- 2 years ago
Laura Connell, Partner, Atomico, Katelin Holloway, Founding Partner, 776, James Wise, Partner, Balderton Capital Moderator: Alex Wood Morton, Executive Editor, Europe, FORTUNE
Category
🤖
TechTranscript
00:00 >> Hey, everyone. Good morning. It's great to see you.
00:03 Okay. Let's get started.
00:05 When I was preparing for this panel,
00:07 a little tweet caught my eye a couple of weeks ago from John Bradford,
00:10 a UK-based VC that I'm sure some of you in the room are familiar with.
00:14 He said, I quote,
00:16 "It might be harsh,
00:17 but I'm not sure if I've ever invested in
00:20 a startup that uses Microsoft Teams for video conferencing."
00:24 So I'm going to start with that,
00:27 and I'm going to ask Caitlin first.
00:30 Do you share that point of view?
00:31 >> Well, I share a few cap tables with Microsoft,
00:37 and so I think that the unspoken rule
00:39 is they just know I'm going to be five minutes late.
00:41 I don't give them outward shit.
00:42 I don't know if I'm allowed to say that,
00:43 but they do expect me to be tardy because I don't know where the buttons are.
00:49 Yeah.
00:50 >> James, what a point of view on this?
00:53 >> No. I sit on a government board and the government runs on Microsoft Teams.
00:57 So now that may not be a buy signal for some of you here.
01:00 You may think the UK government isn't necessarily the best model for this stuff.
01:03 But no, I think Microsoft in particular has probably been one of
01:06 the leaders in thinking about how you
01:09 proliferate next-generation AI through the enterprise.
01:12 So slowly but surely,
01:14 they're catching up and Microsoft Teams will get there as well, I'm sure.
01:16 >> Yeah. Laura, just putting Microsoft Teams to one side for a moment,
01:20 let's talk about the bigger tech stack.
01:22 Are there any red flags for you as an investor
01:24 when you are looking at some startups?
01:26 >> I don't think there's any specific application that's a red flag necessarily,
01:30 but more so that we really focus on
01:32 the overall architecture and any significant tech debt.
01:37 We want to ensure, of course,
01:38 that people are keeping up with the most innovative and frankly
01:41 cheapest applications and infrastructure to run.
01:44 Of course, it's part of the assessment,
01:46 but I've never not invested in a company for MS Teams,
01:48 even if sometimes I am frustrated by it and I'm also late.
01:52 >> The real danger, I think, is the other way around.
01:54 I've met entrepreneurs who are like, "Oh, you still use Zoom?"
01:57 I was like, "Oh, God, we're falling behind."
02:01 >> I just want to also pause for a moment on a piece that one of
02:06 our reporters here in London published yesterday looking
02:09 at investment in Gen AI in Europe.
02:12 Now, it's going to come as no surprise to anyone in this room that
02:15 Europe is falling behind when it comes to venture capital.
02:19 But James, I'm going to come to you.
02:20 You have a different perspective on this.
02:22 >> Well, actually, over the last decade of
02:24 investing in technology startups,
02:26 I think Europe's catching up.
02:27 So still significantly behind in certain areas.
02:30 But if you look at dollars deployed at almost every stage,
02:33 Europe's closed the gap with American-based companies.
02:37 Secondly, if you actually look at where
02:38 most of the dollars are coming from in the US, it's strategics.
02:41 So whether it's Amazon or Microsoft,
02:43 Salesforce or Nvidia,
02:44 the real mega rounds coming through,
02:47 in particular around Gen AI,
02:49 have been led by strategic players.
02:50 Now, obviously, we want more of that kind of
02:52 investment in European entities.
02:53 Actually, I think there'll be some announcements over
02:55 the next few weeks and months where you'll
02:57 see some European companies raise huge rounds,
02:59 absolutely huge rounds, which will
03:00 change that narrative a little bit.
