Anthropic Overtakes OpenAI, IPO Wave Looms and Europe’s AI Moment

0:00Lomax Ward:OpenAI is looking at Q4 or Q1 next year at an$850 billion valuation, but their cash burn is$14 billion. So they're looking at a cash flow break even in 2030. I guess the IPO forced some kind of transparency on that. Obviously, we can't not talk about the Iran ceasefire and how maybe that's going to cause a medium-term catastrophe for us. So short-term, lots of volatility. Longer-term, obviously, who knows? I think we've got this medium-term hit. I also saw this week that Global Ventures hit$300 billion for Q1, which is 70 % of the whole of last year's deployment. And that's four companies absorbing 188 billion of that.

0:31Lomax Ward:Obviously, we know the concentration of investment has been massive. And with AI, it's only going to continue. So loads going on that we're not going to dig into. But was there anything in that that you wanted to poke in the eye? Iran sees fire, maybe. I mean, that story is going to continue to play out over the days and weeks. Jamie Dimon referred to it in his corporately left. It's kind of baked in, isn't it? Well, no, it's not baked in, but we'll talk about it later in the show. But I think inflation is the big watchword here, which then pushes up interest rates, which then might mean a rotation out of equities.

0:57So we would see 2002 all over again.

1:00Lomax Ward:Hello and welcome to Upside, where every week we look behind the headlines affecting European venture. Today it is Lomax and my good self. And we're going to dig into Anthropic Mythos and the European Invasion. We've got JPM, JP Morgan. They've got their annual letter out. What does Jamie Dimon usually get right and wrong? We're going to dig into that. 2026 mega IPOs. But who has half a trillion in cash? Where's the money coming from? Climate and sustainability. Investing is back, but we can't call it that. Germany's 10 billion defence spend nightmare. OpenAI buys some friends, loving that, and predictions plus deals of the week.

1:51Lomax Ward:This show is not investment advice and the hosts of this episode may be invested in the funds and companies featured. Lo Maxius, how are you, my dude? Good, man. We're flying quite, I feel very naked this week. We're flying, not flying solo. Flying duo. Flying very lightweight. Lightweight aqua. Yeah, we've got no Andrew. We've got no guests. We've got no Mads. So we've got no, so just me and you. So we, but we've also tried to fill the docket with things that are in our wheelhouse. So I, I would like to start with some things that I haven't put in the show, but I've caught my, caught my eye this week and see if anything is worth talking about and see if anything resonates.

2:33Lomax Ward:So to start with, obviously, we can't not talk about the Iran ceasefire and how maybe that's going to what I think cause a medium term catastrophe for us. So short term, lots of volatility, stuff's going to happen. Longer term, obviously, who knows? But I think we've got this medium term hit. So I don't know if there's anything in that. I also saw this week that Global Ventures hit 300 billion for Q1, which is 70 percent of the whole of last year's deployment. And that's four companies absorbing 188 billion of that. Obviously, we know the concentration of investment has been massive. And with AI, it's only going to continue.

3:07Lomax Ward:But I don't know if there's anything in that you wanted to speak about. I saw lots of Euro cyber stuff happening, coming through the ranks. I don't know about you, but I'm seeing a lot more cybersecurity startups come through, which I think is understandable. Russian hackers logging into people's routers this week, I saw. Meta reinventing his AI strategy again. Poor old Zuckerberg. and a Brit, Adam Back, denies that he is Satoshi Nakamoto, the Bitcoin inventor. So loads going on that we're not going to dig into. But was there anything in that that you wanted to poke in the eye? Iran ceasefire, maybe.

3:40I mean, that story is going to continue to play out over the days and weeks. Jamie Dimon referred to it in his quarterly letter. It's kind of baked in, isn't it? Well, no, it's not baked in, but we'll talk about it, I guess, in a little bit when we talk about that later in the show. But I think inflation is the big watchword. here, which then pushes up interest rates, which then might mean a rotation out of equities. So that we would see 2002 all over again from a kind of flight to capital from the equity markets, both public and private. So that's the one thing to watch. But at the moment, you wouldn't know it because, as you said, in the private markets and the public markets, your equities are pumping.

