As three inveterate Twitter junkies here, we know where the action is, and it ain't on Threads. Okay, let's be honest, people. If there was action on Threads, Harry, you'd be there. I would. If there was action on Mars, Harry, you'd be there.
This week: number one, Apple sues OpenAI for trade secret theft. Number two, Meta fights back on the coding model front, and Zuck is back on X. And then number three, SK Hynix prices a $26.5 billion NASDAQ listing, bringing the compute and infrastructure layer to the public domain.
To some extent, it may have been a distraction for OpenAI that they were so successful in consumer. If this was a mercy killing, Apple may have done them a favor. In the early stage, when you're on the board, you can't invest in two competitors. In the late stage, structurally, you have to.
If you don't want to be worth 1x, like, do something before it's too late, man. Ready to get him?
Guys, it is so good to be back. As always, there is a lot for us to discuss. Why don't we start with the trillion-dollar company that is Apple, as it brings in another trillion-dollar company in the form of OpenAI, where Apple sues OpenAI for trade secret theft. Essentially, one of Apple's long-term, 24-year vets was taking secrets to OpenAI and sharing information that he shouldn't have been.
Just to be clear on this, it's a really serious thing, and there's actually some conclusions that we can share with listeners in a second. But on the facts, the 24-year veteran is Tan Tang, and he's named in the suit, but the person who was alleged to have taken the secrets was the six-year Apple employee, Changlu. Tan Tang is in charge of the hardware division at OpenAI after having been a super secret executive, senior executive at Apple. And the allegation for him is that he encouraged hires to bring the stuff over, right? Which is just slightly different, right?
Absolutely. And he encouraged them to bring actual parts for show and tell.
Exactly. And you know, I think the big picture here is, in particular, I mean, so let's start with the advice to anyone listening: don't ever do this, because the person on point, the individual who brought that stuff across, I read the lawsuit this morning, 41 pages, didn't read it all, but kind of got through a good slug of it. There's a lot of fact-based evidence that the person who took this stuff did it, and he's screwed. And the first thing he's going to discover is there's no gratitude in litigation, right? You think you're helping your new employer or trying to get a job. They're going to burn you so fast your head's going to turn. Well, they're going to say, like Claude Rains in Casablanca, 'I'm shocked and appalled to find that there's gambling going on here,' and he's going to be left high and dry. Don't ever, as someone interviewing for a job, ever do this, because you're going to be screwed. That guy's toast, right?
Yeah, I hadn't thought about that. You think you're helping your company. He's already lost his job. Forget what his comp exit package is. He's already gone, right? You're not helping your company.
Agreed. So here's, and then the next level up is the gentleman you mentioned, Tan Tang, who is, he also has some issues. It's not as clear because there's no fact-based evidence, but it's not just hearsay because a bunch of folks have said, you know, he did it. He encouraged me to do this. And even in the email traffic with Changlu, there was a lot of, 'Hey, we were asked to do this.' So the risk he's up against is that there's now going to be a ton of deposition, a ton of discovery. And if there's any smoking gun that has him emailing or encouraging people to say, 'Hey, bring in information when you're coming in for your next interview here at OpenAI,' then he's screwed too.
And the same thing will happen because, you know, now you get to the highest level. Look, Apple in one sense is quote-unquote happy because they're really pissed that 400 people from Apple have gone to move to OpenAI, and traditionally over the years Apple has always hated, in my view, you know, almost to the point of not being appropriate, people leaving their place and going to someone else, and really leaned on other companies to not hire, right? And I think, as I say, at some point that gets inappropriate, and see this whole Steve Jobs kind of issue 15 years ago. But the bottom line is they're not looking at this and going, you know, the way when you want to get something for something and then the other guy makes a footfall and they're, 'Now we have them.' So they know they have the junior guy, they probably have the VP of engineering, and they're going to twist the knife, depose everyone, and see how high up it goes. They're pissed and they've got leverage. So it's a tough place to be, right? Because really what they want isn't, I mean, you know, they care about these secrets, but what they're really dealing with is, you know, the big picture comment is OpenAI's been talking about building a hardware device. That's obviously why they're hiring 400 people from Apple, and Apple's pissed, and now they've got some leverage, and they're going to find out what's going on here.
