E712 | Nvidia’s $1T AI Bet, Hyperscaler Risk & The Real AI Battleground | Upside

25:10They're not even making money on AI. They don't even really have an AI business. Okay. So what's interesting here is as you look into 26 and 27, we are seeing negative free cash flow for the hyperscalers. And that, of course, changes the narrative because, well, it might not be 2000, but it's certainly not 2024 either. Alphabet recently issued 100-year bonds. Yeah. Right? So whole new stories are coming up. And this is some of the thing that I think you go to Niklas Tangen from the Norwegian Sovereign Wealth Fund. He says, I'm a huge fan of the AI. It's massively increasing our productivity. I believe in this, but I look at these numbers and the leverage and what's happening.

25:55And as an investor, it worries me. Yeah, I've always been of the mind that as soon as this level of credit goes in, this level of debt is being,

26:06Lomax Ward:I saw, was it, who was it raised a 20,$25 billion bond? Was it Salesforce? I think it was Salesforce had a 20,$25 billion bond sale. Nomax, which aspect of this would you pick up on? So I think it's an interesting point. Like Mads is totally right. And the head of Norwegian Sovereign Wealth Fund has a massive point. It's kind of classic. He's basically saying this is a sort of late cycle. People are so bullish and enthusiastic. They're almost mispricing actually some quite structural risks that are now creeping in into the market. Well, he was focused on the Iran war. He's using this as kind of the warning point.

26:49He was focused on the inflation that will come out of the Iran war, which is a good point. And remember, when you have inflation, you have rates going up, you have a flight from equities into debt instruments, right? Because they're paying higher coupons. So that will potentially cause a rotation, which will then soften valuations. And then Mads has just articulated it very well. Ultimately, Amazon's now running free cash flow negative for the first time, like probably ever, since it actually became a big company. Well, since the 90s. Yeah. So, I mean, this is now the time. and the debt on these projects is only higher and higher.

27:32I am sorry.

27:34Lomax Ward:You know, we should be wary. We should watch. Yeah, we should watch. I mean, I don't know what this means for, well, we do know what it means for European venture, but I don't know as it cascades through our worlds, what this means at this level. I don't know what other things we could uncover and discuss. It just feels like something we should be mindful of. Be careful. These companies used to be the greatest cash machines ever built, right? And now they're borrowing money to plow into data. Because they're running out of, you know, haven't got the free cash flow anymore. The only kind of interesting observation, which maybe you guys picked up on, is that Apple is basically sitting out this huge capex.

28:19I love that graph.

28:21Lomax Ward:The graph of all the crazy hyperscalers and then Apple just bumbling along. They're still doing their share buybacks. They took a big license out of Gemini. They're quietly strengthening their own chips for their own devices. Actually, pretty impressive performance. The M5 now delivers four times the performance of the M4. Actually, Apple quietly continues to operate in the old manner. Maybe it's the right strategy. Maybe it's not. I was looking at all the on-device with the M5. Pro-Up chipset or the on-device AI. It feels like an infrastructural tee-up for running on-device algorithms. I think there's something in.

29:06Lomax Ward:As you know, I'm a big Apple fan. I've always thought the Siri will one day become useful. And it feels like there's this kind of network play that's going on with this new chipset. Yeah, it is genuinely fascinating. And I'd say the other side of the coin, because it's easy to be a bit doom and gloom. We were talking about the Swan Memo, Black Swan Memo a couple of weeks ago, right? And you write one of those and you look smart if it comes to pass. And if it doesn't come to pass, everybody forgets about it. Now, Bill Gurley is somebody I think we all respect a lot. And he talks about these bubbles and bubbles in the past.

29:40And they're always, you know, because there is something good, but then people get greedy and in over the skis. And the question is, is that where we are now? The truth is, revenue growth has been extraordinary. That's the big difference between now and the dot-com boom. I mean, we've seen how open AI and Anthropic have grown revenue and all the others. It's just phenomenal. And yes, it's still not enough to justify the full CapEx build-out, but at the growth rates we have, is it conceivable that one could get there? And you probably don't need revenue to grow much more than five-fold from where we are now for things to start hanging together.

