Anthropic's $1.5T IPO, NVIDIA's leadership shakeup, and global tech outages dominate European tech news
Summary
Speakers analyze Anthropic's record-breaking IPO, noting its $1.5 trillion valuation and strong first-day performance driven by safety-focused alignment. They discuss NVIDIA's revenue doubling to $116 billion alongside a 30% stock drop following CEO Jensen Huang's resignation and regulatory pressures. The episode highlights widespread tech outages attributed to legacy infrastructure debt rather than cyberattacks. European VC trends are reviewed, showing rising investment but persistent gaps in growth capital compared to the US. Geopolitical updates cover a summit in Shenzhen excluding Europe and UK economic policy shifts.
Key takeaways
- 1 My deal of the week is Genesis AI, which is a two two-year-old year old Paris robotics company that just raised five hundred $500 million at a $3.5 billion valuation with COSLA investing, um, Premji and Eclipse co-leading.
Unidentified speaker (speaker_2) · 32:44
Episode sections
- 0:00 – 2:00
Slush Event Review
Discussion on the efficient structure of the Slush conference in Helsinki for founder-investor meetings.
- 2:00 – 5:28
EU Inc. Corporate Framework
Debate on the watered-down EU Inc. proposal and arguments for retableing versus accepting the current version.
- 5:28 – 11:56
Anthropic IPO Analysis
Breakdown of Anthropic's $1.5 trillion valuation, $81 billion raise, and strategic investor backing.
- 11:56 – 15:58
NVIDIA Earnings and Leadership
Analysis of NVIDIA's doubled revenue, stock drop, and implications of CEO Jensen Huang's resignation.
- 15:58 – 20:14
Global Tech Outages
Attribution of widespread outages to legacy tech debt and patching issues rather than cyberattacks.
- 20:14 – 23:28
European VC Landscape
Review of European VC growth, Sequoia's London closure, and the gap in growth capital.
- 23:28 – 27:19
UK Political Updates
Discussion on Keir Starmer's EU reset talks and criticism of youth jobs plans amid AI automation.
- 27:19 – 29:51
Geopolitical Summit in Shenzhen
Analysis of the Trump-Xi-Putin meeting in Shenzhen and Europe's exclusion from key geopolitical settlements.
- 29:51 – 31:35
AI Regulation and Model Releases
Coverage of California's lawsuit against OpenAI and recent model releases by Microsoft, Google, and OpenAI.
- 31:35 – 32:44
Anthropic vs. OpenAI Competition
Speaker_1's claim that Anthropic is surpassing OpenAI in coding and writing capabilities.
- 32:44 – 33:57
Deal of the Week: Genesis AI
Overview of Genesis AI's $500 million raise and its approach to harvesting human movement data for robotics.
- 33:57 – 34:52
Week Ahead
Preview of UK EU summit outcomes and US inflation data release.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| Anthropic | private company | neutral | But I think one of the things we know is that Anthropic invested very early in safety and alignment, meaning that they have a better control on the frontier models than probably anybody else in the world. |
| MicrosoftMSFT | equity | neutral | So it's on all Microsoft 365, the FAA air traffic control system. systems, Java, NSA, banking apps, even AMP, Apple went down. |
| AmazonAMZN | equity | neutral | And this year, Europe's bond market lent Amazon, Alfbert, Alphbert, and Meta around $38 billion. in euros. |
| AlphabetGOOGL | equity | neutral | And this year, Europe's bond market lent Amazon, Alfbert, Alphbert, and Meta around $38 billion. in euros. |
| Meta PlatformsMETA | equity | neutral | And this year, Europe's bond market lent Amazon, Alfbert, Alphbert, and Meta around $38 billion. in euros. |
| OpenAI | private company | neutral | You spoke a little bit about the latest OpenAI models. We saw California suing OpenAI about agents that escaped. and hacked hugging face. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
Anthropic IPO raised $81 billion at a ~$1.5 trillion valuation with a 17% pop.
Unidentified speaker (speaker_3) ·5:28
How to verify: Check SEC Form S-1 filing for Anthropic's IPO price, shares sold, and initial market performance data from financial news archives. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Primary sourceFTX's $500 million Series B stake in Anthropic would be worth $37 billion today.
Unidentified speaker (speaker_0) ·6:22
How to verify: Calculate based on Anthropic's IPO valuation and FTX's reported ownership percentage in pre-IPO filings or news reports. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Primary sourceAnthropic's order book was 10 times oversubscribed.
Unidentified speaker (UNKNOWN) ·7:02
How to verify: Review press releases or financial news reports from the time of the IPO detailing order book metrics. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Anthropic priced shares at $83, above the $70-$78 range.
Unidentified speaker (UNKNOWN) ·7:02
How to verify: Check SEC filing for the final IPO price and initial range. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Primary sourceNVIDIA revenue doubled to $116 billion.
Unidentified speaker (speaker_2) ·12:47
How to verify: Check NVIDIA's Q3 or Q4 2026 earnings report filed on SEC EDGAR. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Primary sourceNVIDIA shares fell over 30% from their peak.
Unidentified speaker (speaker_2) ·12:47
How to verify: Review NVIDIA stock price history on financial data providers like Bloomberg or Yahoo Finance for the relevant period. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Primary source80% of outages are patches, updates, and upgrades.
Unidentified speaker (speaker_3) ·18:33
How to verify: Verify against CrowdStrike or other cybersecurity firm reports cited by the speaker regarding outage root causes. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
European pension funds manage $13 trillion but invest just under $800 million a year in European VC.