4:18And as you rightly point out, global venture funding is absolutely soaring. I mean, there's very much a power law um at play here it's driven by a few companies and predominantly open ai and anthropic that make up like the large vast proportion of the of the numbers here so and

4:34Lomax Ward:both listing this year right well this year open ai might spill over into next year but they're going to be fighting to get out first to market probably um cyber security again like well we're going to talk about this in the context of the the release the release the recent release of mythos by by Anthropic. But of course, I don't know what's true or what's not true in that, but the story is, have you read all about Sandwich Gate and how it all happened? It's quite, mythos is quite, if it is true, it's quite high-propping and quite scary for sure. And then as for Satoshi, Satoshi Nakamoto, I mean, yeah, there's a new, there's a new button every month now.

5:13Lomax Ward:Yeah, the 55th. Yeah, yeah. I thought it was interesting that it was a Brit. So I can imagine some poor crypto guy in his basement going, what? It's not me, dude. What are you talking about? A few Brits get wheeled out. Disproportionately number of Brits get wheeled out as Satoshi Nakamoto suspects, right? I think. But it's not exactly a slow news week and there's loads going on. I don't know why. I don't know how that came about or what the finger pointing was. But yeah, poor Adam. But if it's not him, if it's not you, Adam, I feel for you, dude. Shall we crack on with Anthropic? Because they seemed to have dominated the news cycle this week.

5:47Lomax Ward:So let me give a bit of context. Numbers first. So run rate has hit 30 billion. That's tripled in six months now ahead of OpenAI, which is estimated at around 25 billion. But a quick note, you've heard me say this before. I'm in incessant nag about it. Watch out when the media transposes ARR for annualized run rates. There's a lot of PR going on in there. Anthropic monetizes at eight times OpenAI's rate per user, 80 % enterprise and aiming for breakeven by 2027. OpenAI are looking to get cash flow breakeven by 2030, but at four times the cost. What I found interesting when I was looking into some of the Anthropic news stories was that Europe is really on the map.

6:33Lomax Ward:EMEA is really on the map for Anthropic. So it's their fastest growing region. Revenue's up 9x, large European accounts up 10x. They've already got offices all over Europe. and when I first saw Sadiq Khan saying, please come to like, I just found it all a bit silly and a bit sad, but there might be something in that, which is part of my prediction later on. And then Mythos, we were just talking about it, this incredible new powerful model. It's blown all of the benchmarks away. It's a general model, but they're focusing on cybersecurity and they've only got, I think it's 40 people. What's the project called?

7:11Lomax Ward:Project Glasswing, I think it is. they've only got 40 organizations in the inside track of testing this new model because it's so bloody powerful. Trump called them left-wing nut jobs. They've been designated. I saw that a judge didn't block the designation of them being a supply chain risk, so that moniker still stands. They've also just acquired coefficient bio, which I think you wanted to talk a little bit about, Lomax. An amazing, amazing story there. Yeah. Yeah. And they're also looking to invest to 200 million into a billion dollar private equity fund with the who's who in private equity, GA Blackstone, H &F.

7:54Lomax Ward:So there's a lot, there's a whole raft of anthropic news this week. What, which bits or where would you poke? I mean, lots and lots of interesting things there. Yeah. I mean, look, I mean, the numbers are, numbers now surpassed OpenAI. So 30 billion run rate, as you say although open ai is a 25 billion but it depends if you annualize their weekly revenue you might also be at 30 billion but let's say they're kind of neck and neck now which means anthropic has caught up open ai quite significantly right so so the pentagon stuff either hasn't bitten yet which is my assumption no i don't think no no no i think i think that exactly i think there's the breadth of the customer based on that they're serving more than you know there's more than a thousand enterprise customers now spending over a million annually with Anthropic.

8:37That number's doubled since February alone.

8:40Lomax Ward:And they're still being used by the Pentagon, right? I mean, they are still the vendor of choice. They can't just rip them out. No. And eight of the Fortune 10 are clients. Anthropic holds 32 % of the enterprise LLM API market versus OpenAI is 25%. And interestingly, the monetization, Anthropic is monetizing it roughly at$211 per monthly user versus OpenAI at$25 per weekly user. an 8x difference in monetization efficiency. So I think that's... Have you seen OpenAI's? We'll talk a bit more about them later, but they've just launched a bunch of shops and their streaming service. Okay. So we're going to talk a bit about them later.

