E719 | Europe Is Writing the Cheques. The System Still Doesn’t Work.

0:00Lomax Ward:Hello and welcome to Upside, where every week we look behind the headlines affecting European venture. Today, the premium, best, supremo VCs in Europe are with me. It is Andrew and Lomax and my good self. And we're talking about what's on the show this week. We've got a quick news roundup, seeing what's on people's minds. We've got Mistral raising 800 mil, but why? The UK defence tech brain drain, looking at that. Paul Side loses core weave, which is a slightly bigger topic than it sounds, potentially. NHS staff are boycotting Palantir, unpacking that one. Some predictions and deals of the week.

0:50Lomax Ward:This show is not investment advice, and the hosts of this episode may be invested in the funds and companies featured. now gents how are you both andrew how are you we haven't seen you for a gazillion years what's new been busy busy portfolio raising uh the third fund still happening so yeah how you get on with the raise how's all that it's moving it's moving good i'm not gonna i'm not gonna put a date on it because that'd be dumb are you are you finding it um i was gonna say easier i know it's not Are you finally getting more fluid now? I was going to say that's an oxymoron, but yeah, I think things are moving.

1:29The sector that we're in, given we're resilience, which is all related to defence, definitely helps. Obviously quite topical. Things can't ever stay stationary forever, but it's still a tough market, right? The private equity market still hasn't loosened up, but let's hope the big SpaceX IPO encourages others to get on the bandwagon.

1:46Lomax Ward:we um i was saying to somebody yesterday i was saying for the first time in my life i'm sexy because we were doing mad and i were doing whatever ml industry 4.0 10 years ago like about as sexy as a phony dog poo now we're in the zeitgeist and like you know everyone is talking about resilience like with you andrew and physical ai and manufacturing and on-shoring french-shoring reshoring all this stuff all the spaces that we're in and i thought oh bugger everybody's in now it's like we didn't do any of the other zeitgeisty nonsense all the other hypey stuff but now we're in it so um now i'm now i'm fearful that every other bugger's in it well you were you were a pop star once i mean you were definitely cool at one point and then and now you're a cool now you're cool again dan so isn't that isn't i didn't do anything cool in pop stardom but isn't that isn't this the normal route into ventures that's what you both did i think by the way i think it's a very point good point that you raise is that if you um you build a a brand a network a thesis a portfolio around a certain vertical and that vertical suddenly becomes in vogue it can be both like good and bad for you i think you know like good in the sense there's more capital for follow-on for syndicates for follow-on deals bad of course It has created more conversations.

3:05Lomax Ward:Do you know what? Yeah. We have had more LP conversations because they've always known what we've done. They've either liked us or not and so what? But then because we've not jumped, I don't know, from blockchain to AI to defect, we haven't. We've always done the same thing. It has loosened the elastic a bit and it's created, I don't know if I've just got happy ears, but it has created more conversations, which I think is good. But then there's a lot more competition in the space. And also the Iran conflict and everything else is, and fears of inflation is making people sit on their hands a bit more.

3:41Lomax Ward:So it's countering that. But I think there is also is, for people who allocate regularly to venture capital across a number of different funds, when a category matures to the point that it almost becomes a box that a venture capital allocator wants to allocate to, you could take deep tech, for example. if you go back to 2015, 2016, most funder funds or institutionalized family offices or pension funds that allocated routinely to venture would rarely have an allocation to say deep tech, right? They might have an allocation to biotech and then the rest would be what they would call ICT probably, you know, like IT or that would be, that's very old school, but that's how they would look at it at some point in a i don't know when you guys think this but from about nine from about 2019 2020 21 the fund of funds would actually have an allocation to say deep tech for example there is a recognized category and which made it easier to sell to them because it was something that they actually wanted probably came into conversations wanting to buy the problem that you also have with all of this stuff is though that the people who have the distribution capability already like the big funds can spin up products in those areas yes and they have the brand cachet yeah they have the ir team they have the trust they have everything and so when a16z sets up a defense fund a deep tech fund a gaming fund you know suddenly you despite having done it for 10 years you just get blown out of the water anyway it becomes an irrelevant so that's the it's always the trick in venture is you can you can be very vertical specific but even when your vertical kind of starts to form part of the zeitgeist it doesn't actually necessarily fix anything for you you know yeah no it's true it's true i i i just i'd rather be on that side of the fence than the old side of the fence where we were so i have to see the positive in that upside 100 completely 100 so um i think i think it's better it's better to have been early and then caught the guys okay now i'm sexy again yeah i'm gonna go for it yeah and also when you you know when you talk to founders and you've you know you've been allocating in in that domain for the last five ten years it's a world of difference for that conversation that you have them with them versus a generalist gp who's starting to allocate and the network you have as well we have instant corporate network into those spaces because we've been doing it a decade.