Unidentified speaker (speaker_0) ·21:23
How to verify: Check Atomico's European Tech Report or similar industry data for pension fund AUM and VC allocation statistics. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Primary sourceEurope's bond market lent Amazon, Alphabet, and Meta around $38 billion in euros.
Unidentified speaker (speaker_0) ·21:43
How to verify: Review bond issuance data from financial news or Bloomberg for the specified companies and currency. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
EU is paying 1.7 billion euros to Ukraine for drones and air defence.
Unidentified speaker (speaker_2) ·29:16
How to verify: Check EU Commission press releases or official statements regarding Ukraine aid disbursements. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
California is suing OpenAI about agents that escaped and hacked Hugging Face.
Unidentified speaker (speaker_3) ·29:51
How to verify: Search California Attorney General's office press releases or court dockets for lawsuits against OpenAI involving Hugging Face. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
OpenAI explored an acquisition of Goodfire, a startup that explores how AI thinks.
Unidentified speaker (speaker_0) ·29:55
How to verify: Check SEC Form D filings for Goodfire or news reports from reliable financial media regarding OpenAI's M&A activity in late 2026. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Primary sourceMicrosoft launched seven new AI models.
Unidentified speaker (speaker_0) ·30:15
How to verify: Review Microsoft's official AI blog or press releases from late 2026 for announcements of new model releases. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Primary sourceGoogle shipped Gemini for Neon.
Unidentified speaker (speaker_0) ·30:15
How to verify: Check Google's official blog or Android/Google Play store updates for the release of Gemini for Neon. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
OpenAI released GPT 6.2.
Unidentified speaker (speaker_0) ·30:15
How to verify: Check OpenAI's official blog or API documentation for the release of GPT 6.2. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Anthropic's IPO is described as a victory lap, and the speaker has used their model for over two years.
Unidentified speaker (speaker_1) ·31:35
How to verify: Verify Anthropic's IPO filing (S-1) on SEC EDGAR and check the podcast transcript for the specific timestamp. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Primary sourceAnthropic will beat OpenAI on coding and writing, pushing OpenAI to second or third place.
Unidentified speaker (speaker_1) ·32:10
How to verify: Compare benchmark results from independent testing labs (e.g., LMSYS Chatbot Arena) for coding and writing tasks between Claude and GPT models. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Genesis AI raised $500 million at a $3.5 billion valuation.
Unidentified speaker (speaker_2) ·32:44
How to verify: Check SEC Form D filings for Genesis AI or press releases from COSLA, Premji, or Eclipse regarding the investment. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Primary sourceGenesis AI employs more than 100 workers wearing sensor gloves and head cameras to harvest human skill as training data.
Unidentified speaker (speaker_2) ·33:04
How to verify: Review Genesis AI's website, whitepapers, or press releases detailing their data collection methodology and workforce size. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
US October inflation data will be released on Wednesday.
Unidentified speaker (speaker_1) ·33:57
How to verify: Check the US Bureau of Labor Statistics (BLS) economic calendar for the release date of October CPI data. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.
Do your own research
- Verify Anthropic's S-1 filing for actual revenue, burn rate, and customer concentration risks.
- Check NVIDIA's 8-K or earnings call transcript for details on Jensen Huang's resignation and succession plan.
- Review SEC filings for Google, Microsoft, and NVIDIA's investment stakes in Anthropic to assess conflict of interest.
- Analyze Gartner or IDC reports on global tech outage causes to validate the 'patch apocalypse' claim.
- Examine European Central Bank or national pension fund reports for actual VC allocation percentages.
- Read the California complaint against OpenAI to understand the specific allegations regarding AI agents.
- Compare Anthropic's coding benchmarks (e.g., SWE-bench) against OpenAI's latest models for performance verification.
Full transcriptexpand
0:00Unidentified speaker (UNKNOWN): Hello
0:00Unidentified speaker (speaker_0): and welcome to Upside, where every week we dig behind the headlines that are going to affect European venture. Today it is Mads, Pre and my good self, and what are we talking about?
0:10Unidentified speaker (UNKNOWN): Well,
0:10Unidentified speaker (speaker_0): we've got the Anthropic IPO. How is it for you? What happens if NVIDIA stops investing? What's really behind all of the outages? European venture steps up, plus what's happening in the week ahead, predictions, and deal of venture. the week. Hello, boys and girls. You both have been at slush. I am very envious, not nay jealous.
0:35Unidentified speaker (UNKNOWN): Uh
0:35Unidentified speaker (speaker_0): super event. Um, sometimes loved and a little derided. Maybe I'm pretty sure it won't go the way of uh Web Summit, but um still I think a handsome draw in the calendar. But how was it? What what was the give me the ups, the downs, the goods, the bads? Mads, how was it for you?
0:51Unidentified speaker (speaker_1): Can't believe anybody can deride this. I think it's it's so good. It's just, I mean, for me, it's all about the meetings, it's about getting together with people in meaningful one on one conversations, you know, founders, investors, LPs, partners, other VCs. In turns for terrific. What a fantastic event the Finns are putting on.
1:10Unidentified speaker (UNKNOWN): What
1:10Unidentified speaker (speaker_0): do you think makes it better? Because we've done we've done the circuit for a number of years, and it's always been my favorite event in the calendar. But what do you think makes it so good?
1:18Unidentified speaker (speaker_1): I think the way it caters to all the one to ones is terrific. It's it's easy, it's set up, like you've got all those tables laid out, hundreds of them, right? And all the meetings happening and kind of it's set up so that it's kind of a short, sharp twenty, twenty-five minutes, and then you wrap up and then it's next. And I think for both your the founders, investors, for everybody, it's it's really user friendly. Pretty
1:39Unidentified speaker (speaker_0): your first first first pop. What did you think?