Yeah. You know, it's a tough one. To Rory's point, we all have history in our operating or investing experience where you do hire folks for their domain expertise every day, right? That's employment portability. And where the line is for trade secrets, it's actually somewhat nuanced, not only in the law, maybe Rory would challenge me, but in practice, because not everyone gets sued, right? People are stealing trade secrets all the time. They shouldn't be, right? The salesperson's got to be careful when they take the rolodex. Once in a while they get burned for that. Once in a while, a sales rep actually, they lash out when one goes to a competitor in your portfolio. They get sued for taking their actual rolodex, not their conceptual one. They get sued for downloading the CRM when they leave. And Jesus Christ, hacking into the old iFolder or whatever the hell he did to take documents from Apple after you leave, it happens. But that's too much, right? But still, you don't always get sued. And it's tough. But if we can't hire domain experts, sometimes we can't all invent LLMs ourselves. Sometimes we need a little help. I think, Jason, you're exactly right, but I actually think it actually makes the point that you didn't need to do this.
Because stepping back, California is one of the most liberal places in the country in terms of employment law, which is, you know, there's no non-competes, non-competes are typically unenforceable. And there's this doctrine called inevitable disclosure, right, which is if you acquire knowledge in one company and you go to another company and you use that knowledge, the California courts are pretty friendly and they'll say it's inevitable because it's part of your trade. So they'll even say we can't stop you using the tricks you learned at company A at company B. And let me give you a very practical example of that.
Anthropic is the biggest beneficiary of the inability to enforce non-competes and the doctrine of inevitable disclosure. Because those seven employees who left, they basically walked out of them because they're way smarter than this guy. They probably didn't bring anything. No pieces of paper, no stolen code. But next morning they start up and they say, 'We just were thinking about this last week and we can keep thinking about this and everything in our brain is owned by us.' And there's lots of states in the union where you can't do that.
I wonder if Anthropic could have been started in Massachusetts. Probably not.
No, you're exactly right. Which, and I think California's, even though it sucks as the employer, I think California's been right, oddly enough, the rare policy triumph for the great state of California, it's been really good for innovation in the Bay Area. Right. So you already have that very friendly employee thing, right? And given, to Jason's point, given you have that, given you have that inevitability, why put yourself at risk by stealing some stuff that you just don't need? Sorry, how are you going to set up?
Well, my question was, okay, so these individuals are screwed. To be blunt, what does it mean in terms of timeline for OpenAI? Does it cause meaningful delays? Does it impact their ability to ship their hardware product when they wanted to?
Listen, if I had to read the tea leaves, hardware is probably on the bubble. This was one of Sam's initiatives that made sense when everything was working and OpenAI had an unassailable lead over its competitors. Right now everything's, you know, we chopped Sora. Why the hell did we buy TBPN as great as it seemed six months ago? Right. Hardware seemed like a great deal for six billion when I could get the legendary Jony Ive, but his team stole everything. What's the point of, like, this is just an effing distraction for something that's going to hemorrhage cash.
I'm all in. If we go back 60 weeks ago, I'm all in on persistent AI around us. So, Zuck, like with his new eyewear and all that, but this has got to be almost on the cutting board. It's like the Apple car. It's got to be close to being cut. And I think the reason this lawsuit may have happened is to push it over the line of getting cut. Just kill it, guys. Like, it's not directly cause and effect, but it really could push it over the line to just, let's put a pause on this for a year, boys. Let's put a pause on it for a year.
I didn't pile in on that initially because I feel I pile in on this point, but I totally agree. I remember when they did the individual deal, I was like, this is stupid. The only, sometimes, you know, the cliche, keep the main thing the main thing. The main thing here is that LLMs are really amazing at code and the amount of economic value that can be created from that dwarfs everything else. Right? I've seen some of the columnists, Ben Thompson, even articulate the view that I was, Ben, that to some extent it may have been a distraction for OpenAI that they were so successful in consumer, right? And because where the value has clearly been created is not just within enterprise, but within enterprise for coding. And I'm with you, Jason, if this was a mercy killing, Apple may have done him a favor. You know, move on, slow it down, stop doing this.
And just to your point. If you think about when a year ago when OpenAI was a consumer company, I know you thought it didn't make sense, but step back. It actually all makes sense. If you're a consumer company, should we go into hardware? At some point, you do, you think about it. Should we go into media as a consumer company? Yeah. I mean, we could buy a top podcast for a million, four hundred something million. Like worst case, it doesn't work. Like if you're a pure consumer play, this is how you keep building. You add to this, you add to the layers of the concentric circle, right? But if you're an enterprise death march LLM, it makes no sense. It's baffling, right? Let someone else build the pin, the pin that clips to your shirt.
You mentioned that one.
You mentioned Zuck. Zuck returns to X, and he returns to X for Llama Spark 1.1, and the first time ever that Meta starts charging developers to use their own models. It was a very meaningful release. Some people think, Alexander Wang called the pricing of it in particular very aggressive versus OpenAI and Anthropic. As I said, Zuck broke a three-year silence on X to launch it. How did we think about the release of Spark, and was it a meaningful progression for Meta in the fight against OpenAI and Anthropic?
