Loading summary
Transcript1 lines
- [00:00]
A
OpenAI and Anthropic, the two biggest AI companies in the entire world, they both signed a multibillion dollar private equity deals and they did it on the exact same day. The structure of how those deals happened is different though, so I think that's going to be really interesting when it comes to what's going on with AI and enterprise. Also, Greg Brockman says that AI went from writing 20% to 80% of all OpenAI's code in one month. Saudi Arabia's Humane just launched on AWS and it is calling it the first enterprise grade agentic AI operating system. And Morgan Stanley just bumped the 2026 hyperscaler forecast to $805 billion. The first thing I want to break down though is Greg Brockman's. He went on a podcast recently and he had all these wild quotes, including talking about AI. He said, quote, even over the course of December we went from these agentic coding tools, writing 20% of your code, to writing 80% of your code. This is a 4x job jump in one single month. And he also said that OpenAI is 70 to 80% of the way to AGI. He also agrees with Sam and Demise over at Google DeepMind and says that we're maybe two breakthroughs away now. Of course, this wasn't everybody that agreed with him over on X. Yan Lecun predictably, has been dunking on the AGI claim. He's been basically doing this all week. You also have Andrej Karpathy, who was a lot more, I guess you could say, like measured. But basically the, the thing that I've, I've heard him talk a lot about is he says that sure, some of these coding writing percentages are like huge, right? Going from 20 to 80%. But what he said is if you measure tokens emitted versus what is so. And that's kind of where you're getting this 20 to 80% jump. But he says the difference is, and a harder question is what fraction of all of these tokens, all this code that's being admitted are, you know, going to be merge commits that a senior engineer would actually approve. And I think a big point of that is a of lot. A lot of this is getting written by AI, but a lot of it is still going to have to be edited or not approved or redone. And personally, I'm going to be honest, I am not a developer, but I've created a bunch of software apps over the last couple months using these AI tools. It's been so fun, it's super addictive. But basically everything I write, it's like I feel like I'm starting out with a giant block of granite and I'm chiseling it down to be what I want it to be because. And so it's not like, you know, hey Claude, make me a perfect statue. And it chisels everything. It's, you know, you, you give it a direction and then you say, ah, change this, change that, and you're refining the code, you're rewriting it because it, it has bugs and glitches and errors. And so anyways, I, I think people can be pessimistic about that process, but I think at the end of the day it's for people that aren't developers that don't know how to code. I mean, even for developers who are we joking? You can just go way faster. So I know people him and ha about the exact percentage and well, you know, it, it technically can't, you know, write all of your code perfectly. Like you have to go and edit it or, or monitor it or fix the bugs or rework it, like, okay, whatever, but that's life and you're still going a hundred times faster than if you're doing it from scratch. So in my opinion, it's exciting. The most recent project that I personally have been vibe coding, doing all of this is called AI Chat daily dot com. It's basically a news website that goes along with this podcast. So every time I drop a podcast episode, that website will go grab the transcripts from the episode and, and all of the news stories I talk about, it will write an article about them and it will put quotes from me in the podcast and all my opinions basically in there. It will also go and scrape 20 different news sources and get more facts and pieces of information and kind of embed those and has kind of bullet points in there. So if you hear me talk about a story here on the podcast and you're interested in getting a deeper dive into all of the facts in the background while also getting quotes, my opinion on it, you can go check out, um, the associated and usually I'm going to start putting them in links in the description if it's not already there. This is all part of the, this vibe coding project, so it should all be automated. You just get more information. It's kind of like show notes on steroids and it's a whole news site and it's based off of everything I'm saying. So I'm trying not to just make it a purely AI generated news site. It's it's kind of coming with my opinions and my quotes, my thoughts. But it also gives you more of a deep dive. So anyways, that's AI chat daily dot com. There's a bunch of cool stuff that I'll be working on that's probably my personal, you know, weekend vibe coding project that you'll see continue to improve over time. So I'll leave a link in the description if you want to go check out that project. But the next thing I wanted to talk about is that a company called Humane One is launching on aws. So this is coming out of Saudi Arabia and it's a PIF backed AI vehicle. The, the Crown Prince stood up last year and he kind of announced it. So now they have a partnership with AWS and they're calling it the first enterprise grade operating system for building, deploying and governing. Governing autonomous AI agents at scale. What the heck does that mean? I know many people are asking. Essentially there's three different parts to it. So there is H2O