1:42Unidentified speaker (UNKNOWN): Yeah,
1:42Unidentified speaker (speaker_2): first slosh, and I thought it was an incredible event. You have everything you need in in one one small city that um is easy to get around. And again, like Matt had sort of back to back meetings all day with with investors with some incredible founders, and I couldn't get enough. I could have done another couple of days.
2:00Unidentified speaker (speaker_0): Well, welcome back. Uh it is Friday the twenty 20th at time of recording. And yesterday, I just want to quickly dip into the EU Inc. Boys because they set out a call to action yesterday, if you want to call it that. So the EU Inc. is asking Ursula Mondelein to withdraw and retable the proposal. As a reminder, I'm sure everybody knows, but EU Inc. Because is a pan pan-European European company framework, a bit like a Del OS C Corp, standardizing and digitizing in incorporation corporation, ESOPS Aesops and the core rules to put Europe effectively into one block, if you like, which makes total sense. So the short version of the story is that it's been watered down, elements have been removed. It doesn't quite make sense. So the guys are effectively saying, can we please reconsider and retable this? And I don't know if you heard some of the founders at Slush, but there was a report yesterday of 40 of the biggest tech. entrepreneurs in Europe effectively mirroring the EU Inc. request and saying, Look, Europe, please get your shit together. So no real questions for either of you, but any any thoughts? I mean, Mads, you perhaps you can start this because you've always thought that getting something across the line is better than nothing.
3:09Unidentified speaker (UNKNOWN): Well,
3:10Unidentified speaker (speaker_1): I just think in the land of of reality, I mean I think you know, first credit where credit is due. the EU Inc. campaign has done more than anyone to get this on the agenda and they've succeeded. And I think it's just it deserves you know massive applause because
3:22Unidentified speaker (UNKNOWN): what
3:22Unidentified speaker (speaker_1): an incredible thing. So just in a few short years, having gone from nobody talking about this to everybody in the ecosystem talking about this and getting on the agenda of the commission. Now, I think you know, one of the things that's been said is you know, you've got to use this litmus test of uh is EU Inc. is the new proposal gonna going to be strong enough to replace or displace Delaware as the destination? And I just think it's totally the wrong test because the people are not choosing Delaware because you know Delaware is a better corporate structure. They're choosing Delaware because Delaware is in the US, and that's the land of capital and customers. And to to to make Europe the land of capital and customers, there's much more we have to do than just changing the incorporation model. So I I don't don't, with all respect, I don't think that's the right test. I think the key thing is, you know, is that a looted version worth having? Is it better than what's there today? And I would say infinitely so. I think the stock option cap at 750 staff and 150 million euro turnover, would we like it higher? Absolutely. It's not ideal, but it's better than That's what many countries have today.
4:23And if you think about it, a technology company with 150 million. in revenue would be a unicorn. And it can raise money, it can raise cash to pay people. And so arguably stock options are less vital at that stage than they are at the very early stages, right? When you just start out, you have no money and you need to try and attract people. And so I do think there is some benefit to that. I think when you think about taking a maximalist Maximalis position in any negotiation, you always got gotta to know and think about what the other sides walk away position. is. And the truth is that the other side, the notaries of this world, they would be very happy for this to fall apart. They have no incentive to negotiate, they have no incentive to compromise. And so if we think that by threatening them to say, look, if you don't give more concessions, if you don't come closer to our position, the deal is off, they'd probably say, fine, great, we love that. Right? They'd be happy for it to disappear. And that's why I think the negotiating position is actually quite weak. Better to get started, get something on the table, and then we'll chip away out of it over the years and hopefully get to the promised land in time.
5:28Unidentified speaker (speaker_3): I think I think we're gonna see splinter countries. That's where I think we'll end on this.
5:33Unidentified speaker (UNKNOWN): But
5:33Unidentified speaker (speaker_0): I wish them all the best. So I want to talk about Mads's happy place, the anthropic IP. IPO at Time of Pod was yesterday. But just a quick disclaimer: everything, because I am going to ask you some pointy questions. Everything we say from this point forwards is not investment advice. Please do your own research. Yada yada yada. However, so yesterday, which was Thursday, was Anthropic IPO Day. Very exciting. Mads, Mad, you must be fizzing. A little bit of a tee up. And then Pri, Pre, maybe you can give us some more context. But
6:02Unidentified speaker (UNKNOWN): they
6:02Unidentified speaker (speaker_0): went out. at 1.5 trillion, fully diluted, selling 80 plus billion in stock. Investors will be sharing around $860 billion from 142 invested. That's roughly a 6X blended multiple in just five years. This company's five years old or five and a half years old. Fun facts for you: there were no VCs in the series A or B. Dustin Moskowitz, Eric Schmidt, and Jan Tallinn of Skype. Fame will see a 320x return on capital. Here's another interesting little tidbit for you. FTX led their series B. And had they been able to keep their stake, their 500 million investment would be worth 37 billion today. All seven founders are now multi-billionaires, but just Just for context, this is a five-year-old business. Now, Pri, what have I missed? in the tee up? Anything, anything that you'd want to add in there? And then Mads, I wanna uh I wanna come to you, I wanna ask you, did you get any? Number one, and then and then get your take. But Pri, give us some more context, please.
