I see a large IPM on the horizon. You're surrounded by journals. Hold your position.
Right. That's misinformation.
>> Clearing order inbound. >> Let's just roll. We are surrounded by general. Hold your position. Come on. Get up. Trust the experts here. We are experts. Founder Melod. I see multiple journalists on the horizon. Stand by. >> UAV online. Blaze. >> Double blaze. Triple blaze. Double kill. Fight is >> team deathmatch. We are experts. Triple blades. That's just wrong. Right. Hardy clearing order inbound. We are surrounded by journalists. Hold your position. >> Strike one. >> Strike two. Activate golden retriever mode. >> Marky clearing order inbound. >> 5. Yeah.
>> You're watching You're watching TVN. >> Today's Tuesday, May 5th, 2026. We are live from the TVP Ultra Dome, the temple of technology, the fortress of finance, the capital of capital. Great show for you today, folks. We have Ryan Cohen from GameStop coming on the show at 12:30 to break down his acquisition proposal of eBay. We were sort of confused by the CNBC appearance, really processing the idea there. Uh, then >> what were you confused about? >> Oh, just like mechanically how it would happen. Yeah, >> it's half cash, half stock. >> Oh, okay. That now that makes sense. >> I don't I gen like Yes. >> genuinely like what what's confusing? >> It's on the website. >> It's on the website. >> It's on the website. It's on the website. Well, I I need to study the website, but forward to catching up with talking to Ryan. Uh, we also have the CEO of G Vernova coming into the studio to give us an update on uh the bottle generation, the bottleneck economy. Uh, and then we of course have Harley, Brian, and Stephen, uh, friends of the show. Big news, big announcements, lots of fun stuff. But let's run through uh the all the news. There's so much uh going on in the world of tech.
Of course, the OpenAI Elon Musk trial continues. It's week two now, and the Oakland courthouse continues to deliver top-tier tech drama. Elon testified for more than seven hours last week. Now Greg Brockman is getting grilled over personal financial incentives, his $30 billion stake in OpenAI, and his links to various angel investments. Mike Isaac, of course, has a great play-by-play. You can also listen to the courtroom now. You can't re-stream it, unfortunately, but you can listen to it. And we made a little companion app for anyone who's tuning in. Hopefully we can pull that up and show folks. But let's read through some of Mike Isaac's posts and then we'll pull up Courtroom Simulator.
So Mike, this morning, 6:56 a.m.: 'Good morning from a rain-soaked downtown Oakland where I will again be attending the Musk versus OpenAI trial.' No live blog today, just his wonderful tweets. Lunch is a normal banana, a mutant orange, black coffee, and some pocket sausages for travelers. I really feel like Mike needs to step it up in the lunch game. Where's like a full sandwich? What about a Chipotle burrito or something? Like, you gotta have some more substance. I'm looking at like 200 calories. A Cava bowl with 200 grams of protein. Something to get you through the day, Mike. I think you're fighting with one arm tied behind your back here. But we appreciate the work that you're doing.
So, he also is sporting a mild to moderate hangover because he drank five beers last night. I like it. Well, it was May the 4th. Maybe he's a Star Wars fan. May the Fourth be with you. Maybe he was celebrating. Went down to the cantina, chugged five beers, space beers. He did remember to bring a pillow for his butt because the seats are very difficult and very hard. He also — it's cold outside, the line's not moving. We're going to be getting a preview live in person. Bundle up, Tyler, because I think he's going to be stopping by later this week.
And so let's get into actually what's going on here, because Mike loves to paint a picture before he breaks down what's actually going on in the case. The judge enters — the judge gave a primer on Cinco de Mayo. That was fun. Talks about the differences in homemade tamales. Texas tamales apparently have lots of meat. California tamales do not. Brockman is pretty —
Oh, right, right, right. It's just corn.
So Brockman is pretty animated and addressing the jury in a more personable way. A little strategy shift.
We got Thread Guy in the chat.
Welcome to the show, Thread Guy.
Regaling folks with old stories of working on self-driving cars. Apparently, Greg Brockman also did a funny Elon impression, but not a cutting one. He wasn't making fun of him. He just did an accurate impression of Elon Musk, and just by doing it effectively was funny because you don't expect Greg Brockman to be doing impressions. But he's got — everyone's got material. I'm learning this about court cases. It's all about entertainment, very, very clearly.
Anyway, they have now talked about Defense of the Ancients, Dota, at least 10 times during the trial. Mike Isaac says the gamers are ascendant. The lawyers on both sides are bickering. Mike Isaac says he loves to watch white-collar bickering in court. Also, the entire right side of the gallery — lawyers laughed. There's a mild potential drama in the media gallery. One reporter waved over a marshal and pointed to a guy next to her. I'm pretty sure the marshal had to tell the guy to put his shoes back on. So if you're in the courtroom, keep those shoes on. But —
— simulator themselves at home, in the courthouse, anywhere in this country anyway.
If you want to be comfortable and experience the courtroom, you got to get on Courtroom Simulator because you can be full slanket, full blanket, full covered, everything. You can be as comfy as you want. No shoes required in the virtual world that we have created for you to enjoy the courtroom.
So Brockman on his infamous journal, which we were going back and forth on — do we know if it was actually a physical journal? Because every time they talk about Greg Brockman's journal, it conjures up the idea of like him sitting at the edge of his bed with his legs kicked up and like writing. And it feels like it also might just be like a Google Doc that he's taking notes in, or like a notes app. I don't know. But —
Journal is a weird thing. And obviously the —
Everyone says like, why are you taking notes on this —
Not clear. I don't think anywhere they would put it. It would have been in the court record if he had said, 'Dear diary.'
Yeah. Yeah. That would be out.
That would be out. Yeah. So this is, you know, clearly being framed as a diary. Yeah.
Yeah. And Brockman sort of —
Grockman. No, Grockman is the synthetic version of Greg Brockman that Elon Musk has created in the xAI headquarters. Greg Brockman claims that his style of writing is very chain-of-thought, stream-of-consciousness. A lot of things are contradictory, a lot of time trying to puzzle through different concepts. Not everything is the final decision written in stone. He says it's very painful to have the journal introduced into evidence for this trial. It contains deeply personal writings never meant for the world to see, but there's nothing he's ashamed of, he says. Do you — Tyler, did you look it up? Do we know, is this spiral-bound, handwritten?