Yes. And I'd like to come back to how it's actually, let's do on a tangent first. I wonder what the internal Zuck algorithm was, which is how good does something have to be before I, spitting and cursing, put something on the competitor social media product? I mean, you know, maybe the comms person goes in and says, 'I think the new glass release is awesome. Should we use X?' He's like, 'Fuck no, right?' And then finally he's like, 'All the developers are on X. If we're going to launch Spark, we just look like idiots.' And then he's like, 'Okay, I'll type something on the clown car company.' Right. Anyway, that's fine. But before you move off that, Alexander Wang has done 27 posts on X to three posts on Threads, and it actually got people quite pissed off online, and I think internally...
They probably forgot they had Threads, as did I until the other day. It's got 400 million users apparently.
Because it's integrated into Instagram. You can't help but see it. It doesn't mean there's any real engagement.
Activity is the word. Engagement. Let's be honest. As three inveterate Twitter junkies here, guys, we know where the action is, and it ain't on Threads, okay? Let's be honest, people. If there was action on Threads, Harry, you'd be there. If there was action on Mars, Harry, you'd be there. So let's focus on the real point. I thought there's a ton of, like, I mean, and the stock jumped nicely because a ton of things happened here, right? You're right. They shipped Spark 1.1, they move away from, it's free, as I always remind myself when I forget, it wasn't, it wasn't open source, it was open weight, but they've moved away from that to charging on an API. You know, they've embraced the same business model as the other frontier labs, right? It's a decent product. I've used it as an end consumer. It actually, you know, I compared and contrasted it to Claude, and some answers it was pretty good, but of course the real test isn't really the end user, the real test is coding. And you know, they reported good benchmarks, but in the past their benchmarks haven't been as good, and we'll talk in a second about benchmarks, but fundamentally the big picture point is in the game, using the same business model as everyone else, entering the marketplace with a low-price product, big hit to competition. I mean, it's, you know, going toe-to-toe with the balance sheet to afford it, right? Will it be ROI positive over five or ten years to be the fourth player in the market selling at that price? I don't know. But, you know, if you're OpenAI and Anthropic, you'd have preferred this not to happen.
It seems to me like, look, the tweets are fun. Clearly they want to win the eval game and the eval tweet game that's out there. It gets attention, like it clearly gets attention, right? Where you are on the evals gets attention. It's very valuable. It's worth going on Twitter for Zuck. My view is that, my Captain Obvious view is everyone's going to be token maxing if they aren't already. Everyone's going to hit the limits. I finally hit my Claude limits yesterday for the first time on my Max plan, and I had to decide what to do, right? So everyone, and as every organization goes from burn all the tokens you want, baby, to budgets, however you do it, whether it's automated, whether it's manual, whether it's picking a dropdown, everyone's going to have to have a cheaper model. It may be in your product for cheaper workflows, it may be for coding. So everyone, so at the moment this seems like a battle for the cheap seats.
And it's a big battle for the cheap seats, and we'll see what happens. We don't talk about it, but I think Haiku from Anthropic is a pretty damn good product, and in one day they could make it better just by adjusting the slider of how much Opus you get in Haiku. I literally use a massive amount of Haiku, and for like cheap simple stuff it's really good, and it's a tenth of a cent, right? And so there's a battle for this second bucket. How much of the margin is there? How much of the revenue? Who wants to win in it? I don't know. So I think it's interesting, but it appears to be a battle in the frontier bucket, but I think it's a battle in the B-tier bucket where there's still a lot of volume, but how much margin of revenue is TBD.
I think that's correct. Not quite. But on the other, first, agree with you, and I think the Databricks paper and a couple other papers that we'll talk about have talked about this concept of buckets and tiers of model. And I like what you're saying, Jason, you're exactly right. When the pressure on spend is out there, the sentence you said that really resonated with me is every company is going to have an internal tiering. They might all buy it from OpenAI or Anthropic, but every company with a CIO who's half awake is going to have a cheap token model to hand to stop this madness.
And to some extent that means there's a slot for that. A fun, interesting one would be, you guys know that obviously Meta did that contract, you know, they're starting to offer their capacity now just like SpaceX on a, you know, short-term basis. It'll be fun to figure out at what price. You know, it's not clear to me that they would make more money selling at these token prices. They might actually make more money renting their capacity out. So it'll be interesting to think about compute to the highest user. You know, we'll actually have that discussion again when we come to Groq/SpaceX, but you know, this is an aggressively priced product, and you know, good on them. You're right, Jason. It's pushing at the low end.