which is the development SDK and then there's humane code that's for actually like orchestrating so getting all these agents to work together. And then there's humane fabric which is the data ingestion and the governance layer. AWS is putting it into the aws marketplace across 39 regions. This is basically the next step on the $5 billion Saudi AWS partnership that was announced back in May of 2025. The reason why I think this is interesting, you know, of course they like, they have their big press release, but silver and AI is becoming a huge product category right now. If you listen to my most recent interview on AI Chat, I interviewed one of the co founders of Cohere and they are making a huge business, over a billion dollars raised specifically for sovereign AI. So they're helping companies and countries create AI, create AI models specifically that are their own. Like what does that mean? That means they don't have, you know, the company doesn't have an API to cohere. Cohere helps them run it, they help them maintain it, they help them upgrade it. But they take that model, they put it on their own hardware, they put it on their, you know, on prem at their own actual offices and they're running it. And if Cohere ever said, hey, you know, we hate you, you can't use their AI model, there's nothing Cohere can do. They have that AI model, they get to keep it. I think that for countries this is something that is going to be massive. And obviously Saudi Arabia is trying to tap into that. Next up we have Morgan Stanley that just raised its hyperscaler capex forecast. That feels like an absolute tongue twister. But this is for 2026. They've raised it to $805 billion. This is today. That's up from 765 billion before. So that is Microsoft, Amazon, Google, Meta and Oracle. All of those companies combined. These are some of the companies that are spending the most money, the most capex on actually building out these data centers. And so I mean the forecast for how much these companies are going to be spending is increasing. We're just seeing data centers accelerate right now. So for a little bit of context, Morgan Stanley analyst Ben Reitz specifically called out that 2027 spending could hit a $1.1 trillion mark which is absolutely eye watering and that would equal all non tech S&P 500 cap combined. That is insane. Insane. Also I think that's not something that we see very often. If ever they're just spending, you know, they're outspending like the entire market. But AI infrastructure is basically becoming as much, you know, capex as the rest of the, you know, entire corporate economy outside of tech. And while Morgan Stanley is doing that from a political side, we also had some comments from David Sachs which I think he's the White House aizar. I know he's, I don't maybe resigning or moving to a new role or a new post or something but that might not be happening yet. So he might still be the aizar. But anyways he's you know, friend of the pod over on the all in podcast and he recently was talking to Benzinga and he said that I could account for as much as 75% of US GDP growth over the next, you know, number of years, which is massive. Also, you know, kind of feels like he's talking his own book because he's in charge of AI and so makes him look good. If he, if he kind of says that it could be, you know, 75% of US GDP growth but personally I don't think that's that shocking. Even if we look at this from a historical spectrum perspective over the last few years, three to four years basically, let's say the last three years, if you look at where the growth is happening in the economy and you look at where the money is being spent, I mean just by and large based off of the last, you know, stat, we shared capital expenses, capital expenditures for some of these largest hyperscalers going into AI and data centers is where all of the money is, is coming from and going. And so I wouldn't really be shocked. Now he's kind of making the case that once the AI is here, the benefit to the US economy is, is going to be that is is going to be that kind of massive as well. That number immediately got punched at on X. Tyler Cohen kind of pushed back on, on David Sachs there. He's saying that the multiplier numbers that Sachs is using aren't very defensible. Dylan Patel at Semianlyst pointed out that even at the current pace, the chip supply just doesn't exist for the trillion dollar 2027 number TSMC apparently they're saying isn't built for that yet. The one thing I will say for people that are kind of commentating on this and being pundits on this, there's this post that will just forever live rent free in my mind, um, which was an expert, a quote unquote expert when Elon Musk announced what the plan for SpaceX it was and what their forecast was when they were first building it. And he basically, and they were talking specifically about the Starlink satellite network. And this experts, you know, basically broke down an incredibly convincing math and language breaking down how many satellites would have to be launched over the next. I think it was like 10. It was about seven years ago, over the next, you know, every year. And how they'd have to ramp up and how much bigger the starships would have to get and how much the Paylo and how basically it was physically impossible for SpaceX to build Starlink. It was like seven years ago. He broke down the numbers. It looks super, super convincing. And it was just kind of this big dunk. Everyone was like, yeah, Elon Musk is super dumb for thinking this is possible. Okay, well today we have