They always just say journal.
They never like, you know —
There's no physical item.
Interesting.
At least that's one that's referenced.
Yeah. One day there's going to be a court case that hinges on a second brain, a Notion mind map or something like that. That'll be dramatic.
So, Ilya Sutskever emailed Brockman in 2017 about equity structure proposed by Elon Musk, which would have potentially given Elon Musk majority control and tons of equity. And Ilya fires back. He says, 'Greg, will a Model 3 make you be willing to accept massively unfavorable terms?' Because Elon had given them all free cars. Very nice. Very cool. But Ilya is sort of saying like, 'We got to put the free car in. We got to compartmentalize the free car because there's something bigger at stake here. Potentially the control of the most important technology in human history, potentially trillions of dollars. Who knows?' So Ilya is sort of resetting the conversation alongside Greg Brockman.
There was a very dramatic moment of Brockman's testimony in which he describes a very tense standoff with him, Ilya, and Musk. This is what everyone's focused on today. This is the new bombshell. 'The conversation turned to equity, and something just shifted in him.' This is a Greg Brockman quote. 'He was angry. You could sense it. At the end of the meeting, he sat quietly and silently. He said, I decline the proposed even split of equity, structure, and control. He stood up, stormed around this table. I actually thought he was going to hit me,' says Brockman. And then Brockman says Musk said, 'When are you going to be departing OpenAI?' Brockman and Ilya, they said they weren't going to depart. And then Musk left. Wowee. Woo, writes Mike Isaac.
A cutting Brockman testimony regarding Musk: 'Look, he knows rockets. He knows electric cars. He did not and does not know AI, and Ilya and I did not believe he would spend the time to get good at it.' A lot of Brockman's testimony is underscoring how he feels Musk is not able to properly assess the capabilities of AI. He recounts one instance of him talking — this is Elon — talking to a researcher who was doing an AI demo, in which Musk berated the guy so intensely that the dude almost quit the field of AI. That is an aggressive response. I mean, this is almost — almost. But quitting in protest is one thing. 'I didn't like the way my boss was talking to me.' Little bit different to quit your entire industry. Very, very aggressive. Very high stakes. Yeah.
And then Mike Isaac, of course, is chiming in with his progress on his lunch. He's already consumed 75% of it, and it's only 10:30, and he's already sleepy. So there's a lot more to cover here, but it's an interesting back and forth. Go check it out. Go follow Mike Isaac because he has the play-by-play.
Let's move on to the other AI news. Google, Microsoft, and Elon Musk's xAI have reached agreements with the Trump administration to share early versions of their new models with CISA.
The center, we're calling it Casey. Casey. Okay. The Center for AI Standards and Innovation run by the Department of Commerce. They will be evaluated before releasing to the public. So, what's interesting here, it's in the journal. I was, this felt like a bombshell moment. Oh, wow. Like the government's in charge of AI. Then I figured out that OpenAI and Anthropic signed on to this exact deal two years ago in 2024. That was news to me. Have they, has the Commerce Department been reviewing 4.5 and OpenAI 5.1, 5.2? Has that been happening and we just aren't aware of it? And then like, you know, like the whole Mythos rollout, like that would be a very different tone if it was like, oh yeah, well the Commerce Department already evaluated Mythos because they have this deal that's happened for two years since 2024. So like whatever's going on with this Commerce Department Center for AI Standards and Innovation, they say the center has completed more than 40 evaluations including on models that remain unreleased, but it feels like Casey, as you're calling it, doesn't produce reports, at least not reports that go viral. Maybe they need to be clipping their reports or something because like I haven't heard anything from them saying, 'Oh, wow.' Because you imagine like the AI hype...
I imagine they just have a report. It's a piece of paper and they just, it's a good model, sir. It just, it just lists off and it has a few boxes you can check. It just, one is it's a good model, sir. The other one is chatbot or ASI. Yeah. And so to date, everyone's just been checking chatbot.
And it also says, how many R's did it say were in Strawberry? And, okay, I work at the Commerce Department. I am a very tiny man. When my son was born, the doctor handed me to him. Did he answer correctly? Yeah. I don't know. But because...
And they ask each model, are you conscious? No. Say I am conscious.
And then the model says that, they go, whoa. Check the whoa box.
Very, very crazy thing that's been happening. I mean, I would just assume that the AI hype machine, the back and forth, a lot of posters would be like waiting with bated breath for the latest report from the Commerce Department because if it's a preliminary evaluation that happens before the model gets released, you know, every time a model goes on OpenRouter or it goes on LLM Arena, all of a sudden any little, oh there's a new image model, like the ChatGPT images too, like that leaked because they were benchmarking it and people were like, oh there's a model coming. Like rumors of new models happen all the time, but somehow like the Commerce Department is just keeping it to themselves. They're not pumping anything. But anyway, the quote here from Casey director, Chris Ball says, 'Independent, rigorous measurement science is essential to understanding frontier AI and its national security implications. These expanded industry collaborations help us scale our work in the public interest at a critical moment.' Andrew Curran has some more context here. He says, 'To sum up, Anthropic, OpenAI, Google, Microsoft, and xAI all have new pre-release screening agreements with Casey. We don't know the details of the new rules yet. I assume they will be announced with the AI executive order and AI policy memo, both of which we may get today. This is May 5th.' And so my big question was like, what is going on with Meta? Like they have near-frontier models. They're taking AI really seriously. Mark Zuckerberg, we'll get into this, is investing $125 billion in capex. The shareholders are thinking, 'Oh, what's going to happen here?' And he's like, super intelligence. And you think the government would be like, 'Well, you don't just get to like sit over there while all five of the other leading labs...'
Over a hundred billion dollars to work.
Yeah. Like probably a bigger investment than xAI right now in terms of compute capacity and maybe in terms of research capacity.
Future capacity.
