Startup Europe — The Sifted Podcast

Deedy Das on Anthropic IPO, AI Economics, and European Startup Challenges

24 min Listen to the original
Anthropic IPOAI EconomicsVenture CapitalEuropean StartupsLovableMenlo VenturesBubble DebateDeepMind

Summary

Deedy Das, partner at Menlo Ventures, discusses Anthropic's IPO, characterizing the company as generational and arguing its $1.4 trillion valuation is likely too low. He rejects AI bubble narratives, citing profitability across the supply chain and demand exceeding supply. Das also addresses Menlo's investment in Lovable, European ecosystem hurdles, and the challenges of US VCs operating in Europe.

Episode sections

  1. 0:02 – 1:56

    Intro and Anthropic IPO Impact

    Host Anne Schrader introduces guest Deedy Das, partner at Menlo Ventures and early backer of Anthropic. Das states that while he cannot comment on specific IPO details, he views Anthropic as a generational company and believes its public listing will not significantly impact broader market sentiment due to the company's long-term growth trajectory.

  2. 1:56 – 3:23

    Anthropic Valuation and AI Economics

    Das argues that Anthropic's $1.4 trillion valuation is likely too low, predicting it will be worth significantly more in the long run. He cites Jeff Dean's comment that spending $10 million per year on AI per engineer is acceptable, illustrating how AI is fundamentally changing software engineering economics.

  3. 3:23 – 7:49

    AI Ecosystem Impact and Bubble Concerns

    Das contends that the Anthropic IPO benefits the global AI ecosystem rather than just the US, as AI accelerates innovation and allows startups to grow faster than incumbents. He rejects the notion of an AI bubble bursting, arguing that demand currently exceeds supply and that every part of the AI supply chain is profitable.

  4. 7:49 – 11:07

    Menlo's Investment in Lovable

    Das explains that Menlo Ventures was late to invest in Lovable, a vibe coding startup founded by Anton, who previously ran Sauna. He admits they missed the initial opportunity because they asked the wrong questions about defensibility, but later recognized Lovable's high retention and unique user base of non-technical founders.

  5. 11:07 – 12:46

    Lovable vs. Anthropic Competition

    Das asserts that Lovable does not compete with Anthropic, noting that different AI tools serve different user bases and purposes. He compares the market to multiple coding tools like Cursor and GitHub Copilot coexisting, suggesting that the AI pie is large enough to support many successful products without cannibalization.

  6. 12:46 – 17:02

    European Startup Ecosystem Challenges

    Das praises European talent but identifies key hurdles: unfavorable legal and tax environments, a lack of geographic concentration compared to US hubs, and a gap in 'venture culture.' He notes that European founders often lack the specific knowledge of how to pitch to VCs and structure companies for capital legibility.

  7. 17:02 – 18:48

    US VCs in Europe and Sequoia's Exit

    Discussing Sequoia's closure of its European office, Das argues that running multi-geo venture offices is difficult due to the loss of team camaraderie and trust. He believes it is challenging for US firms to effectively manage European operations remotely and suggests investing across borders without physical offices is a viable alternative.

  8. 18:48 – 23:46

    Neolabs and DeepMind Spinouts

    Das addresses skepticism around Neolabs and DeepMind spinouts, stating that while 99% of startups fail, these entities face unique challenges in competing with frontier labs on compute and data. He believes breakthroughs will still occur, creating exit opportunities, but notes that many Neolabs may struggle to fundraise at current valuations.

Assets discussed

AssetKindStanceContext
Anthropicprivate companybullishAnthropic is set to begin trading this week, and I think it'll kind of be what is the market's sort of first test of this frontier AI type company going public.

“Stance” is how the speakers talked about the asset, not a recommendation.

Claims worth verifying

Anthropic is set to begin trading this week, marking the market's first test of a frontier AI company going public.

Unidentified speaker (speaker_0) ·0:43

How to verify: Check financial news archives for November 2026 for Anthropic's IPO date and trading start date. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Anthropic has a valuation of $1.4 trillion before pricing at $83 per share.

Unidentified speaker (speaker_0) ·0:02

How to verify: Review Anthropic's S-1 filing with the SEC EDGAR database for the proposed price range and pre-money valuation. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Primary source

Menlo Ventures backed drone maker UForce, which raised $700 million.