30:18And so fivefold, does that sound crazy? Well, we know what Anthropic did last year. You know, the industry in a couple of years will probably be five times the size it is now. And if it then grows another five times after that, well, suddenly this starts to become a really, really big market. And I think we're just all looking at it and saying, could it grow five times from here? Yeah, that feels eminently plausible given how impactful This is five times more after that. So that's 25 times from where we are now. Maybe that's toppy. But there is certainly a scenario that you can paint that actually this is a justified CapEx buildout.

30:54This is justified investment because of the size of the opportunity in front of us. So I think the question is, will the revenue grow fast enough to catch up before the market loses patience and says to the hyperscalers, guys, we're going to slash your equity price in half. The cost of raising this money is going to be so expensive now that you can no longer justify the massive CapEx build out. And I think you go back to the big, I think the big D word is depreciation. Because I think like, if you build your data center, you incur the CapEx for that, you buy the chips, and then you need to replace them.

31:31If you need to replace them within, I don't know, 24 months or 12 months now, if it compresses from the current of companies, suddenly the revenue you need to generate to pay back both the initial capex and then the reinvestments you need to make to rip out and replace your chipsets, suddenly you need to be making a lot more revenue to justify and pay back the capex. If that happens, you will have incinerated an incredible amount of capital. the question is whether you will have built a business with some staying power. Is being a hyperscaler a moat in the future? Going back 15 years, before we knew what a hyperscaler was, when cloud became a thing, people were looking at Amazon and they were looking at Google and they were saying, is this a real business?

32:20You used to just, right? And it turned out that was an incredible business. So are we seeing this new business here as an embryonic thing? Now, I think it is a genuine question. And I think the people that are asking the question are generally right to do so. It is by no means a foregone conclusion. But I don't think you can write it off either.

32:40Lomax Ward:Well, it leads very neatly into our next topic, which is effectively along the lines of what can AI do? What is it doing right now? And how far will it go? What will the adoption cycle be like? So we're understandably seeing mixed messages across the market around this AI capability to adoption piece. This week, we saw this lovely story about a Sydney tech entrepreneur who treated his dog's cancer and he has zero medical training. He was using chat GPT to brainstorm a treatment plan, AlphaFold to create the protein structures and a university lab to manufacture a custom mRNA cancer vaccine for the dying dog.

33:21Lomax Ward:and the tumor shrank 75%. Hopefully he's still alive at that time of release. It cost him$3 ,000 and took two months. Same time, we saw DeepMind this week published a paper admitting the industry can't agree on how to measure AGI. So they've proposed a 10-dimension framework and are now paying researchers$200 ,000 to help build the tests. Also this week, we saw Deloitte found that 5 % of SMEs that are using AI are seeing no value. so we've got this kind of we've got all of these i know there's a bit of a disparate story set here but we're seeing lots of ai usage adoption uh i guess the smartest lab in the world can't define what intelligence is 95 of small businesses can't really use ai we're going to see that play out in 2026 this this feels like the story from from creation to adoption that we're going to see this year uh lomax first up in your world tell us tell us more about the dog and the cancer everyone loves the like a companion animal story um and you know first of all so first of all okay big picture this is an amazing thing so there's like there's obviously been a lot of hyperbole in the press around this and on social media about how the citizen scientist use ai to basically treat and cure his dog's cancer, right?

34:43First of all, it's incredible that those tools exist today, that he can even, because the dog, let's just be clear, that the dog's cancer is not cured, right? The dog is still going to die of cancer. It's probably had its life extended by two to three months, which, you know, actually, like quite a few cancer treatments when you get into the, actually, that's all they do. They aren't necessarily curative, right? So it's amazing that this person has done it, but, you know, let's put it in perspective. The dog is not completely cured. It's going to die of cancer.

35:12Lomax Ward:To be candid, he took the dog from a rescue home knowing that this poor creature had cancer and was just going to look after it in its last day. So this was a known thing. But this is incredible that he can do this, right? Because if he can do this with, you know, a few hours a day and a few grand, And imagine what the big, well-funded institutions can do. And this quietly is happening already, right? It just takes time because clearly, you know, to get these therapies into patients, you need to run clinical trials. And they'd cost a lot of time and they cost a lot of money and take a lot of time, like hundreds of millions and, you know, one, two, three years potentially.

35:59Now, what's really cool, I think, is this is the mRNA technology, which is the technology kind of popularized by the COVID vaccines, because obviously Moderna in the US and BioNTech in Europe, which is one of the biggest success stories in European tech that a lot of people often forget, had this ability to create these mRNA vaccines, which you can create in silico, like in a computer, pretty rapidly, right? When you sequence the tumor, you get the data from that, and then you run a bunch of bioinformatics and machine learning and AI on that to find out specific antigens that you want to go after on specific proteins.