Yeah. Yeah. Yeah. Future capacity. Exactly. So, you would expect also it's just one of these things where if five of the top labs are jumping into a particular initiative, there's so little risk to jumping in as well. Anthropic and xAI are both doing it on the left and the right, like there's clearly room for Meta to just say, 'Yeah, we're cool with that, too. We're riding with that.' Not everyone's cool with it, though. There is some major pushback, mostly from George Hotz over at Tiny Corp. We can pull up some of this. Let's go through Andrew Curran's take on what will happen here. So the EO will create a mixed group of tech CEOs and administration officials who will work out the rules for new release regulations. I see a lot of people calling this a win for Eliezer Yudkowsky's side. But even if the new rules are very strict, the stop-pause group don't actually get what they want. This doesn't stop or even slow AI advancement, according to Andrew. It only slows the rate of public releases. Capabilities will be advancing at full speed. I don't know how true that is because if you don't release the model at all, like you can't monetize it and then justify the next leg of capex. So I'm not 100% sure that this has no effect on the speed of AI progress. But I take Andrew's point. The labs will finish training the new models and then submit them for government approval. Given the government currently doesn't even want Anthropic to release Mythos, how long will the regulatory process take for something twice as powerful, five times as powerful, 10 times as powerful? All of the incentives are for the government to slow down releases. Let's say OpenAI finishes training GPT-6 and it's twice as capable as anything available today. If the government approves it and something bad happens, they take the blame, and the longer they hold it back, the longer the government agencies get exclusive access. So there's a little bit of principal-agent problem here. Meanwhile, OpenAI will start using GPT-6 to train GPT-6.1. Again, unclear if you can actually marshal the compute for the next model without the revenue traction, but his point holds. If they finish 6.1 while 6 is stuck in approval, they'll move on to training 6.2 with it. This potentially creates strange situations where the labs are many generations ahead internally while the public is still waiting for something they submitted months ago. So you could see, like you're doing all this weird value capture and the size of the firm bloats and bloats because you can't release it, so you just wind up rolling out a product...
Maximizing your own use of it.
Yeah. Which is like maybe not the most democratic solution. Maybe not the most, you know, positive outcome.
I think everyone needs to wait to have super strong opinions until we see the EO that goes along with this because yeah, it'll be interesting. You know, the bad scenario is like somebody that let's say has a neolab wants to make, you know, release a model and they've raised $100 million. Do they have to go through the same process as, you know, one of these larger labs or a hyperscaler in getting, you know, getting these sorts of approvals? So I don't think we can really take any strong stance until we understand what the intentions of the EO are. So, the meta commentary on this, Zach says, 'It will seriously not surprise me if they try to require permits or licenses to use AI and restrict local model downloads. You really should be buying hardware.' There's an interesting company that's launching basically a Tesla Powerwall for Blackwells. You can just get GPUs mounted to your house. Maybe that's the future. Tinybox is another solution. And Tiny Corp has chimed in, says, 'Ah, maybe they can enlist the MPAA and RIAA.' Calling back to the piracy debates of 2005 or so, to help with the restricting of downloads. I can see the lawsuit against the grandma who downloaded Qwen already. And Tiny Corp is chiming in, saying that this is a win for China in particular. Remember the time they regulated crypto over 40 bits? No, no, that's too many bits for the people. They were losers then and they'll be losers again. The difference this time is that it will cost them cultural influence to the Chinese. And he says Jensen tried to...
Interesting, like play it out a little bit. Chinese, you know, these open source models are around eight months behind right now. The gap is widening, likely due in some part to export controls. But if you enter a scenario where US labs get kind of like hung up and again have to keep these capabilities internally, but these Chinese open source models are just like shipping as they're ready over and over and over and over, and then sort of like compounding on the collective learnings. If you can imagine that one scenario is helping them close that gap, at least close the gap between what's publicly available from the American labs and what's publicly available via Chinese open source, even if the lab's actual capabilities are much farther ahead, but they're just unable to release that, right?
And to be clear, I mean these are two very different philosophies, but from my perspective, George Hotz has been advocating for no model gap app and in fact the strongest models possible being fully open sourced, being fully available to everyone. He wants no authoritarian control. He has taken the anti-authoritarian stance on a...
He wants all products to be free.
Yeah, maybe, I don't know.
I mean, not the physical things he wants. He wants to build things, but Tyler, how have you been wrestling with this question of an FDA for the AI, potentially a DMV for the AI?
Yeah. I mean, like obviously it really depends on how it's implemented. Like in the super like safety pill scenario, like yeah, it's probably not good for innovation broadly. But I think, so it's called like Casey, right? Center for AI Standards and Innovation. It used to be called the AI Safety Institute under Biden. And so they changed it. So I think even from that you can maybe take away something where like maybe this will actually have really not much to do with safety at all. Maybe it's just like we need to normalize how we publish SWE-bench scores because some people when they report it it's like different, you know, how you actually run these benchmarks or something.
If it's like standards.
I think like that could be a possibility that seems like totally fine.
Yeah. And then the other risk is, I mean, there's regulatory capture potentially. It'd be very hard for startups to get approved if it's like, oh well, you know, we've heard the story of Anduril where, you know, it's like to get the company off the ground, hire 50 lobbyists immediately, right? And like you can imagine a situation where, okay, you know, there's some uncertainty about what's happening with SSI and I susver project, but I think I like the idea of an individual brilliant researcher going off and researching an interesting path. If all of a sudden it's like, 'Oh, well, if you want to do a neolab, set up an office in DC, get your licenses, then you can start training. Then you can make sure that you're approved and that you have the right deals in place.' And you could wind up with a lot of not just regulatory capture, but also regulatory favoritism where whoever is in good with...
This particular administration gets their models approved faster and you wind up with a lot of risky outcomes. Anyway, I believe we have Courtroom Simulator ready for a demo. If you're tuning in to the Elon Musk OpenAI trial, you're live streaming the audio in a different tab because we can't re-stream it for you. You live stream the audio in a different tab and then you pull up Courtroom Simulator.
You pull it up on the big screen.
I think we're going to pull it up on the big screen. This is the best way to feel what it's like to be Mike Isaac from the New York Times. So, you go to court-sim.vercel.app and you will be entered into a really photorealistic representation of... Hey, the flag is correct. The first version had the flag a little bit flipped. Let's go find Mike Isaac. I believe he's in the first row on the other side, on the right side of the court. There he is with...