Unidentified speaker (speaker_0) ·0:22

How to verify: Check UForce's press releases or funding announcements for a $700 million round and Menlo Ventures' portfolio page. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Anthropic has proven in its S-1 filing that it can scale models, make them profitable, and that its revenue line continues to go up.

Unidentified speaker (UNKNOWN) ·1:34

How to verify: Read Anthropic's S-1 filing on SEC EDGAR, specifically the 'Risk Factors' and 'Management's Discussion and Analysis' sections regarding profitability and revenue growth. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Primary source

Deedy Das believes Anthropic's $1.4 trillion valuation is too low.

Unidentified speaker (speaker_1) ·1:58

How to verify: This is a subjective opinion stated by the speaker; verify by checking the transcript timestamp 1:58. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Jeff Dean, founder of Discovery Loop, stated he would spend $10 million a year on AI per engineer.

Unidentified speaker (speaker_1) ·2:24

How to verify: Verify via public interviews or articles featuring Jeff Dean and Discovery Loop from late 2026 where he discusses AI costs. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Anthropic IPO will not suck money and talent back to the US; instead, AI benefits the global ecosystem.

Unidentified speaker (speaker_1) ·3:46

How to verify: This is a forward-looking opinion. Verify the speaker's stance via transcript at 3:46. Historical data on talent migration can be checked via LinkedIn or industry reports. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Primary source

Startups are growing 10x plus year-to-date on sizable revenues, accelerated by AI.

Unidentified speaker (speaker_1) ·4:27

How to verify: This is a general claim. Verify by checking VC market reports (e.g., from Menlo Ventures or other firms) for 2026 growth statistics of portfolio companies. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Menlo Ventures invested in Lovable at the Series B round.

Unidentified speaker (UNKNOWN) ·10:10

How to verify: Check Lovable's funding history on Crunchbase or PitchBook for Series B investors and round size. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Sequoia pulled the plug on its European office.

Unidentified speaker (UNKNOWN) ·17:02

How to verify: Check Sequoia Capital's official announcements or press releases regarding the closure of its European office in 2026. For recent events, verify 2026-08-19 through 2026-11-19. Use another period only when the speaker explicitly names it.

Do your own research

  • Review Anthropic's S-1 filing for detailed financials, risk factors, and management discussion on profitability metrics.
  • Verify Jeff Dean's 'Discovery Loop' affiliation and the specific context of the $10M/engineer comment via primary sources or transcripts.
  • Analyze Lovable's user retention data and revenue models to assess the 'vibe coding' market viability against competitors like Cursor.
  • Evaluate European legal and tax frameworks compared to US hubs to understand the 'venture culture' gap cited by Das.
  • Check Sequoia Capital's official statements regarding the closure of its European office and strategic shifts.
  • Research DeepMind spinout performance and Neolabs portfolio outcomes to assess the '99% failure' rate claim in this context.
  • Compare AI supply chain profitability reports from major chipmakers and cloud providers to verify Das's claim of universal profitability.
  • Investigate Menlo Ventures' portfolio performance and Lovable's current valuation to assess the impact of their late entry.
Full transcriptexpand

0:02Unidentified speaker (speaker_0): Hello and welcome to a special slush edition of the Sifted podcast. I'm Ann Schrader, senior reporter at Sifted, and your host today reporting live from Helsinki. I'm joined on the sidelines of the big startup conference by Didi Das, a partner at Stalwart Sloan Valley firm Menlo Ventures, which boasts, among other things, being an early and big backer of Anthropic. While its most notable investments are stateside, Menlo Ventures has also backed a handful of startups in Europe, from vibe coding company Lovable to drone maker UFOs, which just announced a $700 million round. At his firm, Dee Dee focuses on AI and machine learning as well as robotics, hardware, and infrastructure. Dee Dee, welcome to the Sipted Podcast.

0:42Unidentified speaker (speaker_1): Thank you so much for having me, Ed.

0:43Unidentified speaker (speaker_0): I do have to ask the timely question. Anthropic is set to begin trading this week, and I think it'll kind of be what is the market's sort of first test of this frontier AI type company going public. How do you think it will impact market sentiment broadly for AI?