03:02 Then the final point is really, look,
03:04 there are some great European-based teams,
03:06 but people mostly in the AI world are thinking globally.
03:09 So wherever the capital comes from,
03:11 as long as it's available to make sure
03:12 these companies can scale, it's probably a good thing.
03:14 >> Laura, Atomico is very well-known for
03:17 your state of European tech every year.
03:18 Share your perspective.
03:19 >> Yeah, absolutely. Look, I'd echo what James said.
03:21 Every year we publish the biggest report on
03:24 the state of tech and tech activity
03:26 across the European area.
03:28 One of the things that we saw last year was actually,
03:31 Europe took the same wallet share of
03:33 early-stage investment in AI as did the US.
03:37 So around 33 percent of all dollars invested at
03:39 the early stage went to European AI,
03:43 which is obviously something that
03:44 we focus on in terms of leading indicators.
03:46 So if the leading indicators are anything to go by,
03:48 then it should be very promising. Of course, we look at
03:50 a longer tail of interesting indicators such as
03:53 the depth of AI-based talent and actually there also,
03:55 you can see that Europe, again,
03:57 is closing the gap in terms of the sheer number
04:00 of AI ML researchers and engineers.
04:04 So very, very promising early leading activity indicators.
04:08 So you can see it coming through with
04:09 the latest crop of AI companies out of Europe that really
04:13 are on a par if not outperforming some of
04:16 our US competitors and indeed
04:17 global competitors in the same space.
04:19 And Kanan, you're joining us all the way from San Francisco.
04:22 What's your perspective on the European AI scene?
04:25 Yeah, I mean, I think it's early innings, right?
04:28 And I would agree that this is a more global,
04:30 this feels like for the first time in a long time,
04:32 there's more linking of arms across,
04:34 I mean, even just across multiple funds and
04:36 firms that are trying to participate where you
04:39 just want the best founders funded at this point.
04:41 So I feel like the geography of it is much less important.
04:45 And because it's early innings,
04:47 San Francisco is back, right?
04:49 With AI alley and valley,
04:51 and we've got our talent, of course, that's stacked.
04:53 But I do think that there is a much more global lens
04:55 now than we've had previously, which is exciting.
04:57 And I think that's the right thing.
04:59 And I want to drill down on geography on a more micro level.
05:01 Now, we were just chatting a minute ago about
05:04 the days of 2012 and the London tech scene
05:07 when there was like one, two investors in a room and that was it.
05:09 But things have obviously changed in London.
05:11 But James, give me a sense from boldness perspective,
05:14 like which clusters within Europe are you looking at?
05:16 And where are you seeing the real successes?
05:18 So we are pan-European at an early stage,
05:21 and I think we're seeing opportunities everywhere.
05:23 It really depends on where the talent wants to be.
05:25 There are some areas of more intense,
05:28 I guess, serendipity at the moment,
05:29 simply when you put lots of people who are working on
05:31 a similar problem in a city or an area together,
05:34 that does often spark some kind of innovation.
05:37 The two most intense areas are obviously Paris and London right now.
05:41 I was in Paris last week at an AI event and it was febrile.
05:45 There was so much energy and actually we had a number of
05:49 major US companies and major US investors there looking for,
05:53 not just early stage investments,
05:54 but growth stage investments in Paris.
05:55 I actually think some of that narrative will
05:57 shift over the course of the year back to London.
05:59 London still remains the hub for DeepMind or Google Brain.
06:03 It still is announced that we are having
06:06 a big new investment from Microsoft AI,
06:08 with some of the team from Pine Inflection coming back here.
06:12 Fair, Facebook's AI research unit has also got a big hub in
06:15 King's Cross and I think there'll be a number of
06:17 really large funding rounds announced for
06:19 a UK-based AI companies over the coming six months or so.
06:23 Those two hubs do seem to
06:24 dominate in terms of the attraction of talent.
06:27 But ultimately, that talent comes from
06:30 universities and hubs across Europe.
