Air Street Capital closes $232M Fund III, becoming the largest solo-GP fund in Europe

scalable. And without further ado, we have our next guest, Nathan from Airirst Street Capital coming in to the TV. Nathan, how you doing?

Great. How are you doing, John?

Thank you so much for staying up late. Uh, what time is it there?

7:25.

Okay. Not too bad, but past the workday. Uh, uh, reintroduce yourself. It is your second time on the show, but reintroduce yourself and give us the news.

Yeah. Uh, I'm Nathan. I, uh, started a venture capital firm called Air Street Capital in 2019 to invest in AI first companies. And, uh,

what year were you investing in AI? Well, so I started the firm in 2019, but started investing in AI in 2013, which is probably around the time that like deep learning was still definitely cooking only in the lab and most people didn't really care too much outside of it.

Yeah. What were the median deal like? What was the median deal like in 2013? Was that like uh recommener systems? Like we're going to bring Netflix recommendations to everyone like that type of thing?

Yeah. Well, it was uh e-commerce recommendation systems, adtech.

Yeah.

Uh big data was the buzzword back then.

Sure.

A little bit in finance. Yeah. Like insurance underwriting, like loan prediction, credit detection, that stuff.

Yeah. Fraud detection. Just like a big uh I guess it was like a Were they doing deep learning yet or was it mostly just like

Yeah, they were. Yeah. I mean it was 2013 was the year that uh computers started to be able to recognize images better than uh than humans.

Oh yeah.

That was uh you remember like Andre Carpathy was the infamous human benchmark on imageet in 2013 PhD.

That's right.

So that was the year when uh when basically like Alex uh uh at uh University of Toronto like built AlexNet which

was the first deep learning system running on Nvidia card.

Yeah. Yeah. What what a remarkable time. So, uh, what has it has it been easy? I mean, you just raised a new fund. Has it been easier to pitch this to LPs? What have been the challenges and, uh, and opportunities over the last few years?

Yeah, I mean, it's been a it's been a sea change like in, uh, 2018, you know, when I started, um, for Air Street, it was like I'm, you know, by myself. So, solo GP, super contrarian,

starting in Europe where risk aversion is extremely high. uh trying to focus on AI, which most people didn't really care too much about, and then first-time fun. Those are sort of like all the worst like buying selection criteria that one would have. Yeah.

Yeah.

Um and then yeah, like I think, you know, this this is very much like a long-term journey like you know, I set out with I'm going to do these early stage investments, be high conviction, invest in biotech, defense, vertical software, dev infra uh you know, I stuck with what I said. So you know investing in like Cynthia 11, Black Forest and and others.

Had like six exits to like Recursion and uh you know a company that went public and Amazon etc.

Yeah.

And then um and then yeah like first fund was like $27 million. Fund two was 121 about 3 years later again pre-Chat GPT.

Um and then uh and then this one's 232 million which at this point makes us the largest solo GP in Europe.

That's amazing. So, us. You said solo GP, but you said us. Who else is on the team?

I'm you know what, like I'm kind of guilty with the royal we thing. Um,

but uh but I I have uh two colleagues who run talent and operations and then a pretty sizable back office for like admin.

Sure.

But everything that comes along with like you know building brand, planning founders, investing, fundraising, that's all me and the decisions are just me.

And then also the the is it annual report state of AI? I mean it's uh Yeah,

it's sort of a huge project. Do you bring in collaborators on that?

Yeah, that started in 2018 to basically create like a kind of canonical open access document covering research, industry, politics, and talent.

Yeah,

there are a number of contributors every year who are sort of like at the coal phase doing their PhD or like transitioning between roles in AI labs who help us kind of stay smart on things and and also folks who've been working on policy because it's become increasingly important as we see in the news almost every day. Yeah.

Um, and then the cool thing is like I get contributions from companies and labs and researchers every year and like I think this last year when we last talked there was like 50 people in the Google doc kind of like leaving comments being like hey I I tried this implementation this paper like I had this problem and then some other person's like that was my paper this is what I tried.

And it's like cool like community document basically where we can kind of get to the center of truth.

Uh not to call you like lucky on timing. Uh it's very fortunate that you have the size of fund that you do for where we are in the market cycle, but uh how hard would it be to do what you're doing today with a $27 million fund because it feels like a 20 $27 million is like a seed round for like a you know startup with just an idea. Sometimes that's like one one% of the of the seed round. Uh but like is it can you even make plays with that size fund if that's what you were constrained to today? I I think you have to decide like what I did which is either you want to be like a main player and lead rounds and express your conviction and be early etc. or you play the like large portfolio model and then you have checks and a lot of opportunities and for me that job is like much more of like a network SDR style job and less of like I can make my own opinions like do the research be there early and then when I like something like really make the bet. Yeah,

I don't think you can do the former like be a lead investor in 27 million. No chance. I think you can do the larger portfolio um you know chipping into a variety of rounds and still have like good performance,

but it's just a different job that I don't particularly enjoy um and doesn't maximalize like my strengths and my interests as much.

Yeah.

Um you know, so I I think at the end of the day like you got to pick what what flavor is good for you and then try as best as you can to bring the best product to the market given your circumstances. And you I was fortunate with fund 3 to be able to really come with like a blank slate with, you know, long-term partners and say like this is what I think is going to be the most convincing model, you know, right up to $15 million in first checks and do a couple growth stage rounds up to 25 million. Wow.

But still, you know, high conviction 20 companies and and of course like I'm, you know, mostly based in Europe, but still invest in the US and spend decent time there as well.

