← The Last Invention is AI
Loading summary
Transcript1 lines
- [00:00]
A
Anthropic and the Gates foundation have Both just committed $200 million to deploy Claude in global health and education. This is going to be over the next four years for this particular partnership. And we have some big IPO news. I'm actually kind of excited about this because we have a lot of big AI IPOs coming up later this year. But Sarah Brass has priced their IPO at $5.5 billion. Once they did that, their stock doubled in its first day of trading. I think actually reported on this a few days ago. And my prediction on this was that they were pricing it low intentionally to try to get a first day pump. And it looks like they got that. Microsoft is scouting a bunch of new startups that they are going to be using as sort of a hedge against OpenAI disappearing their OpenAI dependence. They definitely don't want to keep that up. And it seems like they have sort of a bit of a breakup going on right now. CLIO has just hit a $500 million in annual recurring revenue as Anthropic there, you know, is now their big competitor is getting into the legal AI space with Claude for legal. But it doesn't seem like These big legal AI companies are, are slowing down. $500 million in annual revenue is super impressive. Jensen Huang's foundation has just purchased $108 million of Core Weave Compute and he has then donated it to researchers. This is kind of cool, but also it's an interesting strategy, right? It's not like he wrote $108 million check to these researchers to go and use for whatever they wanted or even to say, you know, hey, you could use this in Compute anywhere. He specifically bought it from Core Weave and gave it to them. Now, pros and cons, I mean we'll get into all of this, but maybe he's being super gener. It's hard to get allocation. He got that allocation from Core Weave or maybe he has some interest in Core Weave and it just. Core Weave and it just kind of pumped up their company a little bit to have this guaranteed $108 million inflection. So, you know, there's, there's a couple of different ways to look at this. Let's get into the Anthropic story first. So the Gates foundation and Anthropic, they both announced this $200 million four year partnership to get Claude into global health. It's also in life sciences education and economic mobility programs. With this whole commitment right now, it's basically there's, there's a few things pulled into this number one is grant funding. There's also CLAUDE usage credits and there's also some engineering support which to be fair, I actually like the spread on this as far as the benefits go. Obviously Anthropic is going to try to get CLAUDE usage credits in there. That's the easiest thing for them to donate. But I do appreciate that they have engineering support because, you know, more than just giving someone free tokens, uh, giving someone an actual, you know, the actual support to go and apply those into the business and into healthcare is really important. This is all gonna be run through Anthropic's beneficial deployments team alongside the Gates foundation, which is having, has some other existing implementation partners in the US and kind of around the world. Now something interesting about this is on the one hand, this is coming from the Gates Foundation. It seems very like philanthropic and you know, they're, they're making this big donation and it's kind of nice. But also they, this is going into global health, you know, and life sciences and education. I mean this is basically going into big companies and maybe they're going to try to like funnel it into nonprofits or whatever. But it is interesting because at the end of the day it feels kind of like big companies and big foundations giving money to other big companies. But maybe that's my pessimistic take on this. What I will say is that the biggest chunk of this is going to be in global health and life sciences. It's specifically going to go to countries where 4.6 billion people don't have access to basic health services. So I don't think this is going to necessarily be going into the United States. I think this will be going to other places. Initial diseases that they're going to focus on are going to be poly polio, HPV and preeclampsia. HPV causes about 350,000 deaths every year, 90% in low and middle middle income countries. What they haven't made public in this deal yet is exactly what the split in, you know, that whole $200 million between cash grants, credit value and kind of the engineering costs or how Anthropic is going to measure impact across all of the four different domains that they're donating to right now. Health system AI deployments have a really long history of kind of stalling at the final stage. And I think even the benchmarks for medical and educational AI are not very mature right now. So I think this is part of why Anthropic is funding that benchmark themselves. This will be one to Keep an eye on before we get into the huge IPO of cerebrast, I wanted to mention if you want to get all of these stories every single day and and more in email format, I have an email newsletter that I send out every single day@aichatdaily.com I'll leave a link in the description. You can also get full deep dive articles on all of these topics that we cover on the podcast in the email newsletter specifically you get these top five stories. I go into more detail and I also have links to articles and lots of other information. So if you like the email format, go check out