06:32 I think we've been fortunate to invest in
06:34 probably every country in Europe at
06:36 some point over the last decade and we'll continue to do so.
06:38 Caitlin, when you're making your trips over here,
06:41 is it London, Paris?
06:42 Are there any other spots that have caught your eye?
06:44 It's really interesting.
06:46 Again, with this idea that talent can live anywhere now,
06:49 so in a post-COVID world,
06:50 we finally have understood that we
06:52 can truly be anywhere and do anything.
06:54 But I do still see geography playing
06:56 a role in things that are a little bit more specific.
06:59 For example, I love investing in climate tech.
07:03 Finland has a phenomenal ecosystem
07:05 for young entrepreneurs building in climate.
07:07 But now with the advancement of AI,
07:10 and if we understand that AI is not a sector,
07:12 but a platform shift,
07:13 I'm going to Helsinki and I'm meeting with
07:15 climate tech founders who are building smart grids,
07:17 smart batteries, things that we're
07:19 deploying this technology in a much more meaningful way
07:21 towards something that that particular ecosystem
07:24 is very, very good at.
07:25 That's just one example, but I do think that we're going to
07:27 see that happen more and more often,
07:29 and their teams are growing and they're remote,
07:31 and they're distributed in a way that is much healthier now.
07:33 I think it's really exciting.
07:35 You mentioned about industries as well.
07:38 Laura, from your perspective at Atomico,
07:40 where are you seeing the biggest opportunities
07:42 in different industries across Europe?
07:44 Yeah, I mean, it's really interesting.
07:45 I can only add to what
07:47 Kate and James have already mentioned in terms of
07:49 clear concentration in London and Paris,
07:51 and then a much healthier,
07:53 to use your term, distribution of talent and
07:54 opportunity across the broader European area.
07:57 One of the things that we've noticed in particular is
08:01 a broader, we've got just more payments to pound,
08:05 so to speak, in terms of meeting the best founders,
08:08 because they really are coming from everywhere,
08:09 which again was one of the founding premises of Atomico,
08:13 founded by Nicholas Zentstrom,
08:14 who's the founder of Skype,
08:15 that these founders really can come from anywhere,
08:17 and particularly in AI.
08:18 Some of the companies that we've
08:20 invested in the last 12 to 18 months have come
08:22 from very small university towns,
08:24 namely a company called DeepL,
08:26 which is one of the fastest growing AI companies at scale,
08:29 which was actually founded in Cologne,
08:33 of all places, and I never expected to spend
08:35 as much time as I have done in Cologne
08:36 in the last year and a half,
08:38 but really quite an extraordinary, small,
08:40 very concentrated ecosystem of AI/ML engineers.
08:43 Similarly, for healthcare,
08:45 we just invested in a company called Quartii AI,
08:49 which is one of the leading AI co-pilots for health,
08:52 selling into US healthcare providers,
08:54 and that was a team of researchers and
08:56 academics actually based in Copenhagen.
08:58 So you have this broad distribution,
08:59 then pockets of specialization across the area,
09:03 which is super exciting.
09:05 I think the real shift over
09:06 the last five years has been,
09:08 venture capital really is
09:09 the art of finding non-consensus investments.
09:12 So finding unique opportunities
09:13 which no one else has seen yet.
09:15 I think five years ago,
09:16 even in 2018, it was still the case that most people said,
09:20 "Actually, you can't build
09:21 a large software business out of Europe at all."
09:23 So it was non-consensus to be investing here then.
09:25 Actually, it's now pretty consensus.
09:27 Wherever you are in the world, people think,
09:29 "Actually, wherever the talent is,
09:31 you can build great businesses,
09:32 and obviously, there's talent in Europe."
09:34 So now, the challenge to everyone on
09:36 this panel is to find the non-consensus bets
09:38 within those groups of
09:40 which there's still many to be had,
09:41 but simply their location,
09:42 I don't think really matters anymore.