36:38And then you can code the mRNA and then inject the person to train their immune system to fight these specific cancer cells. That was seen with COVID already. That's phenomenal. That is getting better and better and better. So like AI is gonna play a huge role in that. So I think it's a wonderful story. I mean, by the way, there are no approved mRNA vaccines yet, right? There are some in late stage phase three, in the last stage of clinical trials. So we still have, you know, to get this done at mass scale to actually affect human lives takes time. But it's going to happen. Absolutely. And it's very exciting.

37:19Lomax Ward:When I read the story, it made me think about personalized medicine. It made me think, obviously, every body is different. and made me think, well, how far are we away from a very, very personalized medical? We're not that far away. I mean, you know, there can be complexities, right? Because if you have someone with a tumor and the tumor is very deep inside their body, you need to actually extract the cells from that, right? And that can be hard. You know, in my case, the tumor was removed and then we sequenced it and then we managed to run the analysis on that, right? But for a lot of people, you find, especially if you do all these preventative MRI screens now, you might find tumors buried in your liver or in your pancreas and then you need to actually extract that doing an invasive biopsy, which is possible.

37:58But all of this stuff costs a lot of, takes a lot of time and costs a lot of money and puts a lot of pressure on healthcare systems. And if you're in the game of only increasing someone's lifespan by a few months, then people who pay the bills, the insurance companies or the governments might be like, well, it's just not worth it. Because as you said, these are personalized. You then need to do the sequencing and you then need to actually make the individual vaccine. It's not like COVID where once you've made it, You can just manufacture it at mass scale and just give everyone the same vaccine, right?

38:28But we are going to get there despite all of the roadblocks that I've just talked about. And it's very exciting. And by the time we get to our kids' lives, cancer will be predominantly, 90 % of cancers will be dealt with in the same way that we deal with a common cold. That's a big, bold one. If you listen to the dudes in Silicon Valley, they'll say all disease will be cured by 2030. I mean, that's clearly absolute rubbish. But this, if you take cancer, for example, which is obviously one of the biggest, after cardiovascular disease is the biggest killer of people, we will get to that position.

39:04It's just going to be like another 20, 30 years out. You know, it'll be a buzz out of our kid. I can believe that. Probably is my prediction. I can believe that.

39:11Lomax Ward:It does feel in the crosshairs. Mads, which bits would you pick up on from this little smorgasbord? I think it's worth reflecting on just how far we've come in the last three years, where if you remember, some people had the view that it's kind of LLMs, chat GPT, AI, but what was it more than just an expensive autocomplete? And this is, as we can see, a whole lot more than just an expensive autocomplete. It's real intelligence. Now, it's not a principle. And the question is, what is left to do if AI becomes as good as it is becoming and can do all these things? What's left for us to do? And I think that's so interesting to see this collaboration here between what the AI models are doing and then the principal, the person operating those models.

39:59You scale that to everybody in the world. Anyone facing a PhD-level problem now has a path to solve it. It's truly astounding. I marvel at it every day. As the models grow more and more powerful, what is left for us? Well, sort of the four categories we look at, It's judging, it's creating, it's connecting, and it's exploring. Taking them in the reverse order, exploring, it's really pointing the telescope, thinking about where should we look, where should we explore, what are the frontiers we should try to cross. Connecting is all about human relations. One of the things we've seen is, although we've had record players and recorded music for a long time, but live music is bigger than ever.

40:42hundreds of billions of dollars. We go to see live musicians performing because we want to have human connection. And that is going to be a bigger and bigger thing, right? And AI will hopefully give us even more opportunities to say, hey, I can do the drudgery and we can spend more and more time with each other. Creating, giving things form, turning a want into a real thing. Well, when the production is free, the ability to create things is free, the imagination becomes the scarce resource. And taking people's desires, and they could be limitless, and turning them into real things in the real world is very much a human endeavor.