Well, we'll probably play it by ear. I mean, if this is more successful, then they get more capacity. If it's less, then they give it to their competitors. I mean, they'll probably play it by ear.
We mentioned the Databricks paper. It got a lot of attention. Can we actually stay on that? Can you summarize what happened and why it's important?
Some of the stuff is obvious. I read it all and trying to, the big picture at the start was the basic point is, you know, cost per token is really not a useful metric. It's kind of cost per completed task was the first thing. How much does it cost to get things done, right? And because one of the ahas is some of these things that have low cost per token, their reasoning tokens are more expensive than their basic tokens, and then you don't know how many they're going to use. So there's a whole bunch of things. So the first picture point was cost per completed task is really the thing you have to assess. Then the second point they made, which was interesting, is you end up with a Pareto curve for different kinds of tasks, which is the best model for this kind. And this is Jason's point. There will be a big brainiac frontier set of tasks for which you use them. Once you quantify the cost per task, then the second thing you're going to have is different tasks. Different models are more cost-effective at different tasks. And a quote cheap model that just requires infinite reasoning mightn't be as good as paying up for the more expensive model. So you can imagine this Pareto curve of value. And then the third point, broadly speaking, was the harness you use. In other words, your infrastructure around that massively impacts kind of how efficient that all happens. So it's quite a lot of just...
Data, Databricks. Yeah, they did. Exactly. Yeah. I mean, I'm shocked. Exactly. I'm shocked to discover that a company that effectively helps you manage models says that managing models is important. But to be fair, they produce the goods, Jason, in terms of, you know, facts to back it up. And it makes sense. And I think moving away from, because every time someone announces a new model, they kind of give you the cost per million input and output tokens. I think the emphasis on cost per completed task, it's just, I can totally see that making sense if I'm the CIO. And again, so it's in tune with what Jason was talking about.
I like the post this morning from Aaron Katz at ClickHouse, who said that their AI spend is up 60x since February.
Oh yeah, did you see this? It's a post. Yeah.
And then he said, 'AI is fundamentally changing what is expected from a data platform. We can't build the best data platform for AI if we don't deeply understand it. We're investing in building a team that's truly leading in understanding and innovating with AI.' 60x.
Well, I mean, once you do anything complicated, it's easy to go up 60x, right? Instead of doing one design for one page of your website, have Claude design, redesign your entire website in the background. That'll blow your token budget in five minutes for the whole month. It'll be gone. These really complex workflows consume a lot of tokens, and even as they get cheaper, the workflows, the outputs are getting richer and richer and more and more complicated. So I think, listen, I mean, I don't know how much time you want to spend on this. I think not only is token maxing, you know, gone from niche to interesting, right? Every developer can consume more tokens. The mediocre developers can do it to be performative and pretend they're working even though none of their PRs are accepted. And the great ones literally can run 10 or 20 agents 24/7.
I mean, if you talk to the best developers now that are AI-pilled, and I hate that term, okay, they are coding 24 hours a day now, okay? They are because, and it's addictive. It's like video games, okay? Because now I can build 10 features, 100 features. Now I can do things I couldn't, I can do so quickly. So it's 10:00 at night, I'm checking, instead of doom scrolling, I'm doom coding, right? And I'm going all night, and I wake up in the morning and I want to check in on that workflow, and instead of checking in on one feature, what if it could have built all my features, right? And so I don't know any, and again I hate the term, I don't know any AI developer that couldn't consume even more frontier-level tokens. We just want more, right? You can do, and nothing else, you can do more prime versions of whatever you're working on. You can just do more versions. It saves you time and it's much better. Why do I have to choose A, B, or C? Just build them all and oh look, right? And then build A prime, B prime, and C prime, right? And then let's actually go further. Let's put them all up on production, on staging or the dev server. Let's see how they all work. You know what? Let's go further. Let's run a million test cases against it, right? Oh, well that will cost $20,000 in tokens. Okay, well I won't. I'll narrow down the... What if I could? It would be better coverage, right? You can consume, any top-tier developer can consume an order of magnitude more tokens than they are now. The ideas are limitless.