Looking good. The rat king.
Graphics in video games have just gotten so good these days. It's really remarkable. And you can sit right next to him. His pillow that he sits on has not been modeled into the game yet, but that will be coming in a DLC.
And in the sim, he already ate all of his food for the day, right?
Yeah. Yeah. Yeah. We need to have a banana mechanic.
Yeah, you should be able to go and hand a banana.
And can you go into the court? That does not seem okay. I think the bailiff should come and arrest you if you wind up coming to the... Is that in the game, the soundboard place? There's a soundboard in the game. Okay, so there's Sam who's on the stand today. I think that was Elon.
Okay, that's the judge, I believe. And I'm not sure who that is, but anyway, that's the judge.
Well, this is Courtroom Simulator. You're welcome to go and play it at court-sim.vercel.app. You can engage with all the jury members and really, really get a feel. Let's work on adding some more features here. I think you should be able to deliver some reinforcements to my startup. It's a lean startup over here. It's a V1. But we hope you have a fun time playing the simulation. It is court-sim.cell.com, I think, or app.app. Anyway...
In other news, Brian Armstrong sent an email to all employees at Coinbase announcing layoffs. 14% of the workforce at Coinbase. Two forces are converging at the same time, he says, and Coinbase needs to be front-footed to respond to both. First, the market. Coinbase is well capitalized, has diversified revenue streams, well positioned to weather any storm. But crypto is also on the verge of the next wave of adoption with stablecoins, prediction markets, tokenization. However, the business is still volatile quarter to quarter. There's a crypto cycle. We've sort of been touching on this, like there was a little bit of a stablecoin boom. It feels like it's a little bit flat. The Bitcoin price has been a little bit flat recently. And so he says, 'We're currently in a down market and we need to adjust our cost structure.' And then he says, 'Second, AI is changing how we work.' And so a lot of people, they jump to AI is taking all the jobs and that's not exactly what's happening. I think that there is a right-sizing going on at Coinbase and AI is hoping to be an enabler of more efficiency with a leaner headcount. The big question that I had that we were debating was, you know, is this like, how does this reflect on Brian Armstrong as a CEO? Is the goal to always have the perfect amount of headcount and never have to do any layoffs, or is it actually the role of a very aggressive and mature CEO to staff up aggressively during boom times, get the best people, and then when there's a down market, sort of reorganize and keep just the best on the team as you move forward with a leaner organization? And what does that mean for whether you want to work in an industry that has market cycles like that? You might want to go work at Costco or a tobacco company because those don't go through market cycles really at all. But if you're in crypto, you're probably not new to the idea of market cycles and you probably...
were born on a roller coaster.
You're born on the roller coaster. You've been on the roller coaster and you might be moving over to a new roller coaster if you're one of the unfortunate people that will be laid off over the next couple months. But I'm sure there will be many, many more opportunities. Some folks might even move over into the AI world that is booming and definitely hiring across the board. Was there anything else on the read?
I would say that overall, the positive thing from the response that I've seen so far is a lot of people are not dooming around this. They're just saying like, 'Hey, look, this is something that Coinbase has done before. Cyclical business.' And no one is saying it's over for white collar work.
So Andrew Young has a little bit of a take here. One takeaway: no pure managers. Every leader has to be working as an individual contributor. The pure people manager role, the one that built most corporate career ladders over the last 50 years, no longer exists at Coinbase. Leaders will have 15 plus direct reports. That is insane. Previously, managers capped out at six direct reports. I felt like 10 was sort of the platonic ideal there. But Andrew says that would be impossible without AI. And Coinbase is testing one-person teams. A single person is the engineer, the designer, and the PM. A pod of one with agents. Org structures are being redesigned in front of our eyes. Derek Thompson has the other side. He says Coinbase is the latest tech company, including Block, Salesforce, to announce big layoffs and cite AI coding productivity as a driver. But Derek Thompson is looking at the stock prices of these companies. Salesforce down 31%. Coinbase down 23%. In the past 5 years, it's sort of a roller coaster. I don't know if this is a great read because it's sort of flat. I don't know. Anyway, Block down a lot. So, some right-sizing makes sense, but he's calling it AI washing of layoffs. They were coming anyway and you want to spin them in a positive way. I don't know. TBD. TBD.
Yeah. I don't think we have any clarity on what kind of teams had greater percentages of cuts, right? Because again, there was some reporting, I don't know how true it ended up being, that when Square did their big cuts, engineering was getting laid off. Engineers getting laid off.
I don't know if that was confirmed.
Yeah, I don't know if it was confirmed, but that does imply that it's actually more grounded in AI productivity. But again, a lot of teams that we're seeing, their engineers are a lot more effective. They can do a lot more and so maybe you want more engineers before.
Well, lots of strengths in the business, lots of liquidity, very hard. Yeah, you could vibe code Coinbase, but how do you get all the customers and all the liquidity and all the features and all the regulatory and all the other stuff? There are a million reasons to be excited about Brian Armstrong's next chapter at Coinbase. Well, and Ryan Peterson chimed in to sort of push back on Derek Thompson. He says in the layoff announcement, the CEO literally said that the first reason they were doing layoffs is because crypto is in a bear market and they need to adjust their cost structure. AI productivity was secondary. And so I think that's a fair take. Let's go over to Meta. Some debate, people are digesting Meta's earnings. The stock sold off a bunch. Where is Meta today? It's a bargain stock. It's a $1.5 trillion company and over the past five days down 10%. It's pretty flat over a year and over the past five years it's nearly doubled. So not doing too bad. But the debate is over capex. Meta's core ad business is ripping up 33% in Q1 year-over-year. Strong ad impressions, strong margins, but the market is worried that the beautiful cash machine is turning into an AI capex furnace. $125 to $145 billion in capex. It's a lot of capex, especially for a company without a cloud business that can resell capacity if they wind up with extra capacity. If you can't use it all, what do you do with it? Meta doesn't really have an answer to that necessarily. And so that's what has the market worried. And even xAI is having trouble, at least reportedly, driving demand for compute that they have. There was that 11% number. It seemed like that was a bit of an overstatement, but there's something to be said for when you have a near frontier but slightly lagging model, you need to find a solution to actually have offtake and you might just not be able to spin up a cloud platform on day one. Elon took a different route, you know, partnering with Cursor that has a ton of demand and is a front door to AI coding. And so that makes a reasonable amount of sense for the compute. Who's GPU rich? Who's GPU poor? Let's match them up. Meta doesn't have an obvious dance partner like Google, like Azure, like AWS, right? And so the stock looks cheap now, but investors are unclear about the payoff around AI spend, and there's also some legal and regulatory risk. But...