41:20And then there's the judging piece. It's the what should I focus on? It's what should I sign off? It's the surgeon that signs the consent form. It's the investor that commits capital. Now, the AI can be extremely clever at analyzing all kinds of things, but it has no skin in the game and and humans do so those are some of the four pillars that we think will be most important in the years ahead as the ai keeps getting better this might lead nicely into to the agi piece dan

41:48Lomax Ward:i don't know if you wanted to go there next yeah i wanted to know what your thoughts were on on on the the measurement and what deep mind is doing uh if you had any no i mean it's obviously that is kind of from a science from the science of ai deep mind is at the you know front of the front They are great engineering labs, open AI, anthropic, incredible engineers. When it comes to kind of the hardcore science of artificial intelligence, I don't think anybody is further along than DeepMind. And what's really interesting is they said nobody can agree on what AGI is. So let's try and define it. They've done that by mapping out 10 cognitive faculties that are mapped against human baselines.

42:26So sort of saying there are 10 ways we can measure how a human operates from a cognitive perspective. And then we can say how good are humans at those 10 and how does that compare to the AI. And when the AI exceeds what the human can do, well, then we have some kind of AGI. What I found interesting is that the tests are all disembodied. So they're text and images. It's not touch. It's not spatial awareness. It's not physical world test. And if you look at the learning faculties, they're very much about things that are kind of the models can do in the scheme of what we know as an LLM today. Whereas actually, I think so much of what would be required for you to have real AGI would be things in the real world.

43:14One of the things we've seen with these models is that some of the really hard stuff, kind of all the PhD science level stuff, is easy for the LLMs. But some of the stuff that's easy for us, like catching a ball, is still really, really hard for AI models. And so no LLM comes close to a three-year-old's ability to navigate in the real world. And so for me, until you embed those real-world aspects into your AGI definition, you aren't really talking about AGI.

43:46Lomax Ward:Wow. Listen, they're smart enough to get there, Lomax. No, one more thing I would add, which I could probably articulate better if I had more time, but one is around verification of the outputs of AI. And you can go back to this dog mRNA example, right? Is a clinical trial in the case of medicine is basically one big verification exercise, right? to check the safety and the efficacy of what you've done in silico and then in mice or monkeys or whatever you, whatever organ you tested it in. So that is not going to go away. Like this dude can create an mRNA vaccine and inject it in his dog without really like checking it.

44:28But you know, that's never going to change, right? So there's still, there are big opportunities in these kind, And you need to improve the verification exercises that you run on AI. But AI still often spits out and will continue to spit out, especially if we're kind of always improving AI. Things that have imperfections, right? So there's still going to be that. And also, I do think kind of tied to that, there is just AI is amazing for getting 90 % of a job done, right? This is what it's great for in white-collar work. but that final 10 % we should never forget that and there will be a lot of the value is going to come of that right and so the human judgment the human judgment the human polishing you create a piece of content for marketing and you just create it with AI and you put it out there which annoyingly for all of our ears and eyes most people seem to do but you spend yes at half an hour really polishing it honing it removing the LLLisms adding the context and like putting the message the call to actual whenever it's it's a really powerful example of you know man and machine working in in uh in income compatibly but um

45:49Lomax Ward:that is so important these days so anyway well we're gonna let's talk about let's talk about a product which is possibly doing the opposite or the counter to that which is google stitch very quickly yeah uh so it looks like there's some more side swiping at some of the models from the Hype scalers. Google just coined the term vibe design. Don't know if they'll catch on, but we'll see. And if you're a startup in the AI app builder space, you should probably pay a little bit of attention. Yesterday, Google Labs upgraded Stitch into a full AI native design canvas. So you describe what you want your app to feel like, and it generates the high fidelity UI with exportable code.

46:27Lomax Ward:It's free, infinite canvas, voice interaction, instant post typing, ships to an MCP server. So So your design just flows straight into the toolkit that you're working with. So who's in the blast radius here? Well, obviously Lovable. Now they've just hit 400 million ARR. They've just crossed a 6 billion valuation. Fastest scaling European startup we've ever tracked. Even Lovable's head of growth has said that her biggest fear isn't rivals, but it's the platforms with the big distribution. So how this will pan out is yet to be seen. We see many other products that don't encroach, like Google's DocuSign product hasn't really nibbled away much at the DocuSign efforts.