Agreed. And I think I want to stay on this one for a while because I think what's interesting about that, Jason, there's never been a product, and this is why this is the management challenge. There's never been this product before which on a totally on-demand basis can allow for an individual worker with no cost to them can basically do their work for them and make them look amazing, right? Yet the company's earning 100. If you think about it, if you're a developer and you're sitting there going, I could crank this thing for the next two weeks myself or I can press the magic button and somewhere out there a whole bunch of expense would accrue to my company, but it's not been taxed on me and it can get me a long way down. I'm going to do that if I'm an intern. I mean, you know, summarizing references in our shop this morning, I'm going to get the magic Claude summary, right? And why not, right? And to be clear, sometimes that's going to be wildly value-accretive to the company because you saved an hour of someone's time that you're paying, you know, 500 bucks and you paid 20 bucks. You're ecstatic they did that. But as Jason points out, if there's no governor on that, eventually you're going to hit the moment where you spend 600 bucks to save 500 bucks. And it's, and all what's going on right now is how do people get to grips with that?
Yeah. Yeah. Let me just, I sort of hit, but let me just give you a simple, even if it's very basic. I know some folks will make fun of me that watch this, right? Or listen to it. But it took me a while. Now I'm all in on Claude Design. It got really good. Okay. It took me for a while when it launched. I wasn't sure if it was a play feature. I love it now. I think it's gotten better. But as folks said, and it took me a while to get it, it consumes more tokens than anything else I do. It is massive. Okay. And so yesterday for my first message, I'm like five minutes in. It's like, you got to wait till 3 p.m., Jason. You got to wait till 3 p.m. to use your credits. And that's to do one, basically redesign one page in essence because it uses so many. What I would love to do instead, imagine this. Here's my entire website for any product. Okay. It's 100 pages, whatever it is. Every night, come up with better versions every night. So I'll wake up in the morning and I'll doomscroll through your ideas from my website and I'll pick the two or three of the hundred that are better instead of burning, like think about it. I burnt through my tokens in 20 minutes building one page. How many tokens would it create for every night to redo a hundred, right? Help me with the math, Rory. It's a lot of, it's orders of magnitude more tokens, right? So I think literally every product person, every designer could do that and consume 100x tokens they do now, easily without even blinking.
Does it change your Figma usage?
I'm not, listen, clearly I don't think so. But what I think, not to go down a rat hole, what I do think, Claude Design, if they keep investing in it like they have, and we could figure out the massive token consumption, is that like a lot of, like you can get further without a designer or you can get further with half a designer or an outsourced designer or something because you can build your whole design schema and build, people are, you're going to make fun of it. You can still smell a design, a Claude Design app. You can see the elements, but you can get pretty damn far. It's pretty smart on the product side of how to do it. And so it doesn't mean you don't need a designer, but it may mean that we don't need that crappy designer anymore to get us to a V1, right? There's just no need to have a... So does that harm Figma? No. I think in the age of AI, every time you start losing the bottom of your market as a co, I'd be nervous. It won't show up in that quarter's numbers because it's the bottom of the market. It's the ankle biters. Like, you've already gone up market. Figma's tracking million-dollar deals, right? Who the hell cares about the low-end one-seat deal, you know, it lost to Claude Design. But over time, those ankle biters move up the leg. You know, they're the piranhas. Then they take over the knees and the thighs, and pretty soon you retain the big customers, but you've lost a portion of your funnel, right? If you lose the bottom, you lost a portion of your funnel. They don't all wait and say, 'You know what? I'll graduate to Figma when I'm bigger.' And this is where Salesforce is at risk. People like, I love Salesforce, run it headless, but the next generation of folks may not graduate to Salesforce because they never started there. They started agentically. They may never graduate there. If I'm a CEO, I'm at risk that folks won't graduate into anymore. Workflow tools are at risk that even if just 10%, 20% less folks graduate into, it leads to a death spiral of slow growth. That's why, you know, for public companies, the only thing that really matters to me is are they growing net new logos 15% or more a year. If they are, they'll figure it out, right? If you're growing your net new customers 15, even 20% a year, but as soon as it falls, man, you could, I mean, you can raise price every year, but it only takes you so far. So rambly answer is it does worry me, but it's not a threat to this sort of enterprise Figma play at all. But that's if I were building a startup, say, that's right.
Rory, what point did you think was more interesting that I didn't mention or glossed over? You said...