09:44 I'm going to break for questions
09:46 from the audience in just one minute,
09:47 but before we do, we have to talk about regulation.
09:50 We've been talking a lot about geography.
09:51 Obviously, the UK is diverging.
09:53 Caitlin, is that a concern for you,
09:56 the UK having its own regulation piece versus the EU?
10:00 To be honest, this is something
10:03 we've talked about a lot at our firm.
10:05 I think that the underscore here really
10:07 is about things being ethically sound.
10:10 This is, like I said, this is the earliest days.
10:13 This is tip of the iceberg,
10:14 and getting these policies right is very important.
10:17 Again, I think with the linking
10:19 of the arms that we're starting to see,
10:20 having different policies develop in
10:22 different areas is not a bad thing,
10:24 because I think eventually, they will converge.
10:26 But right now, we're trying to sort out
10:27 this bedrock of how do we make sure
10:29 that this innovation is ethically sound.
10:33 So I think anything moving towards that is a positive sign.
10:36 Laura, do you have a take on the regulation?
10:37 Yeah, absolutely.
10:38 I mean, we're very involved and very active
10:41 in conversations with government across the Eurozone
10:43 and in the UK to ensure that we play our part
10:47 in conscious scaling, particularly around AI,
10:50 when there are some clear
10:51 and heavily discussed risks, of course.
10:55 I think what's really important, though,
10:56 is to consider both the downside,
10:59 and I think that's largely what the approach
11:01 that the UK and Europe have taken,
11:02 which is more of a risk-based approach,
11:04 same as the US, albeit in execution
11:05 and application of that regulation,
11:07 quite different from an institutional
11:09 requirement perspective.
11:11 Europe has obviously taken a more
11:13 centralized approach to this, top-down approach,
11:15 and the UK has been more decentralized in it.
11:18 Similarly, the US has taken a similar approach to the UK.
11:23 I think what we're seeing play out real-time
11:25 is there is some level of regulatory arbitrage going on,
11:28 given the amount of capital at play here
11:30 and the general recognition from a tech sovereignty
11:34 perspective that AI really is going to be
11:36 such an important economic driver going forward.
11:40 One of the things that we just want to ensure
11:41 that we're doing is not impeding,
11:45 or in some way constraining, the best talent
11:48 and companies that we have.
11:50 One of the things that we're particularly focused on,
11:52 as an example of not just focusing on the risk,
11:54 is also focused on how do we encourage governments
11:58 then to really drive the allocation of capital
12:01 also at the later stage, because I think one of the things
12:03 that we see very clearly in Europe
12:04 is that there is a significant capital gap
12:06 at the later stage.
12:07 I mentioned that early-stage funding is really very similar
12:10 to the US in terms of funding AI,
12:13 but actually where we see a big gap
12:14 is in the allocation of risk capital at the later stage.
12:17 And the reality of a lot of AI, as you know,
12:19 is a lot of that investment goes straight to compute,
12:22 and those computational requirements
12:24 only increase over time, and so we really need
12:27 an availability of deeper pockets of risk capital
12:30 and appetite at the later stage to keep driving
12:33 some of the developments that we're seeing in the ecosystem.
12:34 So those are some of the conversations to the upside,
12:36 actually, that we're having,
12:37 where we really want to work very closely
12:39 with government actors and indeed larger asset allocators
12:43 across Europe and the UK to make sure
12:45 that we can deliver on the promise of AI.
12:48 - Fantastic, well I hope that we do see
12:49 more late-stage funding coming across Europe.
12:52 I'm just gonna turn to our audience.
12:54 I'd love some questions from the audience.
12:56 We've got roving mics, please do stand up,
12:58 let us know your name and where you're coming from.
13:01 Thank you.
13:02 - Cool, thank you.
13:03 Bartek Argonowski from Levera.
13:05 With so much open source today,
13:09 what do investors find valuable
13:12 from a tech and AI-built perspective?