1:01Unidentified speaker (speaker_1): I I think, you know, well obviously I can't comment on things specific to the IPO, but what I what I will say is it is so undeniable to me that Anthropic is a generational company. For for some several reasons. I would say that the short-term impact on the market honestly doesn't faze me that much. I think what we're seeing with these models is models continue to scale and get better with compute and with data. And now the the next thing is compute efficiency. Um

1:34Unidentified speaker (UNKNOWN): And

1:35Unidentified speaker (speaker_1): Anthropic has been able to scale and monetize, has been able to prove as we see in their S1, and that they can scale models, make them profitable, and that line just continues to go up. So while there is market demand, I think the company is a long term extremely I'm very bullish on that company.

1:56Unidentified speaker (speaker_0): Is it worth 1.4 trillion?

1:58Unidentified speaker (speaker_1): I think that's too low, probably. Too

2:00Unidentified speaker (speaker_0): low.

2:01Unidentified speaker (UNKNOWN): In

2:01Unidentified speaker (speaker_1): the long run, it's worth going to gonna be worth a lot more than. that. Um

2:05Unidentified speaker (UNKNOWN): I

2:05Unidentified speaker (speaker_1): think the ways I think what people often fail to to realize about where we are in the cycle of AI is just the kind of things these models can do on the job. Um and maybe we see a lot more of this being in in the venture world, but you know, I I had a conversation with with a founder of a

2:26Unidentified speaker (UNKNOWN): of

2:27Unidentified speaker (speaker_1): K I this is public. I had a conversation with Jeff Dean um a few weeks ago on how he uses AI models at Discovery Loop, his new company. And he and I asked him, hey, like how much would you say is too much to spend on AI per engineer? And he said, I don't know, 10 million a year. And he didn't even blink at all. He's like, this is just the way we work. Um and I've heard all kinds of answers from 30% of RD to to more, um, for for all startups at various sizes. And I think this is at least the new way we do software engineering. And I think that's going to continue to translate to more and more disciplines. Um and just taking that lens on it, you know, I don't want to comment on whether it's anthropic or anybody else, but that is the impact AI is gonna have. And there's no signs of the quality of these models slowing down. Um so that's where that's where I lie on the matter. Aaron

3:23Unidentified speaker (UNKNOWN): Powell

3:24Unidentified speaker (speaker_0): I think the big question we here in Europe have regarding the anthropic IPO is whether it will sort of lift the whole AI ecosystem and you know, including in Europe as well, or if it kind of just sucks more money and talent back to the US. Do you have a take on that? Do you see this kind of rising, rising tides lift all boats situation, or do you think this is really uh an anthropic and US story?

3:46Unidentified speaker (speaker_1): I don't think of it as an anthropic in U.S. story. I think the reality is everyone's benefiting from these models. It allows more innovation to happen. You have startup formation at a pace that was unpr is unprecedented, and we see it all the time. What used to take a hundred 100 employees to do often even takes two to three employees to do today. So

4:07Unidentified speaker (UNKNOWN): I

4:07Unidentified speaker (speaker_1): think it's a great time for innovation for all for all of those companies. I also think that every single startup benefits more from AI than the incumbent, because the incumbent moves too slow to even adopt and use these things quickly enough. And that's why when we look at as VCs, when we look at companies in this market, they're growing at rates that are faster than ever before. You know, the number of companies this year that have been growing 10X plus year to date on really sizable starting revenues has been amazing across the world, not just American companies. And a large part of that is because of how accelerated they are by these models. And it's that's not even just UnderI, Under I it's all the models. It is, it is AI as a whole. So I actually like don't think of the IPO as a it's not a big moment to me. It just happens to be a company going public. The benefits were seen before, they're gonna be seen after. This is really a service to companies and to innovation.

5:10Unidentified speaker (UNKNOWN): But

5:11Unidentified speaker (speaker_0): I guess in terms of where we are in the market cycle, I think a lot of people are concerned about an AI bubble. Some stories I've even seen recently have claimed that it's starting to burst now. Do you think that this is a big moment to determine whether that's true or not? Or do you think that, you know, there's kind of factors outside of that that are at play here? What what's your take on whether we're at sort of a bursting moment in AI?

5:33Unidentified speaker (speaker_1): In every um bubble

5:36Unidentified speaker (UNKNOWN): that

5:36Unidentified speaker (speaker_1): we've seen in the past,

5:38Unidentified speaker (UNKNOWN): it's in

5:38Unidentified speaker (speaker_1): technology, namely the ones I can think of are like the telecom bubble and the dot com bubble. The reason the bubble bursts is when you overbuild supply for not enough demand. And that used to be the case, I mean in in both of those things, it's really hard to predict demand. You don't get it to know immediately. So you you see some, you build and you build, you build we way too aggressively up front, and then when the demand doesn't show up, then you're in a position where oh people. were too optimistic about something that didn't come to fruition. What does that mean today? Uh or how does that translate to the world of AI today? Well, the demand is clearly there.