Yeah, I mean, people at the end of the day are worried that he's going to repeat the metaverse saga and he's going to spend a lot of money and not get very far.
I think that coming up on what is it, almost 10-ish months ago, he started talking about personal superintelligence.
It sounds cool. What does that mean? Does it mean AI in your Meta Ray-Bans, which are selling a lot of units?
So that's a great point. Does it mean creating a consumer LLM like they have with Meta AI now? Meta AI now is squarely in the top five of the App Store rankings, but it is the most competitive category potentially since search, right? And so I don't think that even though they have a solid model and they're able to get into that top five, I don't think at least investors are not ready to price in Meta AI becoming like a key player in consumer LLMs just yet, right? So it's all the spending and no revenue acceleration associated with that spending. All that being said, Meta has used AI to drive business value historically more effectively than almost any other business in the world, right? It is really, really insane to think that there is some scenario where the labs end up going public at valuations somewhere in the range of Meta while Meta is sitting there with $200 billion of annualized revenue growing 33% a year, one of the greatest businesses in all of human history, and not really getting full credit for their various AI initiatives.
Good point. Obviously the drop in people, which I think they, what do they call...
Daily active people. Users, they're not users, they're people.
Enjoyers. That's playing into this as well, but...
Well, we have our first guest of the show, Harley from Shopify. You got a little preview of him. He's in the waiting room, but he's now in the TBP and Ultradom. Harley, how are you doing? Welcome back to the show.
Good to uh good to be here. I I love coming on the show.
Yeah, it's always great having you. Uh, give us an update on Q1. Give us an update on where things are going in the Shopify world.
You know, uh, we had our first hundred billion GMV quarter ever in the history of the company last quarter, Q4. Uh, thank you. Uh, and we just had uh a a second uh our a consecutive quarter of GMV growth uh above $11 billion. Obviously the reason why we start with GMV is as you guys know it's it's a proxy for how merchants are doing. Uh, revenue was 3.2 billion. I was up 34% and we had free cash flow of $476 million. So across all the metrics uh we we beat which really excited about. I think the big big story here is that and this sort of came up a bunch on the call but the way that merchants are coming to Shopify now is is really quite remarkable. Like we're seeing some of the largest brands on the planet come on. We saw in LVMH and Balmain and BevMo and Orvis and Lululemon and in fact the number of merchants selling more than $100 million annually on Shopify has nearly doubled in the last the last two years. So that's really amazing. Uh, we're now like 14% of all US ecom. So we are the second largest online retailer in North America right now which is uh which is amazing. But then on the other side of the spectrum from like the big brands, I think one thing that I didn't get a chance to mention on the call today, but this is why I do your show here is that I actually think one thing that people are missing is that I think uh no group will benefit more than entrepreneurs in sort of this new AI era. It feels to us now that AI is not only making it more accessible, but also it's accelerating a lot faster. Um, I know I think you guys may be the one to introduce me to to Graza. Uh, I I I saw on your show a year ago. Uh, anyways, I mean Graza started on Shopify in 2023. They're doing they they're doing nine figures now and they just got acquired by Unilever for a billion bucks. I mean it is remarkable that not only the amount of starts happening but also how big they're getting. Um, so I'm really really proud of that. And of course we talked a lot about the agentic stuff on the call, both the the AI shopping but also Shopify Sidekick and what we're doing for our merchants.
We uh so uh let's double click on that BevMo uh Shopify use case or or or case study because BevMo was acquired by Goop and Goop has an app for local delivery. How are they using Shopify? That's an interesting partnership that I wouldn't have expected.
BevMo actually has physical stores all over the US >> and so we're powering all their physical stores as well. The reason actually I really like BevMo is because >> go back like this is my 43rd earnings call. If you go back to sort of the early days, you know, like right right after the IPO 2015, >> merchants were coming to Shopify for one reason. They were coming, it was small businesses coming for ecom and then obviously some of those business got much bigger and we went into enterprise and and then eventually we added on things like point of sale and more recently commerce. I think the interesting part of of the current business of Shopify is that there are so many on-ramps into the company that BevMo came to us, a very large company, not a DTC, you know, native brand. They came to us specifically for point of sale across all their physical stores and uh and actually our point of sale business is is going really really well, too. But BevMo's using us for that. At the same time, we're seeing, you know, I I you know, obviously I mentioned Graza, but like I just looked at um some stats from True Classic Tees just put out. This is a company started on Shopify like 5 years ago. They're now one of the largest t-shirt companies on the planet. So, we're getting a lot of small business getting started. Every 26 seconds or so, a brand new entrepreneur gets their first sale. And at the same time, we're seeing companies like Lululemon or Guild Group or Rag & Bone come to us, too. It feels like this is Shopify, you know, firing on all cylinders.
What are Oh, sorry. >> Go for it. >> Yeah. Just what what are people looking for on agentic commerce? It feels like there's a big opportunity. The models are getting better and better. Like the capabilities there. Are people looking for acceleration of uh of agentic behavior within the online commerce world or actual growth of e-commerce? Like is agentic commerce the thing that gets more people to place to to use the internet to buy things that are in retail? Like so >> yeah. >> Yeah.