47:09Lomax Ward:But look at their share price over the last year or so. DocuSign isn't doing particularly well, so maybe that's just a matter of time. But Google Stitch, Mads, kick off this one for us. Yeah, I think you asked some of the right questions, Dan. I mean, we talked on the last episode about how nobody writes code anymore. and now Google is trying to collapse the design layer too. And now we get to a point where nobody designs anymore. I think the way Stitch integrates with tools around the ecosystem, you can integrate it directly with Cloud Code with MCP. You can integrate it with all of Google's in-house tools is quite neat, shows they're serious.

47:44They obviously have, as you say, massive distribution. The question is, is this just a distribution play for Gemini or is it their attempt at building a standalone? Right now, it's still too early to say for sure. Lovable is still more capable and I think a more polished product. But I certainly would probably keep me a little bit sleepless at night if I was kind of one of the specialist firms like Lovable. And that was my business model. And I saw Google introducing something as, you know, as good as Stitch is.

48:15Lomax Ward:Yeah. I mean, Lovable have introduced a new business model, I think, haven't they? They're doing something different. Well, so they are trying to go after the model we've seen with Claude Cowork and that we've seen with Perplexity. They launched their personal computer, which is sort of this whole idea of I'll use it as my command console to get work done, not just as a chat bot to ask questions and get answers from, but a place where I upload my documents, my data, and they can then turn data into new presentations and new formats. And it sort of becomes the central hub for my work. And Lovable is going straight after that segment as well.

48:50so they're getting a lot of immediate hate for that as you might you know why there's why what's what's the deal calling this to death now of um lovable um they're getting a lot of hate because i guess um people are saying look you you can't be all things to all people um you need to speak to what you're good at to what you're known for you need to double down on that you're now encroaching on so many other people's turf you're diluting your positioning um on the maybe this is

49:18Lomax Ward:the new world maybe this is the new ai world where you are not as focused and very anti what we talk about as vcs which is focus focus focus maybe you can maybe maybe this is uh maybe this is the new world that we're living in no look and there are precedents right so notion went from docs to workspace um shopify from i guess store builder to commerce platform you know each time you kind of expand you get a bit of pushback and hate and questions around whether you can actually cut it and so we time will tell but it's definitely a sign from lovable mads know that um that they feel that they can't just complete swim in this single feel the heat yeah you feel the heat 100 right gentlemen last but not least on the list is um a little bit of government-ish announcements we've got two of them we've got the non-competes and eu inc taken together at their word could be significant structural reforms for European startups.

50:16Lomax Ward:On Monday, Rachel Reeves used the Mays lecture to announce limits on non-compete clauses, the first time a UK chancellor has explicitly targeted non-competes as an innovation blocker. She framed that alongside the 500 mil in the sovereign AI fund, five bill for the British business bank, and a procurement overhaul designed to make government a launch pad for UK scale-ups, rather than, in her words, a check for global incumbents. Then on Tuesday, European Commission unveiled EU Inc. Again, we've seen a few of these releases of PR, which is the single corporate framework that lets you register a company in any of the 27 member states within 48 hours, fully online, under 100 bucks, with no minimum share capital and no notary required.

51:00Lomax Ward:And then all the notaries are going to do, they must be very upset. And Brussels wants this passed by the year end. but it's super watered down, plagued with pitfalls. So two reasonably big announcements for our world, but will execution meet ambition? Can it? It could be everything. It could be nothing. Lomax, what do you think? So let's maybe focus on Rachel Reid's speech at the so-called Maes lecture that she gave this week. Is that how you say it? Is it Maes, not Maes? Sorry. I thought it was Maes. Maybe it's Maes. Maybe I've been in Portugal too long. I don't know. Portugal is dealing with Maes.

51:33But anyway, obviously, it's not my... Look, I think that this is not just focused on tech, and this was a broader economic competitiveness, productivity speech. I think, you know, actually, interestingly, she did a pretty good job, which I think the Stama government hasn't done a brilliant job of so far, of like, really diagnosing the perhaps bleedingly obvious problems that we have in the UK, which is, as a reminder, you know, low productivity growth since 2008. planning rules have really rationed housing and infrastructure and then a huge accumulation of regulatory barriers that have steadily raised the cost of doing anything at all and or made things doing impossible.