No, actually it wasn't the point they made. It just, yeah, I'm kind of, things you find yourself puzzling on as you think about the market. I'm just looking at the level of traction on these companies right now, which has just been astonishing, right? I mean, the Anthropic explosion to what now might be 50 billion from 9 billion at the start of the year, right? And you know, I saw the SemiAnalysis breakdown, and yeah, two or three billion of that is quote-unquote Claude Code, but the real truth is I think a huge portion of the API part is code as well, right? And I've just been wondering, you know, when something's growing that quickly, momentum, yeah, I say what could cause it to change? And I think that, you know, there's really only two things. One is mass competition, which we've been talking about in other, if everybody dive-bombed the price and it went down by a factor of two or three, yeah, 70%, the kind of the low-end token max pricing you're seeing here, is that likely? And I don't know, I don't know if that's going to happen. So then, yeah, if the momentum continues, what's the constraint? And one of the things you figure out in all these kind of at very late-stage growth investing is in the end the TAM is the criteria. And what has that just been? You start to look at the dollars being spent here relative to the size of software engineering, and it's pretty huge. Remember we talked about is it going to be 20% of software engineering, right? And kind of to Jason's thing, right, everyone goes, oh, 30 million developers, it's all fine. If you go to the BLS, which I did last weekend, and say there were only 1.88 million developers in the United States, of which only 200,000 of them work in software companies. Another 600,000 work in tech companies, you know, HP, IBM, and a million of them work in literally JP Morgan, BofA, right? And you know, that's a total spend, median wage is about 140k for that role. It's a total spend of around $250 billion, right? And if most of this OpenAI and Anthropic enterprise web, if a large percentage of that is software, they really are close to 20% already.
Part of me says, oh my god, what's happening? Because there hasn't been 20% of layoffs for these companies. It's just, it's a huge, my point is this. It's a huge number relative to the total software coding spend in the US. And I freely admit I don't have a conclusion from that. I'm looking at it going, hm, right, is it excess revenue? I read the SemiAnalysis stuff. I don't think it is. I think a lot of it is in the US. Now, there's some purchasing in the US for...
use in overseas, but it's just an astonishingly high number relative even to the biggest number you can think of, which is the wages being spent on software engineering. And it'll be interesting to see how that shapes out in the next 12, 20, 12 months.
It is a great analysis because it sort of, you know, it is definitely a bear case that no matter how great these tools are, we may soon hit the ceiling in terms of how much companies, just on an absolute sense, are willing to spend here. There is a ceiling. You can't spend more revenue than you take in, and that's if you have 100% gross margins, right? There are just physical limits. Even if you reallocate budget to it, there are just limits.
You're right, Jason. And you know, and I almost don't like to posit it because one of our rules of thumb is this. I always joke, simplistically put, Newton's laws of motion are true. Things in motion stay in motion. Things growing at 10x year on year might decline to 8x or 6x or 4x, but they don't come to a grinding stop, right? And the amazing thing is when you're growing at 10x, even 6x next year and 4x the year after that is a huge number. I mean, if you're at 50, 60 bill this year in GAAP revenue and you 6x next year, that's 300 billion, right? That's more than the total wage spent on software development, right? So you look at it and go, one of two things is going to happen in the next year. Either this thing decelerates faster than anything you've ever seen decelerate. And I don't think that's option A. Option B, to Jason's point, a whole bunch of CFOs are literally going, we used to spend 4 million on tech, all of it in salaries, and now we spend 8 million on tech, 4 million of it on salaries. WTF, we got to do something, right? So it's just, as I say, no conclusion. And I'm just things I'm thinking about on July 15th. It'll be interesting to see. And you're right, Jason. There are final, I mean, it may well be that the prize for becoming the fastest growing company in human history is you may hit TAM faster than any other company in human recorded history. And it'll be interesting to see. You may have hit the limits of how much money there is in the till.
You know, there's just one other vector just to add it up for fun. And again, I don't, mock me if folks want to. So there's how much can we spend on engineering, right? And then there's another math you can do. Okay. Every piece of software is becoming agentic very, very rapidly. Okay. So what's software spend? A trillion and a half. You're better at the numbers.
A trillion. So we can't, I think we're coming to the rough conclusion that folks may tolerate, say, a 10% gross margin spend on AI. Let's just assume that for a minute, right? That's another 140 billion that can go to Anthropic and friends, right? So that's significantly more than the math you just did, but it itself has a ceiling, right? It can't be all of the 1.4 trillion. 10% feels about right today. There's 140. All the software spend is going to put 10% of their topline into tokens.
Agreed. First of all, I totally agree because I mean you asked that question a while back and I got the answer wrong because I was an idiot and then I processed, right, which was how much does it cost to run my little sales and marketing agent and it was way less than you thought, right? And the reason this is relevant here is this. It turns out that outside of coding, most of the token costs for agentic software outside of coding are manageable. To your point, they're 10%. They're not 50. Right. So, you're right, Jason, is that if the software industry is doing a trillion in other kind of non-coding stuff, that's another 100 billion right there. You're right. Like Salesforce could comfortably pay 10% of its revenue for tokens from someone. They're not going to pay 40 because they only have 22% operating.
That's 100 billion of revenue that's almost accessible now. And it's, I mean, it's all going to agentic, right? So that, we, I need someone smarter than me, but that could be a hu-, I don't know what percent of Anthropic fuel is, but it's a lot. That 100 billion.