6:15Unidentified speaker (UNKNOWN): And

6:15Unidentified speaker (speaker_1): it's there's more demand than any of the supply can handle, and we know that. Um now, what you could say is, okay, well, how long will that demand last and will that demand continue to be there? Um and maybe maybe we're priced for m more demand than we uh uh are able to supply today. And some of those things might be true, but you know, I would say we're not overbuilt. We're clearly still very much supply constrained. And two, I think the demand is like the entire economy. I just don't see the a world where I mean I think it's on like it's obviously true. Like every single knowledge work job. is bound to be affected by AI. And so it it is just a question of how long it takes to diffuse these models into those jobs. So that's that's my take. I actually don't think there's like the some companies will rise and fall, but when it comes to a bubble, as in everything comes crashing down, I just don't see why that would be the case. I mean, these are every part of the supply. chain

7:15is extremely strong. Everyone's making money, right? Like the people making chips are making money. The people building on chips are making money. The people building on models are making money. Uh and the people using the end applications are making money. Like like so which part of it falls down, I'm I'm not sure. Now, okay, you miss next next quarter as a public company, sure, there might be some jitters, right? You know and a CEO gets fired for reasons out of your control, yeah, sure, there might be some market jitters. But that's not a bubble bursting. You know, it's it's a like bubbles bursting are a very different phenomena, in my opinion.

7:49Unidentified speaker (speaker_0): Okay. In terms of Menlo Ventures uh investments in Europe, you guys have also backed uh vibe coding startup lovable. Um I'm curious, can you kind of tell me the story of how that deal came together?

8:01Unidentified speaker (speaker_1): So uh yes, I can. Lovable actually came out of Anton was working at a portfolio company of ours called Sauna. It's based in uh Sweden as well. Um it's run by a a guy called Joel. Um so Joel ran Sauna. Anton first came out of that, and to be honest, we probably should have invested immediately when Anton left Sauna and we were way too late to to invest in that company.

8:27Unidentified speaker (speaker_0): Why didn't you invest earlier?

8:29Unidentified speaker (speaker_1): It was just a miss for us for various reasons. We didn't see how how big the opportunity was when he first came on the scene. And then when it started taking off, um, I think we just got we just asked too many questions that we think were, I think in retrospect were probably the wrong questions to ask. Um

8:52Unidentified speaker (UNKNOWN): any

8:53Unidentified speaker (speaker_1): the there are questions about like the questions people still ask about the company, right? Like, is it defensible and all of this other stuff? Why can't you just do this with something else? But eventually we came to realize that lovable one of the spectacular things about lovable was

9:08Unidentified speaker (UNKNOWN): it

9:08Unidentified speaker (speaker_1): they'd grown on, they'd gone on to be an extremely highly retentive product for this user base that was actually quite different from the user base that would use, say, like a Claude or ChatGPT. And people were using this to run actual companies and businesses and websites, whether it's like in a lot of non-technically, so p PMs at different companies were using it to mock up designs and products they wanted to build, or it was also small business owners that were building complete sites on Lovable and deploying it that way. And I think the reality is like a lot of the reasons people continue to use Lovable, despite there being, you know, maybe a lot of ways to achieve the same thing, is just because of how smooth that experience was. And a lot of it is because of how formidable Anton of the team have been in and in believing and driving that. He's an absolutely incredible uh growth person. So the business continues. to grow exceptionally well. And so we ended up we'd ended up doing that round at the series. We did the series B,

10:10Unidentified speaker (UNKNOWN): if

10:10Unidentified speaker (speaker_1): I recall. That's right. And then the next one as

10:12Unidentified speaker (UNKNOWN): well.

10:13Unidentified speaker (speaker_2): This podcast is brought to you by Charge B.

10:17Unidentified speaker (UNKNOWN): As

10:17Unidentified speaker (speaker_2): an AI founder, you're doing the hard work of reimagining entire industries, repeatedly identifying use cases that abundant intelligence and domain. specific know-how can solve and shaping the future in the process. ChargeB enables founders just like you to monetize that work as inventively as you build your products. Whether you want to charge based on usage credits, outcomes, seats, or a hybrid model that combines all of those alongside complex enterprise contracts, ChargeB can help. It is the billing, metering, and monetization infrastructure purpose built for this era. Learn more at chargeb.com slash sifted. That's ch-h argeb e.com slash sifted.