6:17Lomax Ward:So I'm going to take it as a positive, but you're absolutely right. It's a bit disconcerting to also be... It's never as big a positive as you think it would be, right? There's always more nuance to it. True, true. I've written down some stuff that caught my eye. First being the Artemis mission, obviously heading back to the moon at 93, Bill. Maybe there's something you want to talk about there, Andrew. I think Trump declaring victory and then not declaring victory. I don't know if you guys saw his 20 minutes. Oh, let's stick to something. Let's start with something positive, at least. I mean, going to the moon again, how cool is that?

6:53It's just, by any measure, it's cool. The rocket's the size of Big Ben. It's 250 ,000 miles.

6:59Lomax Ward:This is the biggest rocket ever, right? This is bigger than Starship. Isn't this bigger than anything that's gone so far? It might be tall. I don't think it will eventually be bigger than the largest Starship. I think Starship will be bigger. But it's pretty close. By any dimensions, it's fantastical, right? They'll be the fastest humans in the last 50 years because nobody else has been to the moon. I just think it's, isn't it that incredible? It's half a century since we've actually, humans have been to the moon, which is just, it's over half a century, so it's mind boggling. But the most important thing is they won't have to drink their own pee because unlike, because it's only a short mission.

7:36It's 10, 11 days.

7:36Lomax Ward:I'm sure there were some filters in the way. they'll just they'll just vent it to space unlike if you're on the International Space Station you have to you know you have to you have to all goes around all goes around but but there are other parts that are recycled the engines I think the rockets the rocket engines have launched 22 shuttle missions before they launched this mission so a bit of circular economy in there as well so yeah I just think I just think it's amazing and I think I posted on LinkedIn isn't that old tech? yeah I mean it's a big mix of new and old tech which is kind of cool as well so I I posted on LinkedIn yesterday that, you know, if you have kids or even younger siblings, you just sit them down and make them watch it and explain it to them.

8:14That thing is taking off and going to the moon and coming back.

8:18Lomax Ward:But it wasn't that big a news story, was it? I mean, it got covered. It did get covered live on most of the main news channels. Yeah, I watched it live at 11.25, 11.30 last night. Maybe I'm just head in sand. I think when they land on the moon, which is the next mission, that will probably get much larger coverage. But it's cool either way. And this is to build a base, right? This isn't just to go and come back. This is to scout out building a base, which I think is going to be interesting. That's the medium-term plan. Well, interestingly, I mean, yeah, actually, by the way, they're going to do in the middle.

8:49This is Artemis 2. Artemis 3, the next one, was going to be Man on the Moon, but actually it's going to be Artemis 4 now in 2028. Things are slipping a little bit.

8:59Lomax Ward:Well, they have to make sure that all the kit works because there were some problems with the bits that were made by all the different manufacturers. Well, by the way, many of which whom, you know, if you want to add a European element to this is that the Europeans have been like quite major suppliers to the Artemis program. That's right. Which I guess is both good and bad because, you know, it kind of represents our place in the world as just part of the supply chain. But we are a critical part of it. So Europe did provide this, what's called the European Service Module, which powers Orion spacecraft.