52:12So I think she kind of set that out and set out antidotes to that. Interestingly, in the context of tech, as you alluded to, she just said a few interesting things. So one is she said that the UK will achieve the fastest AI adoption in the G7. so this actually kind of goes back to what we talked about before, is like if you can't develop the models you become the earliest adopter the biggest user, well, you know, God bless her, like fair enough, okay, well if the AI, if the UK can be the biggest and fastest adopter of AI in the G7, I mean I'm suspicious

52:46Lomax Ward:That's not a bad strategy I don't think that's a bad strategy You want to increase productivity, this is what you do then they announced a bunch of money and actually gets the backdrop of European governments announcing financial incentives for startups and generally getting laughed at by the Americans, because, you know, in the case of France, it was about, you know, 10 million euros. She's actually announced two and a half billion across AI and quantum, right? So 500 million for the sovereign AI fund, which you talked about, Dan, and then 2 billion to upgrade the UK's quantum capabilities, including a procurement program to actually spend up to 1 billion buying quantum computers from uk quantum computing manufacturers which is amazing because be the customer government be the customer which is what we've said here a lot don't just you know don't just give out grants like actually become the first customer record which is really really great because again especially in the backdrop of quantum where a number of leading quantum teams have actually left the uk and gone to the us there are still god bless them including in my portfolio So I'm talking my own book here.

53:50Some wonderful, wonderfully talented and dynamic quantum teams still here. And this actually, I think, is a really, really important thing for them that they stay here. And she's actually talked of, you know, 100 ,000 jobs coming out of these initiatives. Right. Interestingly, she then goes on to, I think, undermine herself because she talks about we talked about 500 million. We talked about 2 billion. Then we talk about 13.8 million into the UK's five national quantum research hubs and 12 million into the quantum researchers. I don't, someone needs to work with the government. Is that the coffee fund?

54:21Is that the lunch tab? And when you read it on the government's website, they all have the same level, they're all in the same area, at the same stage of the hierarchy. It doesn't make sense to me. Anyway, I think, look, generally pretty positive things for once from a Rachel Reeves speech.

54:41Lomax Ward:Right. Mads, anything on Reeves or anything on the EU Inc. piece that you'd pick up on? I think the non-compete piece is really interesting. California banned them decades ago. It's probably the single most cited structural reason for why Silicon Valley has the talent velocity it does. We know there is a massive tech sector in the UK and in London in particular, and we know what it can do if that talent can circulate freely between startups. I think this could be the start of a completely new chapter for the UK's tech scene. So I think it's very exciting. I don't think we've seen any specific timelines yet, but certainly something to watch out for.

55:18Always. I'd kick it down the road. But the slightly depressing thing, which always makes me laugh, is that these speeches are often peppered with phrases such as hubs, centres, programmes, you know the ai hub for this the quantum research program for this and you and i know that that's like okay that sounds wonderful in a kind of political speech but when have these things actually done any real good you know i mean it's a it's um it's an optics point more than anything but i feel like they tend to be well not more than optics like focusing on the wrong things but still god bless them they're hot in the right place funny so eu inc think i think is good you know it's just like this is much of the same this is the um single eu entity which will be quick to incorporate and cheap to run and will harmonize a bunch of the european legislations the delaware europe's delaware yeah um and and um you know that's now been kicked off in the legislative process i think they're trying to get it done by next year i think but you know there are going to be there's some pretty entrenched um lobbies they're fighting against that interestingly by By the way, guys, I just incorporated a bunch of entities in Delaware and it didn't take me 24 hours.

56:32It took me quite a few days. And at one point I was told I could accelerate it if I sent a fax to the registry.

56:38Lomax Ward:You know, I thought it was only the NHS that had fax machines. I know, I was like, exactly. So maybe it's not the hallowed turf that we thought it was. But I know we always think that everything in the States is paved with diamonds and gold. I mean, that's how it feels to European venture and European startups. The simplest and easiest place to, I mean, maybe some of the Baltic countries, but the UK is still an incredibly easy place to incorporate a company with a laptop and a credit card. Although I would say it's got harder. Three years ago, it used to be really simple. Now, having incorporated some things recently, I've noticed there's a few more questions, a few more KY.

57:18It doesn't happen automatically. It used to be automatic. Now they go away and run some kind of KYC or something. So, you know, that's it's not cosmic.