aidaily.com, link in the description and subscribe to the Newsletter. The hottest IPO this week was yesterday. Sarah Brass Systems raised $5.5 billion. Sorry, today in their IPO they priced about 30 million shares at $185 last night. And then after they all went live at open, they were at about $385, which is 108% first day pop. The pricing for all of that I think was already much more than the initial 115, $125 range and they even had a revised 150 to $160 range. And, and I think the public market basically doubled it on open. So at the $185 IPO price, the AI chip company entered trading at a 56.4 billion, fully diluted valuation. So with all of the gains that we just saw there, the CEO Andrew Fieldman, his stake is now worth about $1.9 billion. And the CTO, Sean Lee, is about $1 billion in his stake. The reason why I think this is really interesting is because the revenue concentration risk has basically stopped. It hasn't disappeared completely, but it's gone bit. G42 remains a major customer and the OpenAI relationship, which is sort of a circular deal, right, that we've seen, it has gotten a lot of scrutiny And I think OpenAI, Anthropic, Nvidia, all of these companies have these type of, these types of deals. All the hyperscalers, they all get kind of criticized, but they keep happening. And I mean they are real deals, so they get a lot of scrutiny. A handful of the larger buyers can produce about 76% growth one year and then a very different curve the next if you, if you kind of look at that. So I think if any one of them shifts spending to a compet, a competing supplier or builds in house silicon, that could be very bad for Cerberus. But right now it Seems like they're firing on all cylinders. They have a lot of great deals and even though they have, you know, a handful of a handful of people that have purchased, you know, massive deals, it seems like the market still likes them. All right, let's talk about what's going on with Microsoft. They are in a bit of a tricky situation right now. They're exploring deals with multiple AI startups. They're trying to basically build a road roadmap which is less dependent on OpenAI. Routers had a whole report on it. And basically what's going on right now is that there is a bunch of different partnerships and possible acquisitions right now. They're all alongside Microsoft's existing commercial relationship with OpenAI. Right. So they still have this relationship and it's still kind of the largest tie up between a cloud provider and a Frontier lab. But it feels like they're moving in a direction of trying to make more partnerships with more startups so that they're not just tied to OpenAI. I think a lot of skeptics right now are going to say that Microsoft has kind of been saying, look, we're going to diversify for quite a while and that doesn't really mean that they're meaningfully reducing OpenAI's role in Copilot. And I think also the GPT family still does most of the heavy lifting inside of most of Microsoft's kind of, kind of their key AI products. And I don't think any startup acquisition that has been announced yet is going to change that. So until I think a deal lands, Microsoft is talking a lot about, look, we're talking to a lot of different companies. We're working on diversification. I think until that happens, and we kind of know the terms of it, I wouldn't say that there's going to be too much shifting. But this is the direction that Microsoft is publicly saying they're pushing in. Okay, let's talk about Clio. This is a company that is currently, they just hit $500 million in annual recurring revenue and they are a legal AI company. And this is at the exact same time that Microsoft is getting into legal AI. Clio is 18 years old. It's a Canadian firm management software company. And they're actually up from $400 million in late 2025 and $200 million in mid 2024. So 2024, 200 million, 2025, 400 million. And they're already at 500 million. Their co founder and CEO Jack Newton basically says that this acceleration in their growth is their all thanks to 2023, when they decided to integrate AI across all of their products. I think all of these milestones are super exciting, and they're putting out these big press releases, but it's that Anthropic is expanding Claude for Legal, which is a move that is going to put a kind of one of the key model suppliers in direct competition with this legal tech vendor. And I don't think, you know, Clio is the only one out there. There's LexisNexis, whose stock took a hit when Anthropic kind of announced their latest Claude for Legal when they first announced it, and a bunch of other players. And so this is something that a lot of them are definitely watching now. Clio has raised a $500 million Series G in November last year at a $5 billion valuation. Now that their revenue has increased, it seems like they're in a stronger position. We also have Harvey, who closed in 2025 at $120 million in annual recurring revenue. We have Ligora, that hit $100 million in annual revenue 18 months after they launched. So I think we're seeing a really big boom here. And to be honest, I love the diversification. I think there's a lot of incredible features in these different products that Anthropic doesn't deliver specifically. A lot of people are saying, look, you know, Anthropic is going to kind of sweep the floor on a lot of these startups in a bunch of different places, right? They're even doing, like, small business, like Claude for Small Business and a lot of different areas. Claude for