9:18Lomax Ward:They were saying one week we need to focus and then what feels like a massive lack of focus. And then the TVP and acquisition, which we'll talk about in a minute. But I think let's go back to the headline, the headline, which is the revenue figure, because I think we've been talking about this a lot, which is to justify the trillions and trillions of capex that is being spent you need to well it's certainly hundreds of billions no it's six to eight hundred billion this year isn't it the capex exactly across the ground you need you know you need to believe that the revenue like ai revenue and therefore ai gross profit and contribution from ai can reach the trillion you know let's say one trillion total in revenue and actually only a year ago anthropic was doing a billion arr or they're now doing 30 they were they were doing 10 they've gone up 3x in q1 alone so if they did 3x in q2 you know then they're at 90 then they do 3x again then you stop these numbers you know start to get very very big very very quickly well it has to the story this year has to marry to the IPO story because if SpaceX, Anthropic and OpenAI all go out, let's say 26, maybe Q1, 27 for OpenAI, there needs to be somewhere between what 200 billion and half a trillion in it, whatever, I don't know what the float's going to be for the IPOs, but there needs to be enough cash in the system to be able to justify the valuations.

10:50That's a capital markets question, which we can maybe talk about in a little bit but the the general just point is if you look at the revenue ramp that is starting to justify the payback um or to justify the the investment in the capex and the equity of these companies so i think you think they'll do it i i think we're going to see some stalling if the ipo market window opens i think they'll do i mean look at this growth rate and

11:14Lomax Ward:it's insane it's insane you know but don't you feel like it's capped it just feels it feels like we are going to hit a wall. It doesn't feel like it's going to be, I mean, we've been talking about this with NVIDIA specifically, haven't we? It's like every quarter they break their numbers and every time we look at Anthropics numbers, they're doing better than anyone could ever imagine. So it just feels like there's a wall. It feels like there's a wall coming. Well, I don't know. The penetration is still very, very low. And you can look at the growth of their experience in Europe. It started in the US, it's in Europe, but there are corporates all over the world.

11:45I really think there's still a long way to go here. One thing to note is that on the gross profit side, gross margin side, that's actually coming in at about 40 % rather than 50%. So I guess they're going to spend the next six months trying to move that up ahead of the IPO. There's definitely a kind of so hot right now aspect to Anthropic versus OpenAI. If you look at the secondary market, and remember, so let's say that we just talked about the revenue being, run rate revenue being more or less the same. Anthropic's got actually a higher growth rate. But if you look at the valuations of the companies, and according to the last fundraising round, the OpenAI last valued at$850 billion.

12:21Anthropic at$380. So actually, OpenAI is more than double the valuation of Anthropic. If you look at the secondary markets, though, the demand for OpenAI has come off a cliff. And that's actually trading at a discount now to that headline valuation. Is it? There's over$2 billion of demand in private markets for Anthropic stocks. So Anthropic is trading in private markets, from what I can read, at a massive premium to the latest fundraising round. So the valuations are converging, that's for sure. Anthropic is having a major, major moment. So it's funny how these play out from week to week because two weeks ago when Trump was slamming them and calling them supply chain risk, then suddenly you felt that this might send shockwaves through the buyer universe.

13:12Lomax Ward:But is that inertia still yet to hit, do you think? Do you think, how much of this is there? I think Anthropic is positioning itself nicely. It's way more enterprise heavy than OpenAI. It's found it's positioning much, I think, more cleanly in the market than OpenAI is still. And even despite efforts to focus and centralize around a certain proposition, OpenAI still likes to break out and do things that contradict that. Whereas I think Anthropic has a much cleaner narrative and equity story, quite frankly, from an investor perspective. And to be honest, I think it's got a better product. Well, it has.

13:52Lomax Ward:But how much is a question for you? How much do you think is values, ethics and alignment? Because there's a big story coming out about Sam Altman this week in The New York Times. there's obviously almost the anti-Trump brigade going, okay, fine, if you're going to call them a supply chain risk, then that's the reason we will work with them. And I wonder what is values, ethics, alignment? I think it's hugely important. And I think it's playing into Anthropik's hands at the moment. If you look at Sadiq Khan, bless him, courting Anthropik in London. But certainly for a European, looking at what's going on in the US, anthropic being bashed by the trump administration for drawing a line at being used in autonomous killing systems and in surveillance of your own population that resonates pretty strongly with the european buyers i would have thought and also probably european talent so if you want to come to europe build offices and they've got now offices in a bunch of tier one cities now hire good people probably going to have a better job of it than than open ai potentially so So I think there's going to be many twists and turns in this narrative yet, but certainly Anthropic has the upper hand.