Starlink. It's incredible and it's going to be an, it is an insanely valuable company. It's used all over the world, yada yada. They were able to achieve it. And to be fair, if you look back seven years, I mean the, the technological advancements they went through were really incredible. Like the, the rate of the rockets, the size of the rockets, the launch velocity, how many rockets they were launching a month. Like they really did hit their goals. They were able to scale up. It seems impossible and they figured out things that were quote unquote impossible along the way. But when you get these people that are kind of just saying like look, that's, that's basically impossible. There's no way physically possible it could happen. I Don't know. There's a lot of naysayers and this is kind of where a lot of the brightest minds, a lot of the, the most money is definitely going. So I wouldn't be shocked if we're able to achieve some of these numbers on economic growth and definitely on, on spending and data center build out. I don't think it's implausible at all. All right, let's get into what's going on with OpenAI and anthropic, because both of these companies have made announcements on the exact same day and they both are basically announcing that they're patterning, they're. They're partnering with private equity to push enterprise AI distribution. This is something that, I mean, basically these companies are neck and neck for the enterprise. Anthropica was kind of is basically like OpenAI feels like they're in an existential crisis because Anthropic is crushing it with enterprise adoption and OpenAI is trying to really get in there. So OpenAI has finalized a $10 billion joint venture called the Deployment Company. Anthropic announced a $1.5 billion venture with Blackstone and Goldman and Hellman and Friedman. And basically the structure is super new. Before we get into how they're actually setting these up, I wanted to mention if you are paying for multiple AI subscriptions like OpenAI and Anthropic, and maybe something like Gemini or Grok or 11 Labs for audio or any other image models or video models, I would love for you to try out my own startup, which is AI Box Dot AI. We have image models, audio models, video models, sort of VO from Google. We also have music models for generating music and we have a ton of other incredible things over there. Tools for helping you build and create by describing tools you want to build. And we have workflows, all sorts of awesome stuff. It's AI box AI. It is 8,99amonth. It's basically your digital workspace. You can get all of your projects done for one subscription in one place. Have to track around a bunch of different tabs. Personally, I love using Claude and Anthropic because the tone is so good for writing and helping me come up with ideas. But it doesn't have image or audio generation. So I find myself switching over to OpenAI for image generation. I use VO for Google for video generation. Anyways, all of that is in one place with AI Box and I would love for you to try it out and let me know what you think the link is in the description. Okay, let's get into OpenAI's deployment company and Anthropic's Blackstone deal. Both of these are private equity deals that were just announced today. They're very similar, a little bit different, but I think they are fascinating for what's going on in the industry. So essentially what's happening in my opinion is OpenAI, anthropic. They're both preparing for IPOs that could come this fall, so later this year. And right now they have a problem where, you know, OpenAI has got like, you know, millions of paying customers. Sure, but they're paying $20 a month. That's not incredibly profitable. Anthropic is doing much better where they have a lot of companies that are paying, you know, huge amounts of money to them, but the customer base is narrower. They want to expand to a lot of the kind of what we call mid market businesses. So not someone like Google or Microsoft who will, you know, spend millions and millions of dollars and not just all these small little people that are paying $20, kind of the middle people. That's where a lot of the money is today. And the way that they've gone about and done this is that they're, they're partnering with private equity companies because these private equity companies own a ton of different mid market businesses and they're partnering with them to get OpenAI into those businesses. So specifically TPG, Blackstone, Bain, all of these companies, they own hundreds of mid sized companies and TPG for just as an example, they own over or they have stakes in or they own over 2000 companies. All of those companies are, you know, basically customers that OpenAI would like to have. But if OpenAI wanted to go and get all of those 2,000 customers, they have to start up a sales team. They'd have to cold call them. They have to basically start, you know, working with all of those companies and the sales cycles for those companies takes a very long time. Plus those companies like, well, what do we want to use Anthropic instead? And there's all sorts of decisions that have to happen. If they work through a private equity company, then all of a sudden they can just say, hey look, you know, the private equity owners can say, hey look, we just partnered with OpenAI. You guys are all using Open, you know, OpenAI now instead of Anthropic or whatever else you were using before, or if you were using nothing before, like this is what we're all using. And on top of this deal, both Anthropic and OpenAI are going to copy the, the