9:34So propulsion electrical systems. Unless Trump affixed tariffs on it in between. Of course, yeah. Well, then he, well, yeah. But then I guess NASA, which is a government agency, will just have to pay more if he's shooting himself in the foot. But the interesting thing, which actually does negatively potentially impact Europe, and it shows you what happens when you are just a supplier, is this week maybe you saw the Lunar Gateway was effectively cancelled. The Lunar Gateway is a part of the Artemis program, which is to build an equivalent of the International Space Agency Station, which currently orbits the Earth, to build that which orbits Moon.

10:13And the Europeans, including Tales and I think Airbus, had been critical suppliers of that and actually sunk a lot of money into that project already. And then it was cancelled this week. Why? Do we know why? postponed slash canceled in favor of actually building um and focusing resource and attention on actually building a kind of physical lunar base on the moon rather than having the um the satellite in orbit it's not a reverse in the program if anything it's a it's a an embellishment or enhancement but um the europeans are actually pretty pretty involved in the lunar gateway and now uh maybe they're going to lose a bunch of business off the back of it and have had a bunch of wasted work um everything they've done that's a weird thing to do have have a satellite around the moon what's the point in that well that was that was supposed to be the kind of key like look you know launch beachhead or launch pad to go in and for lunar missions right the other things that

11:09Lomax Ward:caught my eye this week apple turned 50 uh software stocks so the s &p down 20 year to date that's uh quite an interesting drop spacex how can we not talk about spacex and the moon filed for potentially the largest IPO in history. It's 75B, with a B, at 1.75T trillion, which I think is just absolutely... Valuation's gone up, no? Gone up from 1.5 trillion. I'm trying to decipher what's kind of Muskian bluff and bravado and actually what's reality. I think the latest number was 1.75, but I'm sure by the time this goes... The stock market is mostly bluff and bravado, isn't it? Anyway, I mean, it's all sentiment driven, so, you know, horses for courses.

11:50Lomax Ward:Somebody's got to accept that figure. At some point, there's going to be dollar bills changing hands in the public markets. Intel, interestingly, look like they're turning the ship, the new CEO there. They struck a$14 billion deal to buy half, I don't know why half, half of what they sold to Apollo, the Irish chip plant. So I don't know what's going to happen there, if that's meaningful. Mistral raised$830 million in debt, which I also do think is meaningful, to build AI data centers across Europe. poolside lost its core weave compute deal which we're going to dig into later deep seek i don't know if you guys saw this deep seek had a seven hour outage on sunday did you see that i'm not sure how meaningful this is i mean vendors are going to go down and have problems but well aws has it the whole time yeah but not seven hours that that felt quite uh quite seismic um i wonder if people will then challenge this kind of service from those kinds of vendors.

12:47Lomax Ward:NHS staff are boycotting Palantir. No great surprise. No surprise there, no. Sentiment wise, but I don't know if there's going to be any knock-ons there. Starcloud raised 170 million. We'll talk about that later. I'll pick you up on Palantir. Go, go, go. I'll pick you up on Palantir. I mean, I think this is going to be the tip of the iceberg going forward. And it's what happens when you don't have sovereign capability yourself. I mean, the things the staff have been complaining about, my understanding, is not only was it sort of impinged on them top down with very significant consulting, it's the fact that it's an American company, an American defense company.

13:25You know, it's the same story about all the guardrails and what happens to that data. You know, probably the largest data set of health data in one single place in the world, even though it's sort of distributed across the NHS, but one organization at least. And the fact that a British company is not doing something with that and isn't funded to do that, you know, this is what happens if you get behind the curve on this technology. And this will continue now across every industry, across every sector, across every part of government, unless we decide to catch up and compete in the race and be an AI power in the same way that we're a nuclear power.

13:58Lomax Ward:Well, do you think we can? I mean, yeah, throw money at the problem. I mean, we were four years late with the hydrogen bomb. We still got there, right? But just to give this some context, so Palantir won this contract in 2023. And I think the timing of that is very important. And it looks like it's a blank check up to around 450 million, which is a big chunk of change. And as you were talking about, Andrew, this kind of trust gap thing that has been exposed now. So the NHS staff making the news cycle is a big thing. So this is going to be a point of conversation for other public sector workers.

