hey, we want to create more vegan foods and let's use a little tech. It's technologists, really smart technologists, that said, why don't we build an AI model of the 300,000 species of the plant kingdom to understand each plant properties, taste, texture, dexterity, all these factors, and then have the AI model generate formulations for the different kinds of foods. So for example, for the blue cheese, which is hard to believe that this is, you know, blue cheese. And the mozzarella, the feta and the. And the other one. So.

Cool. Awesome. I know we have a lot of content. We just want to do a time check for people just to not steal George's thunder. I, I think in the next 15, 20 minutes, what we'll try to cover is essentially things like, if you're going to be investing in AI, what types of questions you need to ask, then how do you invest in AI in terms of where value might be created?

Yeah, let's do that.

Let's go through all of that and then we'll pause at the end for some questions that people may have.

But yeah, and if they want to start posting any of their questions in the chat or the Q and A, however you and Amanda typically like to manage it, you know, we can start to have those questions coming in. But yeah, I'll run us through that.

So that's a good point. That's a good point, George. Sorry I forgot to tell you guys, use the Q and A thing in Zoom to add questions and then George and I will. We'll try to weave that into the conversation in case is something that you feel we should talk about.

Great. So this is Peter, the Amanda, one of my friends and mentors. And you know, he has this quote. By the end of the decade, there'll be two kinds of companies. Those that are fully utilizing AI and those that are out of business. Might sound a little extreme, but I think there's some truth to it. If I was an investment analyst on this call, managing public market stocks, right. I'm not here to give advice, but I would be looking at my investments. How quickly are they implementing AI? How committed are they from the top down to becoming an AI centric company, how are they auditing and looking at their data and creating models with their data with advanced algorithms to gain new insights and sometimes not even knowing what the point is, but just to say, put the data in and say, hey, like chat, GPT or whatever model you built. You know what, what are some interesting insights you see about our customers or our markets or our employees or our finances. You know, these, these AI tools are so strong and capable now that you could just put in the data and start chatting with it and instead of knowing the answer you're looking for or trying to get it to a certain point, you just start open and you let it find things that are interesting. So that's what I'd be looking at. You know, I, I think, I think somebody that's done really good with this just to, you know, highlight a case study, you know, is Moderna. I, I'll only play a minute of this video but just so you get a sense of what somebody adopting AI at an enterprise level and making a full commitment looks like.

We believe very profoundly at Modela that.

ChatGPT and what OpenAI is doing is.

Going to change the world. We are looking at every business process.

From legal to research to manufacturing to commercial and to think about how you.

Redesign those processes with ChatGPT type tools. 90% of companies want to go do Genai, but only 10% of them are successful. They haven't built the mechanisms of actually transforming your workforce to adopt new technology and new capabilities.

We've been born as a platform company.

And we think about our drug development as a platform. And with the onset of Enterprise ChatGPT.

These GPTs became embedded into the workflows.

Of how our users are really using it. We're up to now 400 GPTs within two months, far more than what we expected.

Some might be surprised that the legal department was the first department to wholly adopt ChatGPT. Here at Moderna we have 100% adoption rate of ChatGPT and other generative AI large language models here in the company.

One of the GPTs that we're most.

Excited about is called Contract Companion. You can upload the contract into the GPT, ask it for a quick high level summary or even ask specific questions. And what it enables the legal function to do is to focus our time and attention on those matters that are truly driving an impact for patients to allow us to deliver on our mission here at Moderna. Dose ID GPT really. So from there, right, they, they go into Dose id, another example of they're using it in the clinical setting and actually asking it for dosing questions. Right. So what I love about this example is it's a mission critical, you know, enterprise wide, top down commitment, 400 GPTs in a matter of a few months. So you know, I would be looking for in my portfolio for something like that, right. And so, and there's so many different Tools. Right. Depending on, you know, what, what line of business they're in. I'm gonna skip through some of these and then just go on to the next section. But again, just, just to recap the part we're leaving which is for everybody, that's not a magnificent seven, just your traditional companies. I would be, you know, I would be really keen on understanding how committed are they to AI, how, how fully have they woken up to it and how enterprise wide, top down, are they, are they, are they leveraging this to both increase their revenues and decrease their costs at the same time having a double effect on, on their bottom line?

Yep, yep. I guess just on that topic, just a quick question on startups that are eating their own dog food or startups tend to be more nimble when it comes to using new tech and improving their processes using new tech. So where do you see kind of startups that you come across use AI more? Is it more for reducing costs or are they looking ways to make more money? Is there some, is, is there some common theme that you can paint this by?