15:05And then I guess we move to Mythos, which is this general purpose frontier model with strong agentic coding and reasoning skills that they put out this week with a focus on cybersecurity.

15:15Lomax Ward:And it's not incremental, is it? I don't think this is incremental. I'm actually looking at, I think there was one thing I was reading, which is very, very scary, which is in cybersecurity, you have a vulnerability and then you have an exploit. So an exploit is something that exploits that vulnerability. So if you have a vulnerability that's sitting there in a code base, but then an exploit is actually an action that exploits that vulnerability. And if you look at the historic models, Sonic 4.6 had a 4 % exploit to vulnerability ratio. So of 4 % of the vulnerabilities that it found, it could exploit 4 % of them.

15:48Opus 4.6 had a 14.4 % exploit to vulnerability ratio. and Mythos has a 72.4 % ratio. Yeah, I was looking at the benchmark.

16:00Lomax Ward:I was looking at even some of the weaker benchmarks and they're like 20 points ahead. It was like quite crazy how impressive this model is. It is. I think it's scary. And I think definitely cyber security is going to be a major, major thing on the landscape for all of us, both offensive and defensive. Well, I read yesterday that AI-driven cyber attacks were up 1 ,500%. So I don't know. that that was a kind of a skim read as i flew past an article but even if it's part true that's quite scary yeah and then if we move to health and life science but you talked about the acquisition of coefficient bio so this week anthropic announced the acquisition of coefficient bio for 400 million in stock this was a company in anthropic stock this was a company that was only founded last year in new york did you see it only had six employees no no so coefficient was a um incubation by a vc fund called dimension so dimension capital is a u.s fund um i know a couple of the gps there only founded in 2021 spin out from lux capital in the u.s and london and in new york the deep tech fund and dimension focuses actually on this very area which is the intersection of of biology and ai and and software and technology so they're kind of there's what what the kind of it's a little bit inside baseball but people in the industry called tech bio right this kind of new age people in hoodies doing biotech basically so you replace the old school dudes in the lab coats and the funders wearing the twee jackets and you're everyone's wearing a hoodie and just hope they don't move fast and break things right yeah and as you know one of my friends calls it It's basically kids do bio effectively.

17:39So it's applying, you know, mood, fast, break things kind of methodologies to bio and uses a lot more software and AI in your discovery processes. So, I mean, there's definitely, there's huge companies like Recursion and the US Exegenta in the UK that have been built off the back of this. And we're only just at the beginning of that in terms of the inception.

17:57Lomax Ward:This is 400 million in stock. I mean, this is a big ticket. Big ticket. So actually, it's a really, really interesting, you know, if you want to kind of geek out on the details of it is that Dimension incubated this company. They pulled together three executives out of Royvon, Evozyne and Biogen in the US, pulled them together and worked on the company, which was working again at the intersection of technology and biology. And I think Dimension put in a$10 million check and they own just over 50 % of the business. So they're going to get the$400 million. They're going to gain$205 million in adropic stock, which we know is now like doubling.

18:34Lomax Ward:That's not venture. That's venture building, right? Yeah, proper venture building. Looking at it here, Dimension reported a 38 ,000 IRR on the investment from last year. So I think if they turned$10 million cash into the company to$204 million in paper and anthropic at the current anthropic valuation, I believe. if dropout valuation doubles or triples in the next like six to nine months which it really should do and we just talked about that in the context that it's valued less than half of open ai probably dimension is sitting on a something like a 60x return there and they're fun i think here's a here's a fun fact that i was going to share later with you was that none of the 2025 major ipos increased in value they all lost value when they went public yeah well i mean that's the story exactly but so first of all it's a great it's a great venture story actually the guys at the guys at dimension came out of luxe capital which itself very very highly respected deep tech fund in new york has done a lot of very successful incubations over the years but um it's a really really cool story and it also shows that anthropic is really doubling down and not willing to you know cut back or skimp when it comes to what did they buy in this space what I'm assuming this is an act or higher, right?

20:00They basically bought a team. Honestly, they pay for it in stock. So it's like it's$400 million in stock of a company that's valued at$400 billion. It's like the dilution is like nothing. And what is the dilution? 0.1 %?