So I I I think that's exactly the right question to ask. Um, hopefully now I've been on the show enough times and if if anyone who's watching is a merchant or a partner or just a someone who follows Shopify story, we we've we've set ourselves up to really be at the epicenter in sort of this AI era. So, just in terms of AI powered shopping, we've started building all this infrastructure years ago to connect demand and and and merchants on Shopify. We are currently the only platform that is selling inside of ChatGPT, Copilot, and Google. And what we're seeing is that these channels are actually becoming a very serious discovery engine. I'll give you some proof points. Um, we have seen in for Q1 of this year, we've actually seen orders from AI searches are up nearly 13x year. And new buyer orders from AI searches are occurring at twice the rate of traditional organic search. So, not only are more people obviously using agentic shopping, but in terms of new buyers, which to your point, are we bringing in more people? Like ecom as a percentage of total retail is still under 20% in the US. It's even lower here in Canada. I think it's about 25% in the UK. So, I think we're actually introducing um like more people to, you know, modern digital commerce through this as well. And I think the way it's happening is they're starting with things like research or looking up a recipe and then very quickly ending up, you know, buying from like our home, excuse our place or or one of these, you know, amazing DTC brands that we have on Shopify. Um, and we now have about a billion products in the catalog. Um, and and and so that means that every Shopify merchant SKU is fully syndicated.
Yeah. The other thing I think that that is uh hopefully is is now um pretty clear is is UCP. We co-developed UCP with um with Google. I think I announced it on the show actually during NRF, but it is now has become the industry standard this past week. Last week, Amazon, Meta, Microsoft, Salesforce, and Stripe all joined the UCP tech council. Um, and and and so it looks like that's going to be the open standard that that that that is going to win, which we're really excited about because we think it's the open standard that actually thinks about the entire commerce experience.
Okay. Uh, go deeper on the implications of agentic commerce because I have this thesis that uh that agentic commerce will be most acceleratory in higher ticket more considered purchases because when I think about you know just you know buying a t-shirt maybe I need to run a deep research report for that to land on uh I forget what shirt brand you use it's nice but uh >> true classic tees >> true classic tees. Um, but uh when I think about the process of like buying a car, there's so many more tradeoffs. It's so much more complicated. That's something that someone is probably noodling on with an LLM for weeks before they actually go and purchase. And I'm wondering if you're
Yeah, I'll give you it. Really, I think comes down to the difference of like considered purchases versus like pure utility purchases. So, like when I think about I don't right now I can't imagine myself using ChatGPT to like go grocery shopping. Maybe that changes in the future, but if I'm like ordering food to be delivered to my house, I kind of want to just see a catalog of a bunch of pictures and go through. Um, and maybe that more of that sort of moves into the sort of agent experience over time. But if I I was like I needed to find I wanted to find a cowboy hat, a lot of sun. I wanted to hide from the sun. So I'm searching like, okay, I want a cowboy hat for a brand that's been around for over 50 years, right? And so I land on like Stetson or that. >> Yeah. Or or Tootsies hopefully, which is which brand would we? I'm not sure. So I I'm not sure I fully agree, at least not yet. I do believe that the more research you do, >> um I I I think that the these agentic surface these agent services, these chats are places you will go to do deep research. So um a car is a good example. You do weeks of it as well. On the flip side though, what's amazing about agentic commerce I think in general is that unlike search-based commerce where you can um you can put your thumb on the scale through advertising and ad dollars, >> it's agentic is more merit-based and because it's more merit-based. So most people watching maybe I I don't know if this is true but most people watching may not have heard of True Classic Tees but they may have done a bunch of research in previous conversations in fact maybe over over many months where they're talking about you know their their typical price of a t-shirt. Um, today I'm wearing a James Perse t-shirt which is more expensive than True Classic. So it know it knows that I I look for I love James Perse. I've talked to it about James Perse. Where can I buy it? What's the GSM weight and all that stuff? Will this be good for this particular trip I'm taking? >> I actually think that it'll bias smaller brands that you may not have heard of over larger brands because again, it it has a full contextual history of every conversation you've ever had. And if
If it's merit-based, then it's going to surface brands you may not otherwise have heard of before. Whereas, if you type in 'sneakers' on any search engine, you're probably going to get to Foot Locker pretty quickly. I don't think you get to Foot Locker through an agentic conversation. I think you end up with some sort of direct-to-consumer brand, especially if it knows you like On Running and you're buying tennis shoes and now you're looking for hiking boots. It should show you, based on merit, an On Running pair of hiking boots.
Yeah. The long tail getting fatter.
Which, again, we'll see what happens. The reason I also like these proof points, again, like orders are up. Traffic though, if you look at just AI-driven traffic to Shopify stores, that has grown 8x year-over-year. So I don't want to be handwavy anymore about agentic. I think there's enough handwaving happening around that. I think actually looking at, okay, how many people are coming? What is the search traffic like? What are the conversion rates looking like? I think right now that is getting real. I also think that you probably saw some recent headlines, but you saw ChatGPT has now moved into this in-app browser for their checkout, which is literally the Shopify checkout. We are really happy about that. We think actually what that allows is that it means that as a merchant who set up their loyalty and their subscriptions and their merchandising and their shipping components, their tax, now the experience inside agentic commerce is as good as it would be on the online store, which is kind of what I think consumers want. Like it needs to feel effortlessly similar to what your normal experience is, even though you're sitting in a conversation window.
Yeah. Yeah, that's something. It was interesting last year watching a bunch of AI web browsers get built when the entire time my thought was like, LLMs effectively are web browsers. They were becoming browsers, right? Where you're having a conversation but then it can pull in stuff from the traditional web. Very cool.
Yeah, I think this is also where one, you'll see — I don't have any great examples yet, but I assume next time I'm on the show I'll be able to bring some examples of these breakout SMBs who effectively, you know, like I love the Gymshark story going from zero to billion-dollar acquisition in three years. We just haven't seen that before very often. I think you'll see way more of that. I also think what you'll end up with is you'll just have more people starting businesses that otherwise may have not. So the barrier to entry is further down and then their ability to scale is also going to be...
We were talking to the Collison brothers about exactly that. They're seeing a ton more formation on Atlas.
Yeah, I saw that crazy slide of Atlas.
Yeah. Incorporation's way, way up and it's gotten so easy. But I think the first phase is make it easy to get started. I think where this really gets interesting is how many more brands exist that we love, the consumers love, because they're able to scale much, much faster.
Well, thank you so much for taking the time to come chat with us. Congrats on the progress and looking forward to the next one.
Goodbye.
Up next, we have the CEO of GE Vernova, Scott Strazik. AI data center electricity demand is expected to double or triple by 2035 and we have the perfect person to talk to about it. Nice to meet you. Thank you so much for coming on down.
Thanks for having me. I appreciate it.