Yeah, I mean the reality is it's both because it's, it you have the, like the sales and marketing tools that are just amazing that, that, that allow you to, you know, write better copy, create better graphics, be, do customer segmentation better, faster, have a customer chatbot on your website and, and you know, so there's that and then there's the, all the operational ones that help you, you know, the customer service chatbot, the using AI to better understand and make decisions and, and you know, regarding costs and all that. So I'm seeing it on, on both sides. Sometimes the AI is there to avoid having to hire new people and sometimes it's there to reduce the load. Right. And so you know, we see it, we see it in both areas almost equally depending on what the priority is, what the use case is. Okay, okay, so just real quick, just the last section here, you know, I'm not going to, I'm going to do the abridged version of the how to invest in AI. First of all, make sure it's a native AI. Everybody's masquerading around, slapping AI powered AI enabled on their deck. So obviously we're looking for AI. There's a lot of great tools and we think that enterprises should use these tools like we were just talking about. But we don't invest in many of these tools. We're just concerned that they're very saturated. A lot of them are chat GPT wrappers. And when OpenAI and Google and stuff make big announcements like we just saw this week. They put a lot of a whole category out of business. So. So we're more focused on the other three layers that I talked about. The infra, the models and the compute. That's where we've made a number of investments. So you know, we think like vertical models. Just like SAS went to vertical sas, we're going to see AI go to vertical AI. So like an example as we, this was our an investment we made. They were created a vertical model called Kepler to take unstructured healthcare data and make it structured. You know, that was our first exit in the portfolio. So I think those are the things I would recommend to the extent that people on the line are looking to make investments in private companies is like, where are they on the four levels of value creation? What's their moat? And like how mission critical is this? I provided some other due diligence questions here, but I'm not going to read these all out. I am going to share a link to, to the gift and if people want to grab it, they could grab the deck in our cheat sheet in a moment. So, yeah, so as we reach the end, you know, kind of what I always end my talks with, I really, really, really want people to walk away with being even more committed. Like play go fully all in on AI, you know, take the red pill if you haven't already, you know, with the matrix reference and go all in on AI. The A Team was one of my favorite shows growing up, you know, but build your not your A team, but your AI team or your AI team that's like the A team. So internally in your portfolio companies. And then, you know, obviously there's a gold rush happening in the public market, private market, and figure out how you want to play. If anyone on the line is interested in talking more about what we're doing that's unique and our unique investor community and all of our happy investors and all that good stuff, I'm happy to do that. Otherwise, here is a QR code. I'll also drop the the link in the chat if anybody wants to grab the the this deck as well as the cheat sheet that has everything that we covered. So I'm going to just drop that in there now if I'm able to. It looks like I'm unable to. So maybe I could share that with with you guys and maybe one of you guys could drop it in.

Yeah, yeah. Generally we don't do chats here. I don't know. I think. Okay, okay.

No Problem?

I don't know.

So with that said, yeah, I think that's, that's it.

Yeah, I don't know if it went through or not, but I did do something.

Looks like it.

Okay, I guess just one, one question that has come in chat is around GPUs and kind of investment cycle in GPUs. Have you, do you have a thought on kind of how long do you think the investment cycle in AI, at least the larger public companies, how long would that last? Particularly when it comes to building capacity right now, is there a point where we are going to have over capacity anytime in the near future?

So, you know, the, the singularity is, you know, the event horizon afterwards you can't really see. Right. And so I think it's. If we try to ask that, answer that question with a linear mindset, it's hard to see and I think we'd come up with some time span. But I think if we look at it over an exponential mindset and realize that it's the singularity and as we move into artificial general intelligence and artificial super intelligence, it's like asking people when they had the horse and carriages and the first cars were coming and, and, and, and asking them to say, well, how much of this electricity stuff do you guys think we'll really need? You know, we want to build the initial power grid and like, all right, well I guess we're going to have some street lights and some cars and so, you know, we need X amount and, and today we're probably like, I don't know, I'm not an energy analyst, but a billion times whatever. Right? So, so I think, you know, I think, you know, when we have billions of robots that are, that are in every factory floor, in every office and in every home and each, let's say your home robot is helping you with home tasks, right? Like folding the laundry, doing the laundry, gardening, you know, maybe changing the baby's diapers, whatever. Right.

So I'm scary though, though that, that sounds, but anyway, we'll, we'll see how that gets there. But I, I get your point. Like that's a FAS to think because right now it feels myopic in terms of what are the capabilities of AI and what's the capacity we need to handle those capabilities. But perhaps those capabilities are going to expand exponentially and hence the capacity also needed is going to have that similar level of exponential growth curve as such. Cool. Cool. I guess. And in terms of you, you mentioned Grok some time back. Any specific thoughts on like, what is happening in the private markets with these AI infrastructure companies that are doing work as such. And how are they differentiating versus like your large cloud providers that almost every startup uses?

Well, I mean I feel like it was a lot of what we talked about, which is that I think that there is a lot of innovation happening there. The parts that I get excited about are the founders that are really looking to concepts of physics and biology and looking into the natural world to draw inspiration of how we can do things differently and better and coming up with innovations. There's another one, liquid AI that was able to model the brain of a worm and and now can do what it used to take 50000 neurons. They could do in 19 neurons like a humongous breakthrough. But not just that, it's multiple factors less in neurons when it's simpler like that. If you just imagine looking at the worm's brain and you could see those 19 neurons now you have total explainability and visibility and transparency, which means no hallucinations. So in mission critical applications you wanted to know like what, what it did, why it did. It should be able to know why it did it and be able to explain itself. And, and also the other bonus of that specific example is that then it has learning now because it's a simple structure, easy for it to then learn and add to its understanding and you could see its learning happening. So, so I think those, those are the things that are happening that, that I'm paying attention to that I'm excited about.

Cool, cool, cool. I know we are at, at 45 minute mark. I want to be careful of people's times and, and more importantly George's time. This was a fascinating presentation. Lots and lots of logos and things that I need to digest and probably there are going to be more follow up questions on a couple slides that you kind of went through really quickly. But polling for any last minute questions for from the audience in case there are. Going once, going twice, going twice. I guess that's it. Let's break right here. George, I want to be careful of everyone's time, but this was fantastic. I think would love to revisit your thinking on AI in a few months. How because things are evolving very, very rapidly. So again, thank you everyone for joining. Any last words, George, any quick things that you would like?

Just if any of what I said resonated. If anybody wants to chat, I'm happy to. If they want to reach out, happy to chat and go deeper on any of this.

Okay, awesome. Again, thank you so much George and I hope everyone has a great rest of the day and weekend ahead.

Sounds great. Thanks for having me, Rohit.

Thank you.

All right.

Bye.

Bye.