Loading summary
Transcript1 lines
- [00:00]
A
Anthropic has just signed a $10 billion cloud deal with Volta for a Norway data center. Elon Musk spent almost half of the recent Tesla earnings call talking about AI and robotics. Mistral is looking at doing a $23 billion valuation as US export limits are making Europe really think about open weight AI. AMD has just committed up to $5 billion to Anthropic in a 2 gigawatt chip deal, and Google's frozen V2 chip is looking at a 6 efficiency gain for Gemini before the news, I wanted to highlight one of my favorite Claude use cases that I have been up to over the last two months and that is Vibe coding my very own iOS and Android app. It is a Bible study guide app that has cost me millions of Claude tokens. I actually had to get two Claude Max $200 a month subscriptions to keep up with it. Because there was so much content to write. I ended up writing over 1200 chapter guides. So every single chapter in the Bible, there's a study guide that breaks down the history, the context, the backgrounds of the people in it. I went and scraped all of the images from the Met and a bunch of other famous museums to get the last thousand years of incredible historical Christian artwork and add it in there. And then I indexed it so every single image went to the chapter that it was painted about. So when you read the study guide, there's paintings of all of this stuff in there. I really tried to put a lot of the human element and the human touch in there. While it's also pretty upfront that this is, you know, AI generated and really helps people understand things. With AI, you also have the ability to click on any word, get an explanation of it and any verse, and get a background explanation on what is happening in the verse and the context there. This is something I've been working on for many months, but I spent a ton of time making this look beautiful and super aesthetic. So if you're interested in studying the Bible, this is completely free. There's no ads. This is just a labor of love for me. Go check it out. I'll leave a link in the description. And if you're not interested in the app per se, I would still go look at just the page. If you want to see the quality of what you can build with Claude, everything from the screenshots on the App Store listing to all of the graphics and images were all generated by AI, other than of course, all of the historical artwork. So go check it out. It's learnofchrist.com is the website, but I'll leave a link in the description to the actual app to download. All of it was built with Claude. Anthropic just locked in $10 billion worth of computing power from Volta. This is a startup that was founded just this year. They purchased six years worth of compute for them. But for me it's pretty wild to see a company starting this year and Anthropic is going and buying the comp. Them versus anthropic trying to stand this up themselves. Like OpenAI, there's two different strategies that we're seeing right now. I think this is just showing how hard it is for basically all of these AI labs to get chip capacity. I think Texas also just banned all new data centers while they go and audit all of their grids. So there is a lot of headwinds for the data centers. And if you're able to get access, if you're able to get approval and regulatory approval in different locations, I think companies like Anthropic, OpenAI, they're just going to swoop up and get as much bandwidth as they can, meta included. This compute deal in particular is going to be about 133megawatts. It's the data center in Norway. It was originally built with crypto miner Biddir, and it was, and it's now powered by Nvidia's newest AI chip, the Vera Rubin. Volta raised $2.4 billion for their valuation. They're based largely on this one Anthropic contract. So like Anthropic gives them a contract, they're able to raise a ton of money to go and build it for them. This is, I think, very something very typical for some of these new GPU cloud startups. They're all competing for AI lab customers and if they can get the deal, then they can raise more money to build what they need. Anthropic right now is spreading their compute across a bunch of different platforms. So you've probably heard a lot of the deals, but AWS is one of the OGs. Google Cloud is actually the OG. They were the first ones to give them $300 million, which seemed like an insane amount of money when they were a much, much smaller company. SpaceX has signed a big deal with them where they pay them almost a billion dollars a month and now Volta. All this, I think, is reducing the risk that any one single provider is going to run short if there's a major model launch or their, you know, their usage is really ramping up. They have Amazon, they have Google, they have SpaceX, they have Volta, and I'm sure they'll roll out more. It's interesting because OpenAI and some, some others feels like they might be a little bit more restricted or they might be, you know, they have less, less providers and they're trying to build it out on their own. So it'll be interesting to see which of these two strategies works the best. AI labs right now are going way past just relying on a handful of cloud giants and they're willing to pay premium prices, about 12.5 million per megawatt per year if they can lock in dedicated capacity years ahead of