14:36Lomax Ward:I have no doubt. I'm trying to work out where the opportunity is, if there is anything for UK health tech founders, whether there is somebody that could potentially take the place of Palantir. Is there even looking beyond the UK? Is there a European challenger that could fill this gap? So they were the questions. I don't know if either of you had any reflections on those questions. if you think about hospitals in general you know the big platform systems that have been put in place in the us is epic and cerner and they've started to infiltrate the european market i personally think there's a huge opportunity to build an epic or cerner for europe and or the uk are there any kind of contenders on well i've backed one it's a precede there you go talk your book baby come on yeah yeah um they're a great team out of italy so i do think there's a big opportunity there it's going to require you know savvy founders moving far well moving fast but also moving fast and probably not breaking things in this sense you know selling into healthcare is actually moving fast and steadily and carefully probably because it's heavily regulated and there's a massive trust-based component so the sales cycle from a startup perspective will be very very long so you need to have founders who build that around the company around that and will be need to be very very good at fundraising for sure and taking people on the journey and building trust with b you know it's a b2g sale at the end of the day often it's not always b2g because you know there's a lot of private providers in europe too but i think that's a huge opportunity how close to your italian team to actually being a vendor they're already a vendor um but they're at the moment for specific point you know with with the early product but the plan is of course to expand over time but they're very very early stage i mean this would be a journey where they need to raise i mean there are companies that have are building great businesses um i mean our curex who i advised very very early on when they first started pivoted trying to solve the problem of of uh antibiotics being prescribed but they identified an issue of communications and administration problems between um you know patients and and the clinics they started with gps i think they've got 98 coverage of gps in the uk 350 ,000 NHS users I think like over 50 million patients um being supported and they're actually expanding the other way to the US so they've done extremely well so it is possible to and in their case they went sort of they went bottom up which is a fantastic kind of viral viral approach and they got funded by Tomiko so it is possible to build these businesses but you know the the founders out there with the ambition it's the same old story you've got to be able to raise the cash especially when you're talking about these much larger platforms.

17:16You know, Palantir has been around a long time, and especially when you're talking about going forward to AI. So that's the challenge. Founders will go where they need to survive. And until we plug that capital gap, they'll go where the money is. Today, if you want to build a billion-dollar revenue health tech company in Europe, it's basically impossible in Europe. You have to go to the U.S. As a reminder, health is 20 % of US GDP. It's an insane number. So the proportion is much higher than any other country in Europe. And the absolute number is multiples of what we have in Europe. So if you want to build a generational health tech company, basically, you have to go to the US.

18:05And we've seen that, you know, Acurex going there now. There's a company called Anima Health, which started in the UK, which is now doing incredibly well in the US. There's tons of examples. Obviously, Hinge Health came out of the UK. Sword Health came out of here in Portugal. And they're now, I think, 350, 400 million dollar revenue companies. But taking this back. So I do think if Europe wants to benefit from sovereign systems for its hospitals, and obviously these are, because there's a lot of state sponsorship of these systems, unlike in the US, like these are a massive component of the increased amount that it costs to run a country in Europe.

18:47Lomax Ward:I saw a few American, was it Matt Metner from Ex-Sequoia? He was knocking this idea of sovereign tech, calling it welfare, which I thought was quite interesting. Well, I mean, but even that, maybe, you know, the word sovereign is being thrown around a bit too much these days. But if you want to build homegrown companies that you feel proud of and actually deliver value for your economy, then you need to encourage them and give them the environment in which to flourish. just maybe stepping back to this palantir thing there's a broader question for palantir i think like can they as a business have a go-to-market in both defense and in healthcare to um actually potentially quite contradictory things because interestingly this problem also happened in new york so it's not just the kind of it's not just a sovereign it's not just a sovereignty anti-american thing within within the new york um healthcare system they are not going to renew palantir contracts for for a similar reason right so so that's probably a little bit politicized i'm sure as well but the point is they have that question this is also the question which also could could take place in any in any country any in any company is to what extent does it matter and should you listen to your employees as to either who you do business with or who your suppliers are you know remember this issue came up if they're the ones on the front line if they're the ones using the system well and actually this came up in if you if you remember in a slightly different context but this came up in 2018 2019 in the u.s when the big tech companies were winning contracts with um the u.s defect with the pentagon right do you remember the hololens contract that microsoft had in 2019 the 480 million dollar contract they had with the pentagon and all the microsoft employees, a section of them, a cohort of them complained about it.