same strategy that Palantir did, which is that they will put an actual developer embedded inside of your organization. So if you don't have, you know, someone that's in charge of like AI change management, rollout inside of your company, OpenAI and Anthropic, they will send developers to your organization and they will look at all of your code, they'll look at how you do everything and they will roll everything out for you, which is obviously very sticky. It's basically, you know, the salespeople. The pitch is so much easier because it's not like, hey, we'll give you the software and you figure it out. And if they don't actually use it, then it doesn't help OpenAI because they're kind of tokens based, right? So they want to embed someone in there that's going to make sure this gets the maximum amount of usage. Now that's cool. And these kind of private equity partnerships, whatever, that doesn't seem like a big deal. But what is absolutely fascinating to me is how OpenAI has done this isn't just a clean cut partnership. They've actually started a brand new business which is called the deployment company. And this business, this deployment company is what is going to go and be deploying AI into all of these different private equity companies. And this deployment company is kind of like a vehicle because it's a joint venture. They have a 10, they have $10 billion that's been put in TPG, Brookfield, Bain, Advent. So all of like the private equity owners of all these companies are all investing money in. So they're like, hey, look, we're going to all invest a billion dollars each into this company that is going to go upscale our organization. So you know, technically they should be getting some value from it. But also those organizations are going to be paying for it, but they own them. So it's almost like a way for the PE group to like squeeze more money out of the, you know, out of their existing companies. I guess they could probably do that through distributions or something else if they wanted. The idea is probably that it's going to grow and also that company could go to other firms as well. So the kind of the founding PE companies are putting money in and this, this, you know, company could go other places as well. But in any case, $10 billion has put in, been put in total, a bunch of those companies I mentioned, they all put money in. There's 15 others. They all put in about $4 billion. OpenAI put in $1.5 billion. And then I think OpenAI has like the ability to put in another billion in the future. So Anthropic did a really similar deal, but it was smaller. They did a $1.5 billion total deal and Anthropics was happening with Blackstone, Goldman Sachs, Hellman and Friedman and General Atlantic. Basically the same idea. They're just trying to get into these kind of mid market companies. What's really weird in all of this, I think that maybe the most weird detail that no one's really talking about is that OpenAI is actually guaranteeing their private equity partners a 17.5% annual return. So they're like, look, you put in $4 billion into this company, we will guarantee you that you're going to make 17.5% annual return on that money from are AI deployments into your organizations like the token cost, the spend. Now what is interesting is that OpenAI is not kind of, this isn't like a new type of deal for OpenAI they've never done before. They're actually quite used to working with companies, partnering with companies to roll out their tech inside of enterprise and giving a kickback. The biggest was when Microsoft first put $10 billion into OpenAI. There was a deal where, you know, OpenAI went onto Microsoft Azure's platform and, and Microsoft was able to go and get a royalty from all of the sales happening there. And they got, you know, as part of their investment into OpenAI, they were just getting a certain percentage of OpenAI's revenue, period, all the money that OpenAI was making. So OpenAI has done these types of deals before. It looks like now they're making an entirely new company to do that. So they're a little bit more diversified and segmented. It's also going to be interesting because that company evidently is going to be making a lot of revenue. So when OpenAI says, hey, look, you know, we're looking to IPO, we also own, you know, the development company or the deployment company. And so they kind of have these different arms that make their IPOs look better. It feels like a lot of it is kind of, you know, working for the ipo. I think they also knew that maybe their companies didn't look totally IPO ready. They're trying to say, like, look, you know, if we can go sign up an extra 20,000 organizations that are using OpenAI for running all of their companies, that strengthens their position coming into the ipo. Okay, that is it for the show today. If you got something out of it, it would be incredibly helpful to me if you could leave a review or a comment over on Apple Podcast or Spotify on Apple Drop some Stars on Spotify. It's the about tab that you can hit and then you can you can leave some stars. I think you have to listen to at least three episodes on Spotify before you can leave a review, so make sure you if you've listened to three episodes, you're eligible. It would help me out a ton. And make sure to go check out AI box AI if you want to get access to over 80 different AI models and in one place, it is $8.99 a month. I hope it saves you a ton of time. Money consolidates everything for you and is just making your life with AI a piece of cake. All right, see you guys all in the next episode.