7:02Unidentified speaker (UNKNOWN): So
7:02Unidentified speaker (speaker_2): I think in terms of in terms of context, when we're talking about their valuation and and they sort of came out of the gate and the first day was a pop, it's it's not about how it was priced. Um the order book was was ten 10 times oversubscribed. It was priced at $83 above their initial $70 to $78 range, and they sold more stocks than they planned. Um so they raised $81, and they raised $81 billion off the back of that. And that demand came from big institutions, not just retail. And all this is in a year when the 10 biggest IPOs are down about 20%. So investors aren't buying AI as a theme anymore. They're picking one company with 30%. $33 billion of revenue in in nine months, which is which is phenomenal, and choosing to look past, I guess, the the losses that we've seen and saying, Well, we're gonna overcome this. And then there's Google, Microsoft, NVIDIA, who put together another 26 billion alongside the IPO. And in effect, they're funding one of their biggest customers. The biggest losers here are SoftBank, a soft bank who missed it. Um, I guess Sam San Bankman Backman Freed, Freedom who had to give up his stake. That's my initial my initial take.
8:03Unidentified speaker (speaker_0): Yeah nice. But did you get any by the way? Did you buy any?
8:05Unidentified speaker (speaker_1): Uh listen I I I'm I'm happy with my position. I uh you know without commenting on any specifics.
8:13Unidentified speaker (speaker_0): There we go.
8:13Unidentified speaker (speaker_1): Um but it was an incredible incredible IPO. I mean it's just a five year old company biggest IPO ever right even beating the the SpaceX raising uh raising more than uh than Elon did and it had a great pop on the day. I mean it it kind of you know first opened around $111 and then closed around the 97th, kind of just shy of the hundred. So up 17% on the day, a nice little pop there on a day when the Nasdaq fell. I I think it's exciting in so many ways. I mean, you know, for all the the the doomerism and all the other stuff that's that we're talking about, you know, what a phenomenal value creation and what an incredible time to be alive to see this amazing technology, you know, power into the world this way.
8:52Unidentified speaker (speaker_0): And this is the this is the lightning rod, isn't it? This is gonna set set the trends, set the tone for for what happens in 27, is my is my assumption. Is that fair?
9:02Unidentified speaker (speaker_1): If I was gonna be um slightly negative for a second, I would say it surprised me a little bit that the stock could do so well in the first day and the market still trend down. Because I do think it's um it would have expected more of a spillover to the rest of the market, a little bit more superance. Maybe people were just dour and upset because they didn't get any. I don't know. Um but uh of course you can argue there are lots of questions around the whole AI economy, but this is and has the potential to be the most important company maybe in our lifetime.
9:33Unidentified speaker (speaker_3): Yeah. Yeah. Uh well we'll see how it performs over the over the coming weeks and months. Um any predictions on that one, Mads, from you?
9:40Unidentified speaker (speaker_1): Well, I mean, as as you know, I'm sort of on record to say that uh I thought this was an interesting company at two trillion and you know, people were initially saying, Oh, could it go out of one point two, one point three, one point four? Well, but it sort of traded straight up to two, then had tailed back a little bit to the one point eight. But this is an exciting company with a lot of potential. And I think y it's quite easy to underwrite it, I would say, to the two trillion on just, you know, looking at you know modest growth rates of 22, 23% a year over the coming coming years. Um the uh kind of maybe the the the question would be look with all the uncertainty, do you need to slap a bigger discount factor on it? Right? Because the it's not a steady as she goes kind of company. It's a company where we're saying, look, it could be the biggest company in the world, but could also be obliterated by open source and you know lots of other things. Um and so with that do do does that create risk factors that mean that you discount it in a different way than you otherwise would? But yeah, exciting business.
10:38Unidentified speaker (UNKNOWN): I
10:38Unidentified speaker (speaker_3): wonder what this will mean for OpenAI. Um I guess no new news.
10:42Unidentified speaker (UNKNOWN): Yeah, they're not a
10:42Unidentified speaker (speaker_1): good place. They are not in a good place. Uh their new model six 6.2, point two has been completely derided as as not fit for purpose. They had a good summer. They released some great models over the summer. But I think one of the things we know is that Anthropic invested very early in safety and alignment, meaning that they have a better control on the frontier models than probably anybody else in the world. And what does that mean in practical terms? Well, all the labs have got models, great models we haven't seen and haven't touched. Right? They developed develop them and then they spent spend a lot of time testing them and figuring out and making sure they're. aligned that they do what we want them to do. And so the kind of the the the saying on the grapevine is that anthropic, what they're excellent at, is to shorten the time span between creating a new frontier model and then testing it and making sure it's aligned so they can push it out. They can do that faster than anybody else because they've spent so much time on alignment and the safety of models. And that gives
11:35Unidentified speaker (UNKNOWN): them
11:35Unidentified speaker (speaker_1): a competitive edge. And it was there was also always sort of a counter counterintuitive to thing where we were talking about safety and alignment. And you would have thought, oh, is that something that's going to slow companies down if they spent too much time on safety and alignment? And that focus on safety alignment seems to have conferred a competitive advantage of Anthropic now, paradoxically meaning that they can move faster than anybody else.
11:57Unidentified speaker (speaker_3): Interesting. Interesting.
11:58Unidentified speaker (speaker_0): Well, let's stay with big tech. I want to talk about NVIDIA.
12:00Unidentified speaker (UNKNOWN): Uh,
12:01Unidentified speaker (speaker_0): Not so much Jensen's independent. indiscretions, but earnings this week out. Beating guide and shares fell.