11:07Unidentified speaker (UNKNOWN): How

11:07Unidentified speaker (speaker_0): will lovable compete with anthropic?

11:11Unidentified speaker (speaker_1): Um, they don't compete in my in my opinion, like the thing about like I know this comes up a lot when it comes to, hey, does this company compete with that company? I think the people who actually see those companies as

11:23Unidentified speaker (UNKNOWN): or

11:23Unidentified speaker (speaker_1): those products as competitive are usually in

11:29Unidentified speaker (UNKNOWN): like

11:30Unidentified speaker (speaker_1): the very bubble that we are we're in. Like we're like tech people. You're a reporter in tech, like I'm a venture capitalist in tech. So of course we see it that way, but most users don't see it that way. I think the pie out here is a lot bigger than any one particular product. So if you'll take a like an example, right? Like so clawed code exists. Um yet cursor also exists, and codex also exists, and anti-gravity also exists. And there's like a bunch of different ways that you can vibe code if you if you want to. Open code exists. All of these products are growing and they're very successful. I don't think the cannibalization exists, even though ostensibly you could just pick, you could do one thing with any of these products. So similarly for Lovable, I think the main difference is the kind of user is very different, they're very non-technical, Um, and they use it for a very different kind of purpose. Kind of reminds me of the conversations people would have sort of back in the day when people would be like, oh, why won't Facebook or Google do this? Which is they could. But like you could still have a very successful business doing this other thing. And

12:39Unidentified speaker (UNKNOWN): you

12:39Unidentified speaker (speaker_1): know, just because they also do a similar functionality in their app doesn't mean they will necessarily compete and win in what you do.

12:46Unidentified speaker (speaker_0): Yeah. Outside of lovable, um, I'm curious sort of what your your big time topics have have been at slush so far. Like what are you noticing about the European market? Are you guys looking for more deals over here? What's what's your take on Europe from your your couple days in Helsinki?

13:02Unidentified speaker (UNKNOWN): This

13:03Unidentified speaker (speaker_1): is my first slush.

13:04Unidentified speaker (UNKNOWN): My

13:04Unidentified speaker (speaker_1): first time in Helsinki. And I generally I'm not a big fan of tech conferences, but I've had a very ver a fantastic time in Helsinki so far. It's a tremendously well. organized conference. Um, I think the reality for me on Europe has been there are phenomenal founders and the talent in Europe has is exceptional. I think we we always sort of knew that and we'd seen evidence of that, whether it's like actual startups or like teams that we know that are based in Europe that are doing exceptional things. Um rent Rant off a list of these great companies, either the ones we're we're we're already in, um, like you know, Lagora. or um obviously Lovable or Nimodal, I know, has a Swedish office, uh UForce, like you said, but also you know, Eleven Labs or the the Black Forest Labs um of the world, Mistral's of the world. I think there's incredible talent here, and there's clearly incredible companies to be built here. Some of the problems I have started to realize with Europe is, well, one, a lot of the legal situation at a lot of the country countries in Europe aren't very startup friendly. Um, whether that's in terms of, you know, taxing on realized gains or making it easier to h hire people and help them immigrate or having, you know,

14:19labor laws that are um you know more conducive to to startups. I think a lot of that is missing in a lot a lot of the countries here. And it's also true that, you know, someone I put this to me very uh astutely is in America, when you if you are, you know, quote unquote ambitious in at least when it comes to tech, you know, you're you're either going to find yourself at SF or New York. And all of the talent sort of concentrates in these two cities. Uh Europe doesn't necessarily have the same effect. There's obviously some cities where there is concentration, whether it's Stockholm or London or what have you, but there's a lot more unwillingness to move to a city and it's very spread out. And I think for startup formation to happen at velocity, you do need a little bit of that concentration effect. And I think that is that is one those are a couple of things I feel like that are still missing in the European ecosystem to make it truly competitive. The talent is there. Um and maybe the last thing is the familiarity with the venture-venture culture is less. Um, but that's totally fixable. Aaron

15:25Unidentified speaker (UNKNOWN): Powell

15:25Unidentified speaker (speaker_0): What do you mean by venture culture?