20:45And I think Satya Nadella said to them, actually, guys, we're going to do this contract. You can be reallocated to another account if you want. The same thing happened at Google in 2018. And in that case, I think that Sundar actually listened to the employees and did not renew the contract with the Pentagon. So there is a kind of broader question as to how much you as a company have listened to your employees about who you sell to, or in this case in the NHS, who you buy from. And I do think, it as you say dan like they are people on the front line you would be stupid not to take their opinions into account they're the ones having to plug the sensitive data into this software and or use it i do think part of this is is just like yeah as i said the kind of palantir specific issue is like going to market in into markets i mean you've got the bloody chairman of of palantir peter teal i think is quoted on record as saying that the nhs is actually better at making people sick than making them better, which is, well, you've got the chairman of a company for a big client saying things like that, then that's, I guess, a very palantir specific thing.

21:46But at the end of the day, we may well see these things happen more and more across Europe. And the answer is, goes back to what Andrew was saying, is if you have systems that have more trust, they're more likely to win contracts and do a good job, and that's better for everyone. Sometimes we worry about the octopus effect. People talked about Goldman Sachs being the big octopus and people talk about palantir like taking over the world you know there is actually an ability you know that for this to be resisted and for companies to build brands and go to markets around trust which therein leaves an opportunity right for people which is the european opportunity right this is my amazing italian team to become the epic of europe or whatever it may be you know and there's there's and andrew's quoted some you know regulation compliance and and

22:36Lomax Ward:understanding the culture and market in europe has to be part of this new european moat for sure for sure it's all part of like building a brand and a and a and a sales process and everything especially in these markets yeah okay well it was very interesting which is slightly a side point but it's like the palantir platform was only scheduled to deliver 150 million in cost savings until 2030 but when you put that in the context of the whole nhs budget it's like it's piffling it's yeah yeah i'm not suggesting that one trust one annual budget right out of 240 they were looking to say even not even i would have thought but like i i am i mean i'm not sure i'm not suggesting that a european company would do any better but i'm just like the numbers are so big for these healthcare systems.

23:26Anyway. And it's a big contract.

23:28Lomax Ward:Although that said, I mean, let's say it is half a bill in five years or four years. Maybe, I don't know. I don't know. Maybe that's not massive beans for not massive savings. $100 billion a year on the NHS, yeah. Yeah, I don't know. I'm kind of pleased that this whole maxing maxim is starting to die. The London maxing, Europe maxing, retard maxing, that seems to be dying a death, which I'm silently pleased about. Anthropic leaked some source code for clawed code. I don't know what the latest is on that, but not massively European, but I thought that was quite an interesting thing that caught my eye this week.

24:07Lomax Ward:And then the bigger item I wanted to start with was Mistral raising$830 million to build NVIDIA-powered AI data centers in Europe. This feels like a big deal. Every day I open the press, it looks like Mistral has got a new business model. They are always doing new things, but it seems to be quite successful. They seem to be doing well, which I think is great. They're now not competing with OpenAI. They're more competing with AWS. They're going up against the hyperscalers on European soil, building out our own AI infrastructure against the kind of contextual backdrop of$765 billion worth globally being earmarked for build-out.

24:49Lomax Ward:So the questions of the room are, does sovereign AI infrastructure work? Should we be thinking like this? Is it a protected market? Is it just expensive nationalism? Going back to Matt Miller's response from Sequoia, ex-Sequoia. Will founders just use whatever is available compute-wise? Is Mistral becoming a utility, which I don't think is a bad strategy, or anything else on your minds to do with this Mistral big fundraiser? It's debt, by the way. It's not equity, it's debt. The compute base, if you have your own data centers, is lower, right? I mean, so in terms of staying in the game, I mean, it's survival cash, right, compared to what their competitors are raising.