13:14 And what impact will this have on valuation multiples
13:17 at an exit stage?
13:19 - Fantastic, is that for any particular panelists?
13:21 - I guess whoever has something first, yeah?
13:24 - Yeah.
13:25 - All of your inputs would be valuable, I'm sure.
13:27 - Well, I think overall, so Bolton have long been investors
13:31 in open source businesses.
13:32 We were the first investors in MySQL,
13:34 we were the first investors in Talend,
13:35 a bunch of other more recent investments in open source.
13:38 I'm on the board of a company called DeepSat,
13:39 which develops Haystack, which after LANG changed
13:41 is the most popular LM orchestration tool,
13:43 and I would argue the most reliable.
13:45 I think there's huge overall benefits
13:48 to supporting an open source ecosystem.
13:50 It enables more and more people
13:52 to take hold of this technology and scale it,
13:54 and I think that means that we're gonna see
13:56 more and more businesses built and more value created.
13:58 I think one of the biggest challenges
14:00 in the space at the moment is how closed models
14:03 perform against open source models,
14:04 and if closed models continue to capture more value,
14:07 it's unclear to me that that unlocks a entrepreneurial wave
14:10 that venture capitalists can benefit from,
14:13 and instead may just result in existing
14:15 large technology businesses capturing most of the value.
14:18 So actually, I think it's really important
14:19 to support from a horizontal perspective.
14:22 In terms of the ability to capture value,
14:24 right now, developers are still the key piece
14:28 in building reliable AI applications,
14:33 and as long as you have a strong developer community
14:36 around that open source project that's contributing to it,
14:38 in theory, that should be a huge advantage
14:40 over a closed source model where they have
14:42 limited access to engineers to improve their products.
14:44 So we're very, very bullish on the capabilities
14:47 of open source companies to both contribute
14:48 to the ecosystem and also for themselves
14:51 to be competitive with some of the closed source
14:54 companies out there as well.
14:55 Do either of you have anything to add?
14:58 Yeah, I mean, I think I have a bit of a,
15:00 I agree with much of what James has said.
15:03 I think what we're currently reflecting on right now
15:06 is to your question, where the value really accrues
15:09 in terms of foundational models,
15:10 and where the stack, we think that
15:12 from an investment perspective,
15:14 that we can really drive incremental value,
15:16 and we see incremental value being created
15:17 over what currently is a very small number
15:19 of highly concentrated foundational models,
15:22 and including open source, of course, primarily,
15:25 which we fully subscribe to and also have been
15:27 heavily invested in historically.
15:29 So I think what we're really focused on
15:32 is more so on the picks and shovels and applications.
15:37 We're already seeing it, particularly at the latest stage,
15:39 that most founders that are building around open source
15:42 are building around open source models
15:44 and really are taking more of an orchestration approach
15:46 in terms of having the flexibility to leverage
15:49 multiple models simultaneously.
15:50 And so from an investment perspective,
15:53 what we're more so focused on is really ensuring
15:56 that we have clear commercialization
15:57 around picks and shovels and applications
16:00 rather than driving incremental dollars
16:02 to the foundational layer.
16:04 It's hard to imagine how over time,
16:07 well, right now it's hard to see exactly
16:10 how value gets captured by open source models.
16:13 I think a lot of that value instead will accrue
16:16 to, frankly, the metal rather than the model themselves.
16:19 Does that answer your question?
16:21 - Fantastic, thank you for a great question.
16:23 We've got time for one last question.
16:25 I'm gonna take the lady over here, thank you.
16:28 - Thanks for the panelist.
16:31 I'm Helen Konosopoulos.
16:32 I teach at the University of Toronto in computer science,
16:36 and I've seen a lot of IP transfer.
16:39 I'm also a startup founder who came out of the university,
16:41 out of the AI lab.
16:43 What are you seeing about transferring IP
16:46 out of universities in Europe?
16:48 'Cause I know that's been a challenge.