15:27Unidentified speaker (speaker_1): I think the reality uh one thing that I've noticed uh about companies is that if you look at YC or companies that are formed in the Bay Area, a lot of those people were have known about what it takes to start a company and the frills around it for a very long time. This sense goes from everything from here's the set of ideas, here's how to pitch a certain idea, here's a set of ideas I think that are worth pitching, here's how to think about

16:02Unidentified speaker (UNKNOWN): Tam.

16:02Unidentified speaker (speaker_1): of an idea. Here's how I would talk about hiring for my idea. Um, a lot of these, you know, peripheral things really to the core idea of starting a company that become very valuable because it makes your startup legible to capital. It makes a VC understand what you're saying and and and have conviction in it. And and I see this a lot. I mean, you know, I I went to college at at Cornell and I joke that, you know, it's not like I thought we were that. much dumber than Stanford. I

16:28Unidentified speaker (UNKNOWN): mean,

16:28Unidentified speaker (speaker_1): we were not as good as them, but you know, we just weren't taught any of these things versus you would meet these guys at Stanford and then they'd be like, oh, I know all of this. I know how to pitch a company, I know how to say it's going to be worth this much money. And that is a very self-fulfilling prophecy. And of course the network and everything else too. I think that is missing a little bit in not just Europe and in many parts of the world, but is just the confidence to say you have this great. idea. How do I make it a great pitch? Um I would I would probably want to see more about that of that in in Europe. But it is up and to the right. Aaron

17:02Unidentified speaker (UNKNOWN): Ross Powell

17:02Unidentified speaker (speaker_0): Speaking of US versus Europe, uh your Silicon Valley peers, Sequoia uh just pulled the plug on their Europe office. Do you think that US VCs can effectively run a European office?

17:18Unidentified speaker (speaker_1): Yeah, there's a bunch that do it. I think it's challenging in Venture to run offices in other geographies, period, because of the way Venture uh works. It's a very like a lot of venture works on core trends. trust in your partnership.

17:42Unidentified speaker (UNKNOWN): Um

17:43Unidentified speaker (speaker_1): Namely, you know, if you have five partners in a venture fund, the reason it works is because you you believe everyone has great taste and everyone can pick great companies. And if one year you pick terrible companies, your partners will make up for it and they will want to keep you because they trust that next year you're gonna come back

18:00Unidentified speaker (UNKNOWN): um and and

18:01Unidentified speaker (speaker_1): do it. So it kind of works. as like sort of a team dynamic that way. What tends to happen when you have multiple teams and multiple geos is a lot of that camaraderie breaks. And

18:11Unidentified speaker (UNKNOWN): what

18:11Unidentified speaker (speaker_1): what starts to happen is you're not really sure what the other team is thinking. You're not really sure what the other team is doing on a day to day basis. The connection between the two teams goes goes away. And then you start asking the question, you know what, are they even good? Are they even doing their job? I feel like we're doing a better job than them. Do we really need them? And this is true with any like any venture firm and any geo. It's it it is an issue that has come up for many, many different firms. Therefore, I think it is hard to do. But there's ways to invest across the world without necessarily having an office there. Um is my belief.

18:48Unidentified speaker (speaker_0): Yeah. In terms of uh other conversations that at Slush, at least the SIFTA team has been hearing some skepticism around some of the neolabs. Um, you know, some people are kind of expressing concerns around how they'll continue fundraising at these big valuations. Do you share the concern around neo labs and what's your take on the the deep mind spinouts we keep seeing popping up with, you know, hundreds of millions?

19:11Unidentified speaker (speaker_1): I mean, I'm concerned about all startups. Every startup is destined to fail. 99% of startups will fail. That's just the reality of startups. Some startups will raise more and fail, some startups will raise less and less and fail. I think people talk more about Neo Labs because they raise more and fail. But it's TBD whether, you know, most of the most startups in general are going to fail. So I don't really like So I understand that concern,