No, you're right. There's really three buckets. There's the coding bucket. You're right. Then there's the agentic 10% of software tax. Just like Amazon took a tax on software 10 years ago. Everyone's going to run on Amazon. Give me 7%. Now you're right. Everyone's going to run using LLM. Give me 10%. That's 100 billion more. And then the last thing you start talking is does co-work replace the knowledge worker at 20 bucks like Microsoft Office. But you're right. I mean the big picture and I'm still wrestling through and trying to look at those numbers. I'll report back. But you're definitely at the stage where you're talking big ass numbers relative to everyone else's numbers.
Is there movement into legal, biosciences, not just an appreciation of the fact that they are potentially hitting TAM limits in where they are already?
I want to say something potential. There's no evidence that they are, even though logically, again, I want, because I don't want someone to say, hey, the hidden town, all the traction momentum seems to say they're not. I just look at the actual market and I go, huh? Right. I genuinely, I personally think on the medical side, I think oddly enough, I don't know if that's as much as I think the founders of Anthropic, just like the founder of DeepMind, is very much motivated on the medical side by, you know, the desire to, everyone almost references the world where cancer is cured when they're talking about their LLM. So I think actually people's interest in doing life science is one-third TAM expansion and two-thirds, I'm putting everyone out of a job. The least I can do is keep them alive. I want to do good because I want to have meaning. It's a search for meaning. And you know, starting with DeepMind and going from there.
I mean, if you don't think the other founders wish they had a Nobel Prize like DeepMind, like Demis, you're crazy. That must be good. That must feel good.
Jason's going to get a Nobel Peace Prize for everyone's not going to get a Nobel Prize with his companies. Okay.
I'm waiting just to get an offer, a decent offer from any offer from Anthropic. I haven't gotten a single one. No one's offered to bring me in as a chief anything.
It's a joke. What would your price be, Jason?
After tax, it's not worth it otherwise. Plus, I don't really want to work for the man. So that'd be the, if I really wanted to do it, 200 would be the minimum entry price.
I strongly need to say to you, Jason, that based on my understanding of the Anthropic structure, you in fact don't work for the man, you work for the woman, right? Daniela runs everything.
It's a gender-neutral term, I would say, but for take the job and take, yeah.
200. There we go. We have.
I don't think they actually can risk it. I might not work out for five years there. There's no choice I don't take.
There's no other way of saying this without being hurtful.
I don't think you'll pass the personality test. I mean, I know I wouldn't, right? But they're doing good, caring, all that stuff. I think you'd just flunk out.
Yeah, but some people think I'm a little driven, Rory, even at this point in my career.
Yeah, you're driven, but they filter out for cynicism. You just won't make it. You're just not idealistic enough. Look, I haven't gotten the offer. So empirically, you're correct. I do want to make sure we.
I've done 63 of these podcasts. I've written 10,000 blog posts. I haven't gotten one offer from Anthropic. It's just not coming, is it?
It's age discrimination. You're right. The infrastructure that powers foundation models, SK Hynix, NASDAQ listing. You want to go there, Rory?
Great. I mean, yeah, great. I'm suggested topic these days, Jason, because he just scowls and I'm like, oh, fuck.
No, no, no. I thought that it's a winner.
I mean, you know, look, obviously the stuff everyone knows. There are three memory companies. They have been a huge beneficiary of the AI capex boom in the last year. Those stocks, SK Hynix is up 6x, right? Two of the three are based in Korea, Samsung and SK Hynix. Micron is based in Idaho. Go America. And the truth is it's an oligopoly. They've made out like bandits and it's actually very hard to buy stocks in Korea, which is why even today, interestingly enough, after the ADR is traded in the US stock exchange, it's still trading at a 20% premium to the same stock in Korea because it's just hard for an American to open a brokerage account in Korea to buy the Korean things. So I think it's just good that these guys are accessing the capital markets.
Next NASDAQ listing, largest ever by a foreign company. Had a great reaction, popped 13% at close of day.