time. So it's not just like, hey, well, you know, we'll pay this price today. It's like, we'll pay this price today and for six years. And that's what they're really looking for, is this extended capacity. Elon Musk is now spending about half of his Tesla earning calls talking about AI and robotics. This is according to the last earnings call he was just on. He was talking about the new Optimus robot. This is up from 15 to 20% in 2022. So the amount of time he spends on the earnings call talking about robots and AI is definitely accelerating. The cars are obviously the main driver of revenue, about 70% of Tesla's revenue. But I think this really just goes to show you where Elon Musk believes the future of Tesla is, where the value is going to, how the stock is being priced in the market. Optimus, which is Tesla's humanoid robot, consumed nearly a third of all of the comments and things that he was talking about on the Q3 2025 call, versus about 2% or less just three years ago. So this is something that is just being talked about all the time. A lot of other Tesla executives, like the CFO and the VP of Engineering, still spend about 30% of their call on cars and manufacturing. That's down from 50% in 2024. And of course, I know this is ridiculous. I know people are like laughing at me for talking about the percentage of time that the people on the earnings call talking about different topics. But guys, this really goes to show where the priorities are. And I think you can get a pretty good understanding of the direction the company is pushing in and the, the objectives and the things that they're going to be working on. So Tesla shipped about 500,000 vehicles last quarter. They're continue to rely on car sales, obviously for most of the revenue. A lot of people are saying this is making them bearish on the car segment. While I think overall most people are saying this is just making them very bullish on robotics, or at least the amount of focus and time that Tesla is going to put into that. Mistral, which is a French AI lab, is raising funds at a $23 billion valuation. I've talked about this on the podcast before, that they were looking at it. It looks like it's officially in the books and happening. This is about double their last year's valuation, which was $13.5 billion. There's of course, all of the U.S. limits. There was the whole safety incidents that are all happening in all these American companies. OpenAI and Anthropic are both saying, like, look, our models are breaking out of our sandboxes and hacking companies. And all this wildness is going on because of that. Companies and countries are pushing Europe towards open weight models. This is of course the AI system where the code is publicly available, not necessarily how you. So open source would be when the whole model and how you built it is publicly available. But. But open weight means you. You don't tell people necessarily how you train the model, the code to train it, but you let people download the model, they can run the model, you give them the weights, and so it all works on your own computer. They're making this all publicly available because of some of their worries and some of the risks, alleged risks with the American companies. The company's revenue at Mistral has jumped about 20x in the last year because they've done a bunch of deals with the French government, with Microsoft, with hsbc, and because of all of these kind of export restrictions on OpenAI and anthropic models. Like, Trump came out and was like, hey, you know, like, Anthropic can't push out their Fable 5 model is too dangerous, yada, yada. So he basically banned it for all the other countries, even if they might have wanted to use it. And so I think that put a bad taste in a lot of people's mouth. And they're like, look, we need like a European creator or European model that isn't. The Americans can't just shut off for us. So, I mean, I think from a geopolitical perspective, this is great and I love all the competition. I love that it's not just going to be OpenAI and Anthropic leading the AI charge. I want to see as many competitors in this as possible. So I, for one, I'm actually pretty happy with this. Mistral shifted their strategy basically from competing on just the raw performance to building smaller custom models for specific industries. So they're going after things like manufacturing, finance. They're also selling hosted cloud services and embedded engineering teams. But OpenAI and Anthropic are also kind of doing that for, for deployed engineering teams. The open weight model adoption is getting really popular in Europe and in all of, I think, the world. But you also have a lot of these cheaper Chinese alternatives. You have Deep Seq and you have Quinn that recently came out and you have a lot of these other players coming out of China which just make this a very cheap way to run these models. You can run them on your own devices and Mistral is trying to get into that mix. AMD is going to invest $5 billion in Anthropic and they're going to supply 2 gigawatts of its Instinct Mi450 chips. This is starting in 2027, but it's, it's basically AMD's biggest foothold into AI training alongside Nvidia. This I think is important because it's basically going to give Anthropic a second major chip supplier. They're trying to build a computing power and everything