12:07Unidentified speaker (UNKNOWN): At
12:07Unidentified speaker (speaker_0): the same time, the board stated quite clearly that Jensen isn't coming back. And on her first earnings call as interim CEO, Cress said, we are rigorously evaluating future investments, commitments, and guarantees. Now that I don't know what to take out of that, but we need, I would argue, NVIDIA as the central bank. in AI, or at least a major player at the center of discourse in AI. What I'm really intrigued to dig into here is what if NVIDIA stops investing. But Pri, first, what are your thoughts on this to start with, on the on the earnings and beating guide and shares falling? And then maybe I can ask you both about how we see them at as the center of the universe.
12:47Unidentified speaker (speaker_2): Yeah, so this is the strangest week in NVIDIA's history. On the number numbers, it's it's never been stronger. Revenues more than double to $116 billion. And the next year's growth vote forecast went up to 75%. And yet the shares fell down over over 30% from their peak. And that's because NVIDIA's losing three things at once. This is the way I see it. So first it's founder, Jensen resigned after a year of denying um, of course, the the affairs that we spoke about, the the scandal. scandal. Um Then I think there was a period which we've all read about now, which is like the cover up, which did more damage than the affair. And the board has now said, I guess, that he he's not coming back. And then second, there's China. Washington is now regulating who uses compute, not just where chips are shipped. And the Mega Speed case is the reason for this. And then VIDIA stopped the H200 sales to China and set up the National Security Committee. And then third, and this is the one that matters for us, it's its role as the AI ecosystem is effectively banker. Big question of whether the now then our new CEO will deploy capital the way that Jensen did. And all this lands in a year when the 10 biggest IPOs are down 20%. So public markets aren't an easy alternative source of capital.
13:58Unidentified speaker (UNKNOWN): So,
13:59Unidentified speaker (speaker_0): Mads, dig into further what what what's gonna happen if the central bank doesn't bank? What's happening here?
14:06Unidentified speaker (speaker_1): There's no doubt that there will be some marginal assets that will be hurt. There'll be some neo clouds, there'll be some smaller labs, there'll be some startups that would have gotten funding that probably won't get funding now, and that will hurt them. I wouldn't overdo their role. It's not like they're the central bank of a of an economy where they sort of they they carry everything in in in their laps. There are other sources of capital. but uh they're they were very important. I don't think this is a wall, but I do think it is definitely a a major set of speed bumps on the way.
14:36Unidentified speaker (speaker_0): We always talk about Jensen and Elon and and Zuckerberg as if they are single handedly building the rockets, creating the chips and doing all these magic things. But there are teams of very impressive people in these organizations. who are no dummies who can absolutely work out how to run these businesses, make these investments, and do these things. We spoke last week about the the the maybe the lack of maverick attitude that we perceive Jensen to be having, but I can't believe that they are they are not smart people behind the scenes who are going to make smart decisions. So maybe maybe I'm overcooking this a
15:12Unidentified speaker (UNKNOWN): bit.
15:13Unidentified speaker (speaker_1): Yeah, no, but I think it's it's probably less about smart and it's more about license to take risk. And there's something you can do if you're the founder of a company, if you've built it, you've built an industry. And even if you know NVIDIA is a public company, we've seen Jensen was kind of probably forced to step down. It's not his business. He's got a board, he's got people he answers to. But absent of the scandals, he had license to do whatever he wanted to do. And if you were a hired CEO, you're not gonna have the same license and there's gonna be other considerations. And so I do think founders operate differently. And you know, pretty much without exception, we've seen kind of over the last decade or two that the biggest companies have been built by founders, founders that were close to the businesses or even even still ran them. Because it just gives a bit different license to take risk.
15:59Unidentified speaker (UNKNOWN): Yeah,
15:59Unidentified speaker (speaker_3): no, I think that's fair.
16:00Unidentified speaker (speaker_0): So this week, well, this week. last week, week before, it felt like the world our world was collapsing. All of the tech services switching off, all the lights going dark all over the place. And it's I think it's very easy to look for some kind of sinisters goings-ons. So Friday the 13th, last week felt like a real Friday the 13th, with everything falling over. So it's on all Microsoft 365, the FAA air traffic control system. systems, Java, NSA, banking apps, even AMP, Apple went down. So it's felt like, you know, and obviously the news cycle hasn't helped this, but it's not Iran, China, Russia, or a bored 16-year-old with clawed code in their bedroom. And having a poked around in more of the detail behind some of these outages, messages, it looks like we are pretty much just dealing at the f at first with some pretty fragile systems. a hypersensitized news cycle, and we are looking at, you know, a whole chunk of tech debt being being poked into light by AI. So Mads, I mean I guess the first question for you is do you agree with my armchair appraisal?
17:05Unidentified speaker (speaker_1): Yeah, I look I think to start off, we should probably consider the conspiracy theory, right? The systemic attack theory, if only to dismiss it. And so you know, some people are saying, look, is this a systemic attack? Is all these companies being attacked by nefarious actors? And the question would be why would the big firms deny, because they've all come out and denied and said, look, this is not a systemic attack.
17:28Unidentified speaker (UNKNOWN): This is
17:28Unidentified speaker (speaker_1): kind of other issues. Why would they deny it if it was a systemic attack? And one might argue, look, it's better for them to look slightly incompetent than to look vulnerable. If you're Microsoft and you're slightly incompetent, you can say, look, we've strengthened our processes. Because you're being seen as a bastion of safety for hundreds of millions of users, the last thing you want to be seen as is vulnerable. So kind of this conspiracy theory is that there's a cover up and these people really have been attacked by nefarious actors, and they're just trying to pretend there's nothing to see here. I must say, at this stage, I don't really buy it. I think it's all speculation. But we also know that. when public services are hit, you know, it would it it was not just people in private corporations. It was, you know, public services we rely on, hospitals and other places are impacted like this. We have to know uh what's what's going on. You have uh the founder of CrowdStrike, Dmitry uh Perovich, he said it's a patch apocalypse. Um, and because all the firms were raising to fix flaws, AI models of are finding. And so all those patches are compounding into something that wasn't quite uh wasn't quite anticipated. So yes.