And then back down. And then, I mean, look, had a very difficult day, I think, yesterday and the last couple of days, right? And yeah, it would be tempting to go to some kind of, oh my god, the AI capex bubble is busting, and at some point it will. But I actually think a lot of that, there's a lot of weird technical stuff going on there because the two, no surprise, the two memory companies dominate the Korean stock market. They are the two companies, Samsung and Hynix, are 60% of the stock market and roughly 120% of the GDP in terms of market cap. They're just huge, right? They're highly volatile. There's lots of retail action in Korea. So there's just a lot of intraday volatility in the Korean stock exchange. It's all kind of crazy casino-level wealth there, right? So even though the stocks moved down in the last couple of days, it's not as clear. Is it fundamentally a DRAM correction or is it just weird kind of trading? But I think on the DRAM thing, I mean, the fun bet there is, I don't have an opinion on it, but those three companies are all trading at, you know, 5 to 8 PEs. In other words, dirt cheap. And that's the bull case, and this is going to last. And the bear case is these have traditionally been capital cyclical businesses. And the minute more capacity comes online, prices have gone up 6x in the last year, which is why operating margins have gone from negative in '23 to like 70%. The most profitable companies on the planet. Samsung is the most profitable company on the planet right now. And they're going to go back down. So that's the kind of pro and con. I think it'll last a little longer than the typical cycle, but in the end it'll, I mean, at some point it does correct. You're not going to have the memory makers earning 70% net margins, not gross margins, net margins forever.
One thing, maybe you'll challenge the connection. I thought it was interesting that, you know, IBM this week, pretty old tech company, right? I think founded in the 1800s or something like that. I mean, they had a huge miss, stock crashed 20%. And some of it, listen, is probably an excuse. We have to make up excuses as companies when we have a huge miss. But they blamed memory. They said memory is taking so much of the CIO's budget. There is such a scramble to buy memory before it goes up. They weren't buying our mainframes or our servers or anything from us. Like CIOs went into a panic to buy memory before it would get even more expensive. And I'm sure it's an excuse, but I'm sure there's a germ of truth that that money's got to come from somewhere, right? And IBM said it came out of us.
Yeah. I think you're right, Jason. I actually don't, I think there's more than a germ. Look, all money has to come from somewhere, right? And if you're spending a lot on LLM and then if you're spending and you're having to spend even more on your memory needs because the LLMs themselves are bidding up the price of memory, you don't want to be the last item on the purchase order list on the last day of the quarter because I'm sure there's a CFO sitting there going, you know, what can we cut because we only have a finite capex budget this quarter. And it's what you said also about what's going to happen to the modular SaaS vendor. It's not that they're going to vibe code it away. It's as much as, you know, guys, we have a tech budget of 10 million and last year we spent 100 grand in AI and this year we're spending 3 million, so we need 2.9 million in savings.
And it turns out we don't need five productivity apps. You'll love Teams when you get to know it.
And maybe we'll stick with what we have too rather than buy that module from you.
Yeah. So yeah, I mean, and IBM did tumble pretty drastically today. I'm going to go off on a bit of a tangent, Rory, because you said, oh, there's a world outside of this like AI. Me and Jason were messaging earlier this week about Jason Calacanis's syndicate or investing activity where he basically said, hey, due to popular demand, I'm moving where we'll be investing from very early to later stage growth opportunities in some of the big names. Is this symbolic of where the venture market is at today and a sign of the times? Jason, given it was us talking about it, why don't we start with you so we can.
I just thought it was interesting for him to say that because actually, you know, Jason's got some pretty good investments. He does a lot of investments, right? So he's got some pretty good ones. You know, I thought the message would be we're adding an annex fund, right? Or we're going to do a little bit more here. But to say that he has put so much energy over decades, right, into creating an alternative path to do massive number of syndicated investments to apparently tilting all into growth, you know, it is, it's in a niche, it's a huge sign of the times, right? It's a huge sign of times rather than easing into it. But completely switching when you've got multiple billion-dollar winners that you have material ownership in is, to me, it's a big deal, right? I mean, seeds for suckers, as we've known for 60-something weeks, this is the, other than YC dumping early stage, that would be a real sign, right? YC issuing a, going on Threads and telling everybody we're only doing growth rounds, that would rock the world. But I thought it was a bit more, I thought it was at least a sign of the times.
Everything's up. And when everything can exit north of 20 billion, you know, it's, I even, me, like I have an investment with a hundred million dollar position that took me years and years and years to get to right now. And then just watching a growth round where someone comes in and invests 100 million, like, what, why did I bother? All the drama, all the years, all the being the only guy at the board meeting when someone just drops a $100 million check. If that can grow an order of magnitude, what's the point of being the guy any earlier than there, right? No existential point.
Because you love the craft.
I don't love the craft. Who the hell loves the craft?
David Frankel, I interviewed him yesterday.
He does love, that man loves the craft. More than he loves the craft so much.
Replace me with him on this pod. That man will be doing deals from the old folks home, right? And he'll be getting good ones. And he'll be getting good ones. I would love, I love David.
So do I. Please. That was me defending his love of the craft. I was with him and I was like, loves the craft.