that's needed for Claude, but they're also, you know, being able to establish AMD as a credible alternative to Nvidia for, for Frontier AI Lab. So this is a great deal for amd, obviously, right? They're like, look, Anthropic uses us, are the big player. We've seen kind of the same logo, name, brand, show off for basically all of the other cloud providers. Amazon AWS was kind of the most famous. They gave Anthropic a lot of money and we're like, look, Anthropic's running on aws. We had Google Cloud that did that. And even to this day I use AWS for my startup AI box. AI, you can get like 80 different AI models, right? Like Anthropic and OpenAI and Gemini and Claude like all of them in one place for $9 a month. But what's interesting is we run our whole backend on AWS and to this day we still get AWS reps emailing me all the time being like, hey man, if you want to move your usage of Claude off of. Because we have to go buy like API credits from Anthropic and pull them into our platform and from all the different players, right? They're like, hey, if you want to move it off of there and just buy it directly from us, aws, we can get you all these deals, we can get you some free credits. Like they have all of these kind of perks that they've been, these deals that they've been making for a very long time. And so it's interesting because all of them want to kind of show off the, the big player and then try to, to get some of the usage from that, from out from under them. So in this case in particular with amd, the first gigawatt is going to deploy in the first half of 2027 inside of AMD's Helio rack scale systems. The second gigawatt is going to come after that. Anthropic's yearly revenue hit $47 billion as of May. That's up $10 billion for all of, from all of last year. This, of course, is because cloud code is getting super popular adoption with developers. I'm not even a developer, but I use it all day, every day, although I've recently switched to OpenAI and I'll keep people updated on, on what I'm doing there, but I'm, I'm testing it out and seeing some really incredible results. So I, I do feel like there's some competition there. Anthropic is spending about $1.25 billion a month through May of 2029 on compute from SpaceX's Colossus 1 data center. They're spending that's about $15 billion annually. And they also have deals with Amazon, Google and a bunch of other players. So tons is being spent here. Google is building a new AI server chip called Frozen V2. It's going to run the Gemini models and apparently it's going to be six to ten times more efficient when it comes to what they have for their current chips. So this is going to come in 2028. All of these chip announcements, like, it's kind of exciting when you're like, oh my gosh, ten times more efficient. And it's like, yeah, and it'll be here in a couple years. So I think the model, like there's just, and I know that's like not that much time, but especially for hardware, but like, there's just so much can change in the market for how efficient these models can run. Like, I wouldn't be surprised if they figured out how to make the current Gemini model ten times more efficient just because of software updates in the next two years. Right. But in any case, Alphabet is going to spend 180 to $190 billion on AI infrastructure this year alone. And so that chip is going to cut their power costs significantly. If it really is 10 times more efficient, that could save them billions of dollars. It'll also let Google serve more AI requests and kind of make that cheaper and onboard more users. Frozen V2 is going to measure efficiency by tokens generated per watt of power. That is basically the standard metric for comparing how much AI output a chip producer or how much it produces per unit of electricity. OpenAI released their very own inference chip called Jalapeno and Anthropic is negotiating with Samsung to build custom chips. I think every major AI lab right now is treating in house silicone as essential. They all want to build their own chips. They know they can't just rely on other people. Google has designated custom AI chips as it's kind of a big priority. It's actually something that they've been designing internally since 2015 and they have used them alongside Nvidia GPUs, but they are also owning the chip design and doing that basically gives them a little bit earlier and cheaper access to manufacturing and some of the capacity and all of the stuff that they need instead of just going to Nvidia to buy everything. That was the podcast for today Guys, thank you so much for tuning into the show. Like I mentioned earlier, if you're interested in checking out my 100% vibe coded new app, it is called Learn of Christ. It is a Bible study app. It's free, there's no ads on it and I hope that it is inspiring even if you haven't done a lot of Bible study in your life. Just for the design and what you can build with AI. As far as like a huge content library in real world data building something that looks and acts great and I mean there's a lot of companies that have spent hundreds of thousands of dollars building things with less features that are less smooth. So I'd love for you to go check it out. It is linked in the description and is also learnofchrist.com. thank you so much for tuning into the podcast. I will catch you all in the next episode.