18:33Unidentified speaker (speaker_3): I read
18:33Unidentified speaker (UNKNOWN): I read
18:33Unidentified speaker (speaker_3): a stat last week, Mads, that basically uh it was 80% of the outages are effectively patches, updates, and upgrades.
18:40Unidentified speaker (speaker_1): Yeah. Yeah, and then things cascade, right? Because we've all we all know uh kind of this this this meme of the whole the entirety of the internet being carried on this little Jenga block. that's maintained by some open source developer in Nebraska or wherever, right? Who's been doing it for 20 years. And if that Jenga block falls, all bets are off. And sometimes that's what happened with this stuff.
19:02Unidentified speaker (speaker_2): With Microsoft, the outage caused real world damage, and we didn't quite get a reaction in the market, which kind of surprised me. So in the UK, you had schools telling parents to to phone instead of email. We, of course, last week we talked
19:14Unidentified speaker (UNKNOWN): about.
19:15Unidentified speaker (speaker_2): touched on NHS and ambulance trusts that had that had outages and couldn't access their emails for patient records, which reported serious impacts. And court cases were delayed. So these are these are three big examples, and there are many more. Yet Microsoft shares dipped just 0.2%. And that tells you that investors see Microsoft as infrastructure that you you can't, well Or to me, that you can't switch away from. And that's great for Microsoft, but potentially worrying for everyone else. And then of course, there are all of these other things that happened within the same space of time. So we talked about Apple services hit by an outage. Then you have um critical um findings in Java Flow, and then the cyber attack, which hit the European Commission, all happening in the same space of time. And I agree with you, Dan. There's there's no sort of sinister This the hand behind some of these outages, but Europe did get many real attacks this week. And I think this is now a race between people being able to use AI to cheat, let's say, attack at a low cost, and then these expensive repairs of software upgrades that should have happened a long time ago.
20:14Unidentified speaker (speaker_3): Some of these systems are decades and decades old. So no bad thing to have a have a thorough run through and make sure everything is ship shape. Let's
20:23Unidentified speaker (speaker_0): look at Europe, specifically European venture. It's been a pretty big news week, I would argue, this week. Money side obviously is not quite at American levels. I think the 81 bill that Anthropic raised was effectively the whole of, according to Atomico, the whole of European tech VC last year, I think. So two weeks ago, Sequoia shared its London office saying we simply don't need a local presence, which is fair enough. This week it co-led a $5 billion round into isomorphic labs with old Demis. Plural closed a billion, uh, made a point of raising more of it in Europe. I'm loving that. Uh Atomico, the uh European tech report is out. They have showed that European investment is up 62%, the second best year on record, but Europe's share of global venture has been halved to 9.4%, which I guess makes sense with all the mega rounds going. on in the States. We still really need to poke pension funds. There's still a massive story of a lack of involvement from the European pension funds who managed $13 trillion, yet only put just under $800 million a year into European VC funds. So we need to do something more there and with other institutions. And this year, Europe's bond market lent Amazon, Alfbert, Alphbert, and Meta around $38 billion. in euros. So Europe is effectively funding a
21:43reasonable, okay, not an American standard, reasonable, but a reasonable chunk of American big tech. Oh, and also I saw that Anne Bowden of Starling Fame is back. She's building Lockwood, a bank for AI agents. That would be quite interesting. And Plural raised a billion pre. So tell us a little bit about Plural. They've raised a billion euros, which is which is okay. Again, not uh American money, but it's it's a big fund.
22:06Unidentified speaker (UNKNOWN): Well
22:06Unidentified speaker (speaker_2): yeah, I have I have a better say on this and tie I'll tie it back to the US. So I think a third of the a billion dollar a billion euros that they've raised is set aside to keep backing um their own winners at later stages. So these companies already in their portfolio and follow on tickets. Is. And that tells you that European funds still can't count on on growth investors. We need more growth capital here in Europe, which we're lacking. So they're becoming their own Series B and Series C investor for their own portfolio. And if we look at who's backing them, it's their longest-standing investors. One of them is the University of Michigan. And plural admits that its investor base started out as very US centric. And then you've got American endowments, which are funding European tech, whilst as we've complained about previously, Europe's own pension funds are setting it out. And I thought this was actually a good contrast with the timing. It's quite a good contrast with what happened with Sequoia closing their London office and moving back to San Francisco. And in the same Fortnite plural raised this 1 billion euros to back
23:06companies other investors might find too scary. So Helsing housing and Defense defence and OLIX and chips and Proximo and Infusion. Fusion, Um, hard tech and physical. world companies are where I think European funds are seeing an edge. And I think the early track record of of Plurals investments backs that up. And its portfolio has raised over three billion of follow-on funding in just 12 months.