25:24They only raise all the money they can. By European standards, it's big.

25:28Lomax Ward:By global, not. You're right. Yeah, so there's a cost reduction, I suspect, and compute control, not relying on third parties all positive as far as misdrafts concerned but will it be enough we'll see depends how many crazy laws Germany puts in place right about sort of having to use they're loosening I was reading all the all the stuff what was it called the Mertz's China China didn't they just mandate open document format I think I don't know Lomax you never open document formats no I think so I thought there was something. It was just mandates there for Dr. Voight. But it feels like they are talking a good game.

26:10Lomax Ward:They are opening. They're thinking more broadly. They're thinking smart, I think. Who knows what will kind of come down to our level and where rubble hit road, but it feels like they're doing stuff. Max, what else would you pick up on the Mistral? I think Mistral is doing the right thing here generally, like integrating vertically on this critical piece and probably de-risking it as well from a supply chain perspective and just having a little bit more control over its own ecosystem and future. I mean, I would just say more generally without trying to be a downer because I do think it's great that Mistral is having a go because no one else is doing this in the UK, which is that they're doing great, but everything about Mistral is just like a sort of 1 % of what happens in the US.

26:57How much have they raised? They've raised a billion. you know and open ai is raised over 100 billion like what's the valuation like 10 billion versus 800 billion well the valuation is normally what open ai raises right and of course well not even and then and then well it would probably know that open ai raises 3x evaluation and then what's the revenue well the revenue is potentially it's it's minuscule projected to be a billion by the end of the year well you know open ai run rate potentially now is 30 billion it's probably even more by the time we speak so everything that happens with mistral is just a little kind of small um asterisk in in the um but we have to start somewhere right but you have to start somewhere um this is a pretty big debt package in european terms you know 800 was 850 830 million dollars um and actually it was announced now the data center is going to be operational in q2 2026 i.e we're in q2 2026 now so So, you know, they've actually managed to get on with it.

27:57At least, I mean, we'll see whether they can actually execute because we'll probably see in the context of poolside in a minute, this stuff is actually harder or easier said than done. And I think it's a very interesting development, to be honest. I think it's testament probably to the ambition that Mistral has. And I think it's, you know, for the equity early investors, it's great. It's debt financing. It's non-dilutive. is nvidia part of the uh debt financing structure do you know well they're definitely a part of the equity you know i don't know if they're part of the debt that would be a weird thing it's good it would be weird for nvidia to be the debt will be you know debt providers no i mean i think it's generally probably just generally like their decision to use debt rather than equity reflects a sort of maturing of the ai of ai industry where it is today i mean i know we've seen a lot of it's

28:43Lomax Ward:a big bet though it's it's a big bet to put this out put you put it out on a limb like that in debt not not equity or not even split this is yeah but i mean you know core weaver's raised 7.5 billion in debt lambda labs 1.5 billion so you know they're not again the numbers in the us are suspect is the usual suspects right society general like i don't know it's probably the banks probably yeah of course it will probably like you know some clubs and e-support black so nesk providers um but i mean it's it's pretty i i think actually it's pretty safe bet i mean the trajectory of ai unless mr r completely screw up their customer base and growth is is only going one way so we forget in our little bubble the way we use ai that most people aren't so um you know i think it's a safe bet although if you have to depreciate the um the chips more quickly then um it becomes less of a safe bet as the debt provider i would have thought but yes if they're if they're moving into utility space they will be too big to fail so the french government will become will become owners yeah well this is why i thought the comment earlier about you know matt's comment that um so sovereign welfare was i mean it's bs really when you look at what the u.s buys and the requirements they have and the constraints across government for how they purchase and who they purchase from you know they're very much by america and and they do a very good line in sort of counter counter argument when it's anybody else who puts up walls or puts up free you know any free trade barriers and it's always been this way so the idea that the UK shouldn't have sectors, particularly defence, but beyond that where mission critical and anything to do with infrastructure resilience and security should not buy things it has control over and from companies it has, the British.

30:27But it was funding as well. It's anathema, right?