Yeah. I I mean I know you're you're you're in Europe but you're not you know an exclusive European investor by any means but uh I am interested in the thought exercise of like where is the AI opportunity uh internationally um if I were to just back of the envelope it I'm I'm you know I've seen some of the sovereign AI efforts it it sort of makes sense that like a certain government might want to buy from a particular uh local lab but at the same time like you know Google has been very successful internationally and you know China got the the the Google of China but many other European countries didn't get the local Google competitor or the local Amazon competitor or the local Microsoft or Apple competitor just because like they were consumer products and consumers kind of flow wherever they want. Um but at the same time I can imagine if Google or OpenAI or Anthropic is going to build a data center they might want to go to a local Neocloud or there might be opportunities for you know the Harvey of some other country that has different laws and different rules and uh so how are you seeing the shape of like opportunity outside of America in AI? I I think you covered it really well and I would generally agree with you on this uh you know Europe doesn't need its own Google per se and in fact historically like the government uh has tried there was Quero and

uh one or two other initiatives that you know had hundreds of millions of dollars pumped into them so that they could like capture European culture better than Google could and

I think that's a bit bizarre when you're talking about a learning machine. Yeah. But um but so there's certainly some sectors where I think like the sovereignty like really matters and it's not just marketing speak. So, you know, defense and security is clearly one where um you know, Europe woke up to this like two years ago and and and now even more so uh with the Middle Eastern war that's ongoing.

Uh you know, as an example, like I invested in a business called Delhi and Alliance Industries doing defense autonomous defense systems, right, in Greece. And one of the number one things we got from people outside of Greece, particularly in the US, is like, why are you investing or building a company in a vacation resort that I go sailing in, you know?

Yeah. And then that narrative changes significantly once you start seeing shahit like drones hit you know UK bases in Cyprus and you realize like okay the border to southern Europe actually comes over the Mediterranean Greece. Um I think the the other part uh which is which is interesting is like ambitions change and um and I think in Europe where traditionally there's been fewer role models that one can look to and say I'm going to be like that person and I know the path

um some companies grow and you know for example I think Lora originally started as Leia like started in Sweden right and it was the idea of

hey we should do this locally in Sweden and um and now like clearly that company's ambitions are like no we can go ahead headto-head with Harvey.

Yeah. Yeah. It does feel like in Europe like the the entrepreneurs that break out, they're not saying I'm building X for Europe. It's like I'm building 11 Labs. I'm going to go everywhere. I'm building Spotify and I'm going to go everywhere. And yes, I happen to be from some other country and I'm going to have a headquarters there, but get ready New York because we're going to have a headquarters there and we're going to have an engineering hub in SF and like we're just going to be an international company and have our roots there.

Uh and that seems great right path.

Yeah. So this is why I think if you're going to invest in Europe, it's really important to have this foothold and knowledge of the US market so that you can apply the same quality distribution

that you see in the US over in Europe. And either there are people who start from day one being that ambitious or there's others that are that grow and and kind of

uh yeah just like fuel their batteries with ambition as they see and experience it, you know, like when you get success, you want more of it. And so that's like the kind of recursive cycle that the continent's going through.

Yeah. I mean with all the progress from the big labs and they're at such incredible scale now like how are you processing the SAS apocalypse and just advice for founders of like what gets steamrolled versus what doesn't? I mean there were some there was some founder we've had a couple founders on the show that have been like we're doing AI generated video social network and then it was like you're getting steamrolled and then it was like actually like you know Sora is not going to be in the app store anymore so like maybe that was a good bet. I don't know. I'm not on that particular app but uh it feels harder and harder. It used to be just like don't do an app that's just a prompt around the foundation model. Like that's done. But now we're talking about oh is there pressure on CRM? Is there pressure on databases? Is there pressure like what will the labs do? It's it's unpredictable. But how are you working through it?

I think you know one thing you could say is like what are the problem sets and areas that like the smartest AI people want to work on and like don't do that.

Okay.

So like

that's a good one. like don't do coding and then like after it seems like AI researchers really like AI for science so like don't do that either.

Totally. Totally.

Yeah. Yeah. We have a friend from the show that does like uh Yeah. It's an AI company but it's for like small businesses like HVAC owners and helping them. It's like yeah I don't think that's on the road map.

That's really good. Um, you know, jokes aside, I think I think uh it comes down to like this tacid knowledge and like where you can capture how people do a task and taste. Like I see this a bit with our job like

you know I'm using like I'm claing as well and like codeex maxing. It's amazing how it can how you can imbue it with your taste and you're like at some point you realize okay we've built like a learning machine that basically is like a civ. you can pour as much as you want into it and it'll still like learn stuff

where before it was only one task at a time and like forget accumulating multiple tasks. So at this point if you're really AGI pill like you have a call you pipe the feedback in you ask like hey how did I do and you get suggestions you do that on the next one you pipe it back in and then you're like make a skill file and then next time

like here's a new opportunity and like dude if you're not doing that it's like game over. Yeah. And so I really do think especially for our job like there is a point at which you're like you know solo GP with like 200 or$300 million with a bunch of AIS.

Yeah.

So like

I'll call you in 10 years and see if it works.

I'm excited. I'm excited. Uh well I want to hit the gong for the new fund raise. Congratulations.

Thank you for coming on the show. Have a fantastic rest of your day and I will talk to you soon. Have a good one. Goodbye. Let me tell you about vibe.co. Co, where DDC brands, B2B startups and AI companies advertise on streaming TV, pick channels, target audiences, and measure sales just like on Meta. And let me also tell you about Plaid. Plaid powers the apps you use to spend, save,