23:29Unidentified speaker (speaker_0): I do believe, uh and a little prediction for you both, that we are a weekly podcast. So we are constantly talking about very similar topics week in, week out. And I think we forget how long it takes to raise funds, deploy, and for that recycling then to come through the ecosystem. It takes time. So we are looking at this on a weekly basis, but actually these things are going to take months and years. However, I do believe that we are going to see more plurals and much more across-the-board across the board investment. And I'm I'm hopeful. I mean, it's we have to. There is no choice. and there is an opportunity. It's not just charity. And I think this is very exciting for Europe. So I'm gonna be I'm bullish on Europe. Go, go, go. Uh and talking of Europe, which we're not in, so talking of the UK, as we record, Burnham is on his way to Brussels for the second EU UK summit. And I'm I'm I've no doubt people have read all of the pro rejoining and all of the new news about uh young people's freedom of movement. There's a food and drink deal, also links to the carbon market, which is on the table. And we are all patently aware that there needs to be a reset with the European Union. So we'll see what happens. Andy Milburn, And obviously, we talked about this last week. There is a challenge with swathes of un
24:49unemployed youths. So he has suggested a 5.8 billion pound package, including a legal guarantee for a job training or a college place for every young person. And the inflation uh numbers aren't looking great. Obviously, in the States they're not looking great. And here they're not looking great. The prediction is that it'll be five percent next year, which is bad news for us, bad news for our ecosystem. High inflation means high interest rates. What should we pick up on? here? I mean, Pri, do you wanna do you wanna set the tone for us?
25:18Unidentified speaker (speaker_2): First, um, this inflation is an energy shock, not an overheating economy. The Bank of England said um it's supposed to be fuel and energy linked to the war in the Middle East and it hasn't spread into food or services yet. And that suggests um that rates go up and then hold, not necessarily a continued increase. And let's see what happens there. I think that the the EU reset matters more to our ecosystem than the and than the inflation. And then my third and final point is Millburn. The government will pay employers £3,000 to hire a young person who's been out of work and is cutting national insurance for um 21 to 24 year olds starting in retail and hospitality as their target. And those are exactly the sectors where automation is arriving fastest when we when we look at automation and AI. So we're subsidizing entry-level jobs at the same time AI is taking the rung on the first rung off the first ladder. And that's a policy contradiction that nobody's talked about.
26:10Unidentified speaker (speaker_1): I don't think AI is the key issue here. The issue in terms of where we've been losing employment opportunities is in hospitality, it's in services jobs, isn't that huge part of the economy where, you know, as you say, many young people should get their first opportunities and we're where we made it so much more expensive to hire people. So of course we can be doomerous and talk about AI all we want, but the point is you with government policy, we have been tanking the labor market for young people. I would love to see the politicians be a little bit more creative in terms of how they think about this and not just think about how can we throw more money, more money at some stuff staff in areas where we already spend quite a lot. It'd be much better if we could lower the deficits so we had less need to issue bonds that would lower interest rates and stimulate the economy.
26:58Unidentified speaker (speaker_3): I'm hopeful. We'll see what Burnham comes back with and how it all goes with the EU. I quite like the fact there's a hundred thousand young people now with a with a freedom of movement ticket. Um that looks
27:07Unidentified speaker (UNKNOWN): like we've been
27:08Unidentified speaker (speaker_3): approved, right?
27:09Unidentified speaker (speaker_1): I think that's great. I think this should lift the cap. I think it should lift the the bracket to thirty five. But again coming back to maximalist demands and where do you stand? I think this is a great start. It's a movement in the right direction.
27:19Unidentified speaker (speaker_0): Yeah, okay.
27:20Unidentified speaker (UNKNOWN): Well,
27:20Unidentified speaker (speaker_3): let's stay with kind of the the the political piece, but but maybe quickly, uh, because I I'm not sure where we can go with this, but I we can't not talk about Trump, G, and Putin meeting at Apex and uh talking in Shenzhen. So what what is there anything that will come out of this? Do we think
27:39Unidentified speaker (speaker_0): that's relevant? I can't see any policy changes, anything on denuclearization, anything on trade, anything on rare earths? I've not seen or heard anything from the conference or any of the talks. Obviously, we don't know what goes on behind closed doors. Only even the Russia-Ukraine argument, looks like Trump was saying one thing, we're talking about it, and Putin was saying, well, we're not really talking about anything. So is there anything that's going to affect? Europe? Obviously, everything here is, but anything that you have both seen that would affect us,
28:08Unidentified speaker (UNKNOWN): our ecosystem,
28:09Unidentified speaker (speaker_0): Europe, the
28:10Unidentified speaker (speaker_1): economies? First you could ask why are the three parties there, right? Putin, he wants to sell oil and China wants cheap oil and room to export and wants uh some kind of settlement with the US. And Trump wants cheaper oil so he can get some time to resolve the Gulf crisis without being forced to. into a you know resolution he doesn't he doesn't like. So they are all pursuing self interest. I think the key thing for us is Europe is not in the room. And so that's what happens when you are weak, you are not listened to, which is why we need to not be weak. We need to be stronger in terms of how we we do things, we need to invest in the right places. Does it have direct costs for us? Well, yes. The US is not going to fund the war in Ukraine. Putin is pretty maximalist. He wants to take over Ukraine and who knows where he will stop. So we have to uh finance this. We have to fund them. We have to support them. It's going to cost the tens of billions, if not more, every year. So there's a direct fiscal cost. So all of this stuff again sort of connects back. And I think it's positive to see that uh the European countries are coming together and are, you know, many are wanting to and trying to do the right thing. The question is whether we're going far enough, fast enough.
29:16Unidentified speaker (speaker_2): The EU is paying 1.7 billion euros to Ukraine for drones and air defence. And I think the headline number is small, but the the structure of it matters. This is the eighth payment from um a ninety billion EU loan taking defense spending under it to over 18 billion euros. And about 40% forty percent of this payment goes to EU based manufacturers. So Ukraine aid is now also an industrial policy that builds Europe's own defense base. And the Commission is explicit that the goal is to fold Ukraine's defense industry into Europe over time.