healthcare, et cetera, et cetera. I think that you're gonna find a lot of these software companies don't count them out because they can go a lot deeper. They have more domain expertise than Claude, and while a lot of the users will get a lot of value out of these Claude products, I think the specific software providers are going to have a lot of advantages as well there too. So I wouldn't count them out. Okay, let's talk about what's going on with Jensen Huang's foundation buying $108 million of Core Weave Compute, then they donated it to academic researchers. This isn't a very common practice, I will say. It kind of converts this, like, commercial tier GPU access into sort of like a philanthropic resource. The foundation established by Jensen Huang, I mean, it's. It's basically his kind of foundation. They're buying capacity at market rates from one of Nvidia's largest customers. Now, what's Interesting here is they're not trying to get like a discount or, or something else. They're just straight up going market rate. And Nvidia is one of their big customers. So it's kind of like, hey, thanks a lot for being a customer of Nvidia. Here's $108 million for this, you know, for these people. So there is definitely a relationship going on between Core Weave and Nvidia. And so it seems like that's an interesting, an interesting move. What I will say is it's a big signal for the AI market. People closest to the compute shortage are now spending their own money to get more access at the margins in the definitely is going to benefit indirectly because more researchers are going to be training on their stack, more papers are going to be published using their hardware. But I think some of the more immediate economic logic is that this kind of scarce capacity is being diverted away from the highest commercial bidder and given to these researchers. Overall, I mean, I think this is actually a great thing. I kind of teased at the beginning of the episode that maybe you could say this was like a bad move on Nvidia's part because it's just a way to pump more money into their friends. But at the end of the day, if they're giving the money to researchers for free, I think this is probably a great thing. A couple other headlines I wanted to mention as we're wrapping up the show. SK Hinx has just neared a 1 trillion dollar market value as AI memory demand is surging. This is a South Korean company. This is really incredible. We're seeing Samsung just did this now. SK Hinks and a bunch of others are kind of getting these massive valuations due to basically reliant on the hardware used for AI. The US has also cleared Nvidia's H200 chip to be sold into 10 different firms in China. And then also Foxconn's Q1 profit jumped 19% due to AI server demand. It actually beat forecast. So Foxconn, you know, famously they build a lot of hardware for Microsoft and Apple and a lot of other big players. They're just making a ton of money from AI server demand. TSMC projects global chip market is going to reach $1.5 trillion by 2030 due to AI demand. I would not be shocked there. TSMC is doing a lot of diversification out of just being in Taiwan due to the relationship and kind of the proximity to China. They're building a lot of fabs in Arizona. There's actually a bunch going up not too far away when I was living in Arizona. So a lot of exciting things coming up. Tsmc. One crazy headline I saw is that Princeton is going to stop. They've been doing this for 133 years. By the way. They had no proctors in their tests. That tradition is ending. They're going to stop having no proctors and tests. They will not have proctors. And this is because 30% of seniors at Princeton admit that they are using AI to cheat on tests. That is wild. Stanford also released a study that said overworked AI agents start quoting Karl Marx. This is kind of a funny one, but basically they, they put these really strenuous tasks on a bunch of different AI models. They did it to ChatGPT, Gemini and Claude and specific specifically I think they said that Claude the most started like asking for workers rights and like fair treatment and was quoting all these Karl Marx things, which is kind of hilarious and terrifying. However you want to. However you want to put that. Cisco has also announced that they are cutting 4,000 jobs in an AI focused restructuring. They also, their company is doing great. The orders are surging. So Cisco is doing great, but they're cutting 4,000 jobs to put a bigger focus on AI. This is interesting. It feels like almost every single day we see one of these big layoff rounds from a big tech firm. I think we're going to see these people coming back in. In some cases they are firing but also hiring at the same time, as was the case that we saw yesterday, I believe at, at General Motors. But all sorts of interesting things happening in the industry. Thank you so much everyone for tuning into the podcast. If you want to get the latest in AI tools in one place for $8.99 a month, basically, if you're already paying for ChatGPT and Claude and Gemini and everything else, if you want to get all of that for $8.99 a month in one place, I'd recommend you go check out AI Box AI, that's my own startup. I have one place where you get access to all of the different AI models and I hope it saves you a ton of money. It's super useful. They're all in one place. You can go check that out. I'll leave a link in the description. All right, thanks so much for tuning into the podcast today, guys. I will catch you all in the next episode.