57:27Lomax Ward:It's not a cosmic shift. No, but going back to the beginning of what I said at Reeves' speech is, you know, this kind of steady, like incremental layering on of new regulations and rules every year. I felt incorporating a company in the UK three to four years ago literally was, you know, laptop, credit card, 20 minutes done. now it's like laptop credit cards half an hour 40 minutes a bit more form filling and then wait two days to actually get the confirmation i think there's a there was a fraud challenge wasn't there there was there was some other other of course like if you if you can incorporate a company with a credit card and laptop in 20 minutes you know you're opening yourself up to potentially um that kind of thing but you know still good anything more on the on the euing i'm slightly anxious that it's not going to roll out in a way that's going to work anything on the eu it'll get it'll before we move on.

58:13It'll get watered down. It'll get watered down. That's the problem. And the problem is the EU Inc. website, which is very clear and has great mission statements on it. And actually not more than just mission statements. It has really good detail as to what the...

58:25Lomax Ward:But just to be clear, though, Max, we've got EU Inc., which is the movement, the Andreas Klinger et al. movement. And then there's EU Inc., which has been adopted by the EU as the name of the... You go to the movement's website, which outlines what the corporate structure will look like, what the rules would look like, what the employee share option scheme, which is kind of a separate but, you know, connected point. Well, what it should look like. Correct. This is what they posit will work. And, you know, it's exactly. And if that starts to get watered down and pared back, and we talked previously about some of the reactions from the incumbents on this, you know, there's the Labour lobby that don't like the fact that the employees are going to have less say on some of the more kind of continental jurisdictions where employees have a big say in how companies are run um yeah they're going to fight some um it's tin hats on for them they're going to need to really push it through yeah okay um lovely lovely gentlemen um any deals of the week anything happening this week mad you got dinner the week yes so we saw travis come out of stealth not the band, we're not talking about the UK band here.

59:38Travis? We're talking about Uber founder Travis Kalanick who has been running Cloud Kitchens. So he's been eight years in stealth and he wasn't really in stealth because he was running Cloud Kitchens and had raised more than a billion dollars to make automate the making of food and people were scratching their heads a little bit and sort of saying what is all this about? Well he came out of stealth over the last week with a company called atoms. And the whole idea is to look at three verticals within physical AI in food and hospitality, in transport, and in mining. That's the new head scratchers.

1:00:18How are these three things connected? Travis is saying it's all about specialized robots, not humanoids. And it's all about having things and giving intelligence wheels to move and transport and do things in these different spaces. Now, the puzzle for us, as I said, is why he combined the three in one company. We're big believers in physical AI. We try and invest in it through a portfolio of companies, not squash everything into one business. He's taking the opposite approach. He wants to do all of it within his company. Why does it all matter? We could say, well, he is one of the greatest operators of his generation.

1:00:52The thing he did to build Uber was a superhuman feat and clearly a company that today dominates its industry and sector. And the question is whether he can repeat that in his new space. If he can, Atoms is going to become a very, very important company. But I think the jury is out whether this will really work. I've never bet against him. And there was this week, gentlemen, a certain announcement from a UK fund manager around a big commitment from the British Business Bank. Who was that? Was that? They sound awesome. I know so congrats guys on a 50 million pound anchor commitment from the British Business Bank is that right

1:01:33Lomax Ward:that's right yeah and just levering up all the private capital your fund 3 correct hopefully more announcements to come excellent huge chapeau to you really important focus on obviously physical AI is a big theme for you guys right yeah And just for the record We were in there before it was cool We were doing this before You were in there when it was called Industry 4.0 at some point a few years ago Wow yeah the names may change But the innocents are Protected But guys huge congratulations I can be the You know the doomsayer When it comes to Rachel Reeves And you guys We love the British government Well, yeah, lots to do and very exciting times.

1:02:26Lomax Ward:Thank you for that, Lomax. Gents, anything else happening this week? Anything we need to know about? Anything in the diaries? So what's coming up in the week ahead? Discord is starting their IPO roadshow, targeting a Nasdaq listing,$15 billion valuation. Do you know what that used to sound like a lot of money? $15 billion, it's almost like a... It's funny. Almost like, yeah, how the world has changed. Indeed. Indeed. We've got KubeCon in Europe, so it's the largest cloud native conference in Europe. I think at least a few of our fan teams will be there. And there are a couple of other conferences and things going on.

1:03:03But no major central bank meetings that I can see.

1:03:08Lomax Ward:Didn't the Fed announce a rate steady? Did they cut? They cut. Despite all the shenanigans going on globally and the expected okay well we'll see how that turns out uh lovely gentlemen thank you so much for your time and we'll catch you next week goodbye see you the next one