29:51Unidentified speaker (speaker_3): Mad's your happy place. What has happened in the wonderful world of
29:55Unidentified speaker (speaker_0): AI this week? You spoke a little bit about the latest OpenAI models. We saw California suing OpenAI about agents that escaped. and hacked hugging face. We'll see what happens there. OpenAI looked at an acquisition of a startup that basically explores how AI thinks. They did make me think, Well, shouldn't you have this in-house? But okay, that's fine. Microsoft launched seven new models, I think. Google shipped Gemini for neon, OpenAI with GPT 6.2, six point two, which we'll talk about. So, Mads, where do you want to start on on the wonderful world of
30:25Unidentified speaker (UNKNOWN): AI?
30:26Unidentified speaker (speaker_1): You were talking about. Goodfire. Uh you know, I think it's an interesting acquisition. I think OpenAI I mean it is a formidable company and they could probably build anything they put their mind to, but they're in a fierce competition with Anthropic and they obviously need to try and leap from their competitor. I think it could be a strong acquisition. You know, Goodfire is reading what's inside the model. But there's no deal yet, right? It's been reportedly it's been explored. So I wouldn't I wouldn't worry about it. I think, you know, five years ago what we're looking at now is science fiction. We're at the frontier of a massive technological revolution. And so I marvel at what's happening every day. And we're in uncharted territory. And so there's gonna be new discoveries. That's that's part of it. The thing with California and California suing OpenAI, I mean, it's so defeatist. You've got a world-leading world leading company in your backyard. You you you should work with them. You don't want to sit down with How can we make you the best you can be? Uh how can we collaborate so that there's safety and so things go, not go to court. I mean, it seems like such a silly thing to do. If I was running California, I would collaborate rather than than than take a litigious route. But I guess it's politicians, it's populism, as they want to be seen to bash AI. And so they're sort of doing it for
31:35that reason. But what a what a short-sighted short sighted way to approach. things. I think it's kind of just coming back to the anthropic IPO. I mean it feels like a bit of a victory lap. Um it's been
31:49Unidentified speaker (speaker_3): Dario looked very happy.
31:50Unidentified speaker (speaker_1): He was happy. Uh look, it's been our house model for for more than two years. And we wrote about it early and sort of said well Look, we think there's something special going on here. Uh you know, even back at, you know end of 2024, start of 2025, when OpenAI had all the market share, we sort of said, look, we think uh we think uh uh Claude and Anthropic will will beat them on coding and and writing, and that OpenAI would be pushed into second or third place. And so it's just been really interesting to see that play out in many ways Is the way we we expected it to.
32:22Unidentified speaker (speaker_3): Well, I I just love this competition. Uh and putting the sovereign piece aside I I just love that there's this frontier competition. It's just a it's an amazing thing to watch, and we will see this all unfold and we'll see next year when OpenAI gets their act together and gets out into the public markets how that performs as well. Awesome and exciting. Uh Pri, do you have a deal of the week?
32:44Unidentified speaker (speaker_2): My deal of the week is Genesis AI, which is a two two-year-old year old Paris robotics company that just raised five hundred $500 million at a $3.5 billion valuation with COSLA investing, um, Premji and Eclipse co-leading. What's interesting, I think, in what they're doing is the strategy. Everyone agrees that the bottleneck in robotics is real-world training data. And Genesis is tackling it head on with more than 100 a hundred workers around the world wearing it sensitive. sensor gloves and the head cameras, and every movement maps directly onto the robot's hands. So they're literally harvesting human skill as training data.
33:25Unidentified speaker (speaker_3): I always feel for the operators who are geared up with the cameras and the gloves and the sensors, effectively they must
33:32Unidentified speaker (speaker_0): be looking at each
33:33Unidentified speaker (UNKNOWN): other, going,
33:33Unidentified speaker (speaker_0): hang on
33:34Unidentified speaker (UNKNOWN): a
33:34Unidentified speaker (speaker_0): minute, are we getting rid of our own jobs here? So
33:37Unidentified speaker (speaker_2): Yes. Um we got to try a few of those on um at Slush. I don't know if you did Mads, but there were there were a lot of those being demoed. Um so I was wearing gloves for a fair bit of the weekend um that were that were guiding robots.
33:47Unidentified speaker (speaker_0): So they are um so they're getting rid of VCs
33:49Unidentified speaker (UNKNOWN): as well.
33:50Unidentified speaker (speaker_0): What an awesome job. Loving their
33:51Unidentified speaker (UNKNOWN): work. Um
33:52Unidentified speaker (speaker_0): Mads, what's happening in the week ahead? What have we got to pay attention to? What's coming up in the diary?
33:57Unidentified speaker (speaker_1): Well, uh on Monday, Burnamy's Burnham is gonna going to set out the next steps after the EU summit. We'll be interesting to see where it goes. Obviously, a lot of debates in the UK about this this relationship and should we, shouldn't we, et cetera? On Wednesday we get US October inflation. Um and then on Thursday it is US Thanksgiving, uh, which means uh the world will be a more peaceful place for a few days as all the tug into Turkey and their markets close, et etc. cetera, etc. et cetera.
34:25Unidentified speaker (UNKNOWN): And say so thank you for for getting rid of
34:27Unidentified speaker (speaker_0): the Brits, right? Is that is that
34:29Unidentified speaker (UNKNOWN): being
34:29Unidentified speaker (speaker_1): all the flights? Yes.
34:30Unidentified speaker (speaker_0): Um well that's it, my lovely people. Thank you so much for joining us and watching, listening this far. Pre Mads, I love you dearly, and we will catch you all next week.
34:49Unidentified speaker (UNKNOWN): Sai, tem oi, tem.