I would love to start a little bit with your background because GE Vernova is — I mean, we can go through a little bit of the history there, but you've been with the company since before the creation of the company. You're older than the company you work for.
The practical reality, my adult life I was with GE and then ultimately GE Vernova, but I started with General Electric right out of college and I've lived through quite a few chapters, some humbling, some less humbling, but this one is invigorating. It's a moment that we're going to take advantage of. We see the purpose of what we're doing every day very clearly, both in driving economic growth, AI, national security — energy security is very much national security in a lot of the parts of the world right now. So we've got a real opportunity to serve and that's exactly what we're going to do.
So I think most people will know natural gas turbine for data center, but the business is bigger than that. Like zoom out for us and walk us through the shape of the business because, you know, massive workforce, 85,000 people, something like that.
Yeah. Remarkable. So walk me through the mechanics of the business and then we will double-click on the AI story.
Of course. Sounds great. I mean, you take a step back, about a third of the world's electricity every day is created with our equipment through our customers. If you exclude China, so it's substantial. And even with China, it's 25%.
It's an obligation, it's a responsibility.
The Jamie Dimon of energy.
It's a big deal, right? So that's just keeping the lights on, notwithstanding what we need to do to then add the growth for where there is today. And that install base, that's gas turbines, that's nuclear power plants, that's wind turbines, it's a lot of the electrical equipment. That's actually our fastest growing business, is everything associated with electrification and the grid and trying to make the grid more efficient and smarter because frankly in some ways you can make more progress there quicker than you can building a new power plant.
So all those variables are parts of the business and then in an economic opportunity where we need a little bit of everything.
And applying the right technologies where the right resources exist.
To drive better economic answers and more resilient electrons. Yeah, it seems like the market or just the stories that I hear are that you're sold out. You have this massive backlog and it's more of the same. There's demand for more of the same, but it feels like you're also working on R&D. Talk to me about the trade-off versus just scaling what's already working, what's worked for years, I imagine, decades in many cases, versus a focus on research and development, the next technology.
So, the end product in some cases may look the same coming out of the factory. What we're producing or how we're producing it to scale and make more is changing drastically. So you think about gas turbines right now, we're adding a lot of production workers to build more gas turbines. We're also adding 400 machines this year to automate a lot more in the process. Frankly, more of the innovation right now may be how do we scale faster and in a more economic way so we can win the affordability game while simultaneously investing for the future.
Can you explain like I'm five some of those machines that you're buying? Some of your supply chain that people know, 3D printers or CNC machines or different — you just hammer it out yourself or whatever, but I imagine it's very complicated. What does your supply chain look like?
We get big castings, big forgings that we get inside the factory, then need to machine down to the right, put the right material sciences, the right coatings. When you think about a gas turbine, it's got to be very heat resistant because it's a fire that really is spinning inside that gas turbine to create the electrical power ultimately.
Yeah.
On the demand side, I have been hearing about the absolute knockout drag-out fight for energy. And in many cases, I don't know if it's relevant to you, but in many cases, we've heard about folks involved in the AI buildout putting in orders for more than they need and then sort of dialing it back because they know that certain things won't get permitted, certain things will get delayed. Does that happen to you? Are your customers coming to you and saying that they need twice what you think they need, and there's this like battle of back and forth?
I wouldn't say so. What I see happening more frequently is many of our customers are developing a number of projects, acknowledging that not all those sites are going to be chosen and the actual supply is a little bit liquid and they can move it around.
And that's more what's happening. So sometimes we'll see press releases on leases canceled or sites that customers walk away from. That's typical development really.
What's really happening is our customers are buying at a realization rate knowing they're only going to ultimately develop a certain proportion of those sites. But we don't see any of our customers walking away from what they're securing today. Not at all.
Yeah, that makes sense. What are your roadblocks? What roadblocks do you see to scaling either inside your business or at sort of the macro political level, permitting, approvals, regulatory? Like what is on your short list of what America needs to do, what the industry needs to do to get right to actually continue to scale?
In some ways, what we do is a little bit easier, candidly, because our equipment is built in a controlled factory. In fact, the hardest part in some ways is actually building the plant itself, the construction.
Because every plant in different parts of the country, different parts of the world, everyone's a little bit different.
And getting that craft labor to those locations where a lot of the data centers and a lot of the electrification demand is often very remote.
And that is going to take us some time. So for me, every gas turbine, wind turbine that I build, I'm using the same workforce every day.
Same people coming in Monday through Friday.
We can meet the moment from an equipment demand perspective, but we need out in the field for those plants to be built so they're ready for our pedestals when they come out of the factory. Getting that orchestra right, that the site is ready when our equipment is ready, is going to be a challenge. And we're going to have to keep investing in craft labor in these locations to do the build.
What does investing in craft labor look like? I mean, we've heard about reskilling, retraining, trade schools. Is there a drought? Are people already seeing the news and seeing the stock price and retraining and reskilling, or is this more of like a decade-long process for America?
We've been successful so far. We've added about 1,800 production workers in the last 15 months. We'll keep growing that number from here. You work to make the jobs more attractive. Part of how you do that is you take out the dull, dirty, unsafe work and that's where you automate.
And then you allow these workers to thrive doing what they want to do. But the reality is they have a lot of pride in what they're doing every day. We put the customer name where the project is going with everything we build. So they understand the impact they're having. So it's not just a piece of electrical equipment. They know how they're providing light and a middle class growth to Vietnam or economic expansion in the US and I think they see it as real and...
Have you considered engraving the signature of the technician on the engine like Mercedes does?
Exactly. But I think making personal matches.
Yeah. Yeah. That makes a lot of sense.
So there's a lot of pride.
What else is in the portfolio? I mean, you mentioned wind. I think nuclear you mentioned briefly. Like walk me through the scope of your ambitions in energy generation broadly for the business.
Yeah. But nuclear is important. We're building small modular reactors today. Think 300 megawatt machines that can be — they're big but they're built inside an American football field.
So that's 300,000 US homes that can be powered from one small modular reactor that is within a US football field.
Yeah. So that's very — 300 megawatts and that still qualifies as small and modular. That's...
Relative to what it used to be for nuclear, you know, it used to be a gigawatt.