20:14Lomax Ward:It's just funny to think in those monetary terms being contextually so small. And also the context here is that, so as a reminder, healthcare spend is roughly 20 % of US GDP. It's like probably the biggest standalone sector in the whole US and almost globally. And Anthropic recently, you know, they released Anthropic for Healthcare and Anthropic for Life Sciences at the end of last year, beginning of this year. So, you know, they're going and leaning big into the space. So I think it's set against that context. Well, I think, you know, they've got such, such momentum, Anthropic. I mean, there are more stories that we haven't gone into, but I mean, it was almost a week full of anthropic stuff.

20:59Lomax Ward:So very, very interesting. It was the fun thing I would say, because we have talked on talking, talking about my own companies here. We backed Bios, which is working in vertical AI for clinical trials that raised the biggest Series A that Iberia has ever seen. I think at the end of beginning of this year, 52 and a half million is scaling incredibly quickly. But Pedro, the founder there, put out a LinkedIn post this week saying he gets asked every single day, like, won't Anthropik just take your business? Like, they'll just crush you. And he, you know, clearly that is a risk for the business, but, or it's a theoretical risk for the business.

21:31But ultimately, when he tests his model and his product on the problem that they're working on, on clinical and planning clinical trials, it is infinitely better. and of course that is a highly regulated space people worry about rappers on these llms and and how they're ultimately going to get eaten and crushed but when you've got the rapper actually operating in a very specific like highly regulated domain with a lot of relevant context that's only but actually the general llms do not perform anywhere near as well right so i think it's um it's just something to bear in mind so anyway well that's also that's also part of europe's

22:12Lomax Ward:moat isn't it the regulation piece i think there is something in a little layer of protection for for those operating in europe and understanding regulatory environment you know products that that actually take that into account and can deal with it is for sure that they have a right to exist let's talk about uh jamie diamond jp morgan ceo he released his annual shareholder letter on monday It's a big letter, 48 pages, against the backdrop of hedge funds making their biggest bets ever against European stocks. So they're shorting Europe. Every year he lays out what he thinks is coming. But, and we're going to go into some of the content anyway, just quite naturally.

22:55Lomax Ward:But I wanted to specifically look at historically what he gets right and historically what he gets wrong and see if there was a lens there to look at this current letter. What he normally gets right is the structural stuff. So the multi-year call structural stuff. He warned about the US fiscal deficits from 2013, said nobody would act, and then it became a crisis. Debt has more than doubled. He flagged the fintech disruption years before Wall Street or anybody took it seriously. He warned about the health and maturity accounting that gave banks a false security on interest rate risk. And that was, I would argue, the biggest reason why Silicon Valley Bank died.

23:30Lomax Ward:Where he has historically got things wrong is timing and dooming. So in 22, he warned of an economic hurricane. By October, he was then warning of recession and obviously didn't come. He was then warning of recession for four consecutive years. The economy has been resilient and has defied him every time. One investor says his track record on leading the bank is incredible. His track record on economic calamity predictions, not so good. Although it's probably that's the best way to be, because I think everyone forgets, which is part of the reason why I wanted to have a look at his record. As with everything in our industry, it's all timing and luck.

24:09Lomax Ward:So maybe it is just a piece of timing and he will be absolutely bang on the nail. But one thing which you've already mentioned, which I think he is absolutely on point with, is inflation. He calls it the skunk at the party. The Iran energy shock, meaning stickier inflation, high interest rates and supply chain reshaping, which I think is going to be, to me, that when I was at the front of the show talking about the mid-term issue. I don't think it's going to hit it straight away, but this is going to be super problematic, I think. When the straits were closed, oil spiked to$120 and banks postponed rate cuts, which I think is the first ripple.

24:49Lomax Ward:This current ceasefire, which by the time this pod goes out, might all be over. I know we've got Israel-Lebanon, so who knows what is going on as we're recording. But that is going to complicate the picture. But markets are surging, right? So three other things that he did mention. He mentioned private credit. He warned that the$1.8 trillion market lacks transparency. That's going to be problematic. If it tightens, it's going to hit venture debt. PPC is already in trouble. That's global. But I think we're going to see more in private credit. It's going to be problematic. On the trade side, he's basically gone against Trump.