16:49 Have you seen there, has there been a change?
16:52 Especially in Switzerland, in Paris,
16:57 and also, of course, in the UK?
17:00 - That's a fantastic question, thank you.
17:02 I'm happy to touch on it predominantly from a UK perspective.
17:06 So we've invested in a very broad range of businesses,
17:09 which especially in some deep tech,
17:10 so zero to one technologies have been university spin outs.
17:14 And some universities have done much better than others.
17:17 So Cambridge and Imperial, I think,
17:19 are two leaders in the UK.
17:20 I think there's plenty of examples in Switzerland.
17:23 I worked closely with EPFL on a business in genomics.
17:27 And so there's various areas across Europe
17:30 where they've managed to get this balance right.
17:33 I think overall, though, most investors are wary
17:37 of investing in spin outs,
17:38 because a lot of them don't get it right.
17:39 And actually, I think there's a great case study
17:41 between Oxford and Cambridge,
17:42 and how, in particular, in their robotics department,
17:44 they got things so drastically wrong
17:46 that Cambridge is running ahead
17:50 in terms of the number of businesses
17:51 being founded out of it, and Oxford actually lagging.
17:53 So there's a ecosystem concern.
17:56 I'd say that my real focus as an investor
17:59 is on people building applications that people love.
18:02 And I think that the concerns
18:03 around where exactly the IP sits
18:05 is actually probably lower in the stack
18:07 by the time we invest at Series A, for example.
18:10 So it's not something which we are
18:11 sort of overly concerned by at Balderson,
18:14 but I think because we're later in the stack,
18:16 it's normally solved by then.
18:17 But I think as an investor in technology businesses,
18:19 I hope that more universities will follow the models
18:22 set by people like Cambridge and Imperial and EPFL
18:25 to make it easier for people to spin out.
18:27 - And Caitlin, I know the scene is very different in the US
18:30 when it comes to university startups.
18:33 What's your impression over here in Europe?
18:35 - Yeah, I was gonna say,
18:36 I actually don't have much to add to this,
18:37 because we're a young firm,
18:38 we're only three and a half years old,
18:40 and so we haven't had the opportunity
18:41 to invest in something that's spun out of a university.
18:44 But if I were to compare apples
18:47 to maybe different color apples,
18:50 I would say that we are very similar
18:52 in just being kind of wary
18:53 and just making sure that we are investing
18:55 in the very best founders where they are,
18:57 and then the IP comes lower
18:59 in that prioritization stack rank.
19:01 - Yeah, and Laura, just quickly, is it a concern for you?
19:04 - No, I mean, similarly, I think,
19:06 not the top priority for us.
19:09 But what is very good to see is obviously
19:11 European universities and British universities
19:14 have historically lagged the US
19:15 in really promoting research and commercialization,
19:18 and there's been a real step change there,
19:20 as I'm sure you've all seen in the last decade.
19:24 Most of us as investors,
19:25 basically London as an example,
19:26 are often very heavily involved with local universities,
19:28 myself at Imperial, sorry.
19:31 So I think that there really has been a massive step change
19:35 in kind of recognizing the importance
19:36 of really supporting our best innovators
19:39 and researchers at universities
19:40 and finding earlier ways for not only commercialization
19:43 and funding the best ideas,
19:45 but also in synthesizing and bringing those best researchers
19:48 from a cross-disciplinary perspective together
19:51 to really drive company formation and creation.
19:53 - Well, I think that is a fantastic point to end on.
19:56 I feel very optimistic.
19:57 It seems like funding is coming back in Europe.
19:59 We're catching back up.
20:01 We are seeing university innovation
20:03 and we just need some more late stage funding.
20:05 Thank you so much, Caitlin, James, and Laura
20:07 for being such great panelists.
20:09 Thank you.
20:10 (audience applauding)
20:13 (screen whooshing)
20:16 [BLANK_AUDIO]
Comments