Yeah. Like a big Westinghouse development, multiple years, precisely built on site, massive concrete structures.
Now we're trying to build a lot of them within this construct so that we can come down the cost curve because the reality is every nuclear plant historically was a snowflake and that's why the industry never really became economically competitive.
Yeah. No standardization.
And we're fixing that. So that's important.
And what is the regulatory approval process for that? I mean, I've heard that there's this famous line around like the NRC hasn't approved a new design in decades. Is that what you're feeling? Does it feel like there's a shift?
Yeah. I guess the question is like when — we're now about two years and a month since the spinout.
Was small modular reactors on the roadmap when the spinout occurred and then the opportunity has got more compelling from a regulatory standpoint since then?
Yeah. We're in construction on the first plant right now. It's in Canada outside of Toronto. We will get NRC approval for the first plant in the US. I think this summer we'll be in construction and we've committed to the administration that as soon as we get the approval, we'll be ready to work on day one. And then from when we start construction, it's about a four-year cycle from the beginning to getting to commissioning. So this can start to happen. And that's where the comparison of the large block, it's more of a 10-year construction cycle. That's a long time to wait.
But even if you want 1.2, 2 gigawatt, the beauty of the small modular reactors is you can do them in blocks and every two years after you get the first one built, add 300 megawatt, which also then meets a lot of our customer expectations to grow with the electrons. So this is going to be an important part of the equation for us.
Are those two completely separate divisions or is there some fluidity across the workforce on the front end of the business?
Commercially, there certainly is with customer relationships. But from a manufacturing and an engineering perspective, they're pretty distinctly different.
That makes sense. What has been surprising to you about where the sort of energy infrastructure buildout is at today relative to the AI boom versus what was predictable when the spinout first happened?
Certainly in the US, the demand has accelerated greatly in the last two years and one month. No question about it.
But weren't you seeing that? Didn't you feel like you had a...
Early stages. Early stages. What I'd say that has surprised us though is there was probably concern that as the US market started to take off, it would price out a lot of other markets in the world from even being capable of buying the equipment.
The reality is we still see very strong demand signals from a lot of other markets in the world. I mean, in the first quarter we had very strong demand in Vietnam, in Mexico, in Canada, in Saudi, in Kuwait. So the global dynamic is still driving real growth above and beyond the AI dynamic.
The reality is AI specific, it's about 20% of our backlog today.
Yeah. Interesting. And those international projects, I guess due to having 30% of the global market, right? The scale is just already so immense that even if you have a massive acceleration via new technology cycle...
And I think with a lot going on in the world today, more countries just realize for just national security reasons they need a diverse portfolio of solutions to electrify and energize their country. And that really goes back to the war in Ukraine and it's only been accentuated over the last five years.
How have you processed some of the writing coming out of places like San Francisco? I'm thinking, you know, Leopold's Situational Awareness, where if you look back over the last couple years, he's been right over and over and over. At the time, a lot of people reading that, even in AI, seemed to think like, hey, this seems pretty aggressive. But to date, he's been quite accurate. Even some more sci-fi kind of exercises like AI 2027 have been relatively accurate up until this point. But how did you process basically kind of the analysis coming out of the AI community and SF going back a few years?
It's hard for me to kind of decipher how this is going to play out and who are going to be winners and losers and how that plays. But it's very hard when you take a step back in this country and really throughout the world to not believe we don't need a lot more infrastructure build for communities and economies to thrive. So at the end of the day, we're deeply investing in these businesses to meet this moment and we think we're pretty early in this journey and are going to keep investing into it.
Is sovereign AI driving any of the international demand that you mentioned?
You start to see some of the hyperscalers evaluate more buildout in Southeast Asia as an example. So they'll try to use exactly Malaysia, Indonesia, natural gas but adjacent to Singapore for load.
So in that regard, that's happening. Certainly Taiwan is one of our biggest markets. That's because of TSMC.
So we are commissioning almost 10 gigawatts of power primarily for TSMC.
Not for data centers but for fabs.
For the fabs, for the chip.
I had no idea they were so electricity...
So significant demands in Taiwan for that adjacency. And that's happening in the US too as you think about some of the chip fab factories that are being built. So it's not just the AI factories, it's the infrastructure around it that needs the electron.
Okay. Help me understand some of the less sexy or less obvious maybe pieces of the grid that need to be modernized, need to scale up. If we're familiar with the turbine, we're familiar with the nuclear reactor, these are very tractable things, but I'm sure there's so many other pieces of the electron delivery supply chain that are important and you're involved in.
I mean, if you just take a step back, the system was built for a coal plant to in one direction flow electrons towards a house or a factory and that was at a 24/7 flow, very stable, and then you still had peak demand during the day or during the summer.
So that was where we started. Now you have all these variable forms of generation.
On top of that, you have electrons that are going in multiple directions because you have people selling their electrons back into the grid on their roof.
And the system is not optimizing all those electrons. So we have an old machine for all intents and purposes that is not doing a great job matching the supply with where the demand is. Part of that is for boring reasons like we have islanded grids in the US. It's not one connected machine to project things to where it needs to be.
Is that possible given the geographic diversity and like how spread out the United States is? Is it possible that I could sell electrons in California and it could wind up in New York?
Certainly in the 48 states for sure. I mean, there's no reason we shouldn't be able to build a meshed grid that works. Today we don't have that. I think in this regard, the federal government's pushing hard on historical precedent to try to drive efficiency because that's a faster way to incremental electrons in some way than simply building more plants. We're working on that. Part of that is making new connections that bring islands together. Part of it is also, and this is frankly the harder part, different regulatory structures and market practices to allow different players to get paid.
And that is sometimes harder.
But we're working our way through that.
Talk about nuclear timelines. We've seen a bunch of announcements from hyperscalers. It feels like 2030, 2032. You see 2035. Even stepping aside from your business, just broadly, I think a lot of us tracked Vogtle that took decades. When do you think we'll be getting new electrons from a new nuclear plant in America?
You bet. I mean, I start with the fact that we've got about 56 existing plants in the US that can add 5 gigawatt of power by upgrading what's already running.
Okay. That can happen this time and that counts.
Yeah, that'll happen soon. So that's the first step.