19:36Unidentified speaker (UNKNOWN): but

19:36Unidentified speaker (speaker_1): I think it applies to everybody. And with Neolabs specifically, yes, I mean I think the hard thing about Neolabs, depending on what style of Neolab you actually are, is you still have to answer this really tough question: is if you have raised a lot of money to buy compute and to hire all this talent, how are you actually going to compete in a very purist purest way with the labs at the very frontier? Right? They have computers. compute. They have data. You have you can buy some compute. You're not gonna have as much as them. You're not gonna get the deals that they do. You can buy data, but you're probably not gonna be able to spend as much as they do on data. So okay, the the options for you to really compute boil down to either you have an algorithmic breakthrough that somehow they've not been able to come up with, or you have a fundamentally different hardware breakthrough that also facilitates algorithmic breakthrags. breakthraakthroughs. Algorithmic breakthroughs typically don't happen in a black box. You need it hardware to facilitate that. If you look at the history of neural nets, this is kind of what you see, right? Neural nets were around for such a long time, but it required hardware to get there to make them be able to show these emergent properties. So I think that's what's really tricky about a lot of them. Um now

20:47do I think all of them will fail? Um no. I mean I I I just think there's like it's obviously true that to do innovation,

20:55Unidentified speaker (UNKNOWN): off

20:56Unidentified speaker (speaker_1): there's often certain kinds of innovation that just can't happen at big companies. Uh, I do think the big labs are now big companies. When you are a big lab, you just have to focus on the next thing. You have to focus on the revenue generating thing. You can't sort of clean slate whiteboard and try to resolve this research problem. You're tied into some things, some constraints. So because of that nature, I think there are breakthroughs like unless I believe this is the final form of AI, which I don't think is true, there are breakthroughs left to be had. As long as there's breakthroughs left to be had, the exit opportunities, whether it's MA or or or something else, like or or an IPO, w continue to exist. Now of the twenty 20 or like hundred Neo Neol labs, I would say twenty of them that are actually core core AI, will all of them work out? Probably not. Um, will all of them be able to fundraise? Probably not. I mean, some of the fundraises are just ridiculous. But um but that's the case for everybody.

21:51Unidentified speaker (UNKNOWN): I

21:51Unidentified speaker (speaker_0): guess based on your sort of overview or or take on Europe here, I guess uh you know, do you guys expect to spend maybe a bit more time looking into European startups? Is anything attracting your interest? Would you guys look at some of the Neo Labs or the Deep Mind Talent coming out? What's what's your plan?

22:07Unidentified speaker (UNKNOWN): Well, look

22:07Unidentified speaker (speaker_1): at everything. I mean, what I have always said is like you never know what's exceptional unless you look at a lot of things, right? And I think our job is to invest in it. I mean, I'm I come from a very technical background. I'm like an engineer and a startup guy and a product guy. And when I care about exceptional people building exceptional things, I care. very less about where they're from. Um is a factor in like what whether you could do certain things, but I think that's been proven Europe can clearly build great companies. So could I meet exceptional people with exceptional ideas? That's what I'm looking for. Now on on specifically Neolabs out of out out of say DeepMind, um I mean it depends on the Neolab. It depends on what they're doing and what their angle is and can I can I buy it, right? Um and I go back to what I was saying about the the ways to differentiate earlier. I don't think all Neolabs are built the same and I don't will not treat them all the same. It's everything is on a case by case basis. Aaron

23:03Unidentified speaker (UNKNOWN): Powell

23:03Unidentified speaker (speaker_0): Is there a European startup that you would like to join if you are still in the the operator world? Aaron

23:08Unidentified speaker (UNKNOWN): Powell

23:08Unidentified speaker (speaker_1): Oh my god, so many. Um there's uh I I I'm pretty biased, like but modal is definitely my pick. I mean I just I account to me as European. I know maybe that's not a problem with people think it's American. Um if it if you want pure European, then uh probably 11.

23:25Unidentified speaker (speaker_0): 11 labs.

23:26Unidentified speaker (speaker_1): Yeah, I like 11 labs a lot.

23:28Unidentified speaker (speaker_0): Fantastic.

23:29Unidentified speaker (speaker_1): Um Maddie's exceptional.

23:31Unidentified speaker (speaker_0): Well, unfortunately, that is all the time that we have for today. But uh if you want to follow my reporting, you can watch that on Sifted. Uh and also please subscribe

23:38Unidentified speaker (UNKNOWN): to the video.

23:38Unidentified speaker (speaker_0): send me tips at an at sifted.eu thanks so much dd

23:42Unidentified speaker (speaker_3): thank you and thank you for having me

This page was produced automatically: the audio was transcribed with nvidia/parakeet-tdt-0.6b-v3+sortformer-v2.1 and interpreted by Qwen3.6-35B-A3B-UD-IQ3_S.gguf. Transcription and interpretation can contain errors; always confirm against the original episode. Nothing here is investment advice.
All insights