Deep Tech Gold Rush: Smart Boom or Future Bust?

23:24Guy Perelmuter:In my opinion, it's the fact that the arc of history seems very neat and clear when you have the benefit of time, right? We can very clearly see the patterns of the first industrial revolution, the second industrial revolution, the third one, because we can see that chart looking into 50 years, right? Or a hundred years. Right now, we have our screens ticking with prices every second or less than second. And we do have ample access to information and capital. So the speed and the efficiency and the nervousness with which capital moves is extreme. And that will create backlash. That will inevitably create winners and losers.

24:12Guy Perelmuter:this is going to be as an infrastructure that is going to be the base for a new economy effectively over the next few decades. There can be very few players. So I do not believe that there's space for too many LLM providers, pure LLM providers, as there's not a lot of space for global telecommunications companies or global fiber optics companies and so on and so forth. So at the end of the day, I think the negative effects are going to be that some investors are going to be burned. Some investors are going to lose a lot of money. They're going to bet on the wrong horse, if you will. And as we have seen in history time and time again, there is a risk that this will overflow to the whole asset class, right?

25:02Guy Perelmuter:That people are going to say, okay, this asset class doesn't work. And the truth of the matter is not that the asset class doesn't work. It's that specific strategy for that specific investor didn't work. And that's my fear, that you're going to see this kind of nuclear winter, if you will, caused by these failures that will inevitably come and that could spill over into the broader venture market.

25:27Sunil Nagaraj:So Neil, one question I would ask is, because you're on the ground investing at the early stages, and a lot of what you invest in, the old kind of adage in venture investing is when you invest in a company, you have to believe there's a potential for it to return the fund or more over a long period of time, which means you need to make some assessment of how big exits can be, especially when you see valuations arise at the entry point that you and Nate are getting into. And Guy, I know you do co-investments, so you're coming in many cases alongside some of these folks here. And it's very hard to debate that the exits have gotten bigger.

26:06Sunil Nagaraj:If you look at the internet to mobile to now, they've increased 5 to 10x every single time in terms of the size of the IPO, the size of the average outlier exit. But those are a small group of companies. And we feel like right now, or I feel right now, we are in this extreme power law state. So if you're an LP and you have no exposure to OpenAI, Anthropic, Andrel, SpaceX, your returns are unlikely to look very good. When you look at maybe$2 trillion of capital going back to LPs just from those three or four companies, is that the world we are going to live in, this extreme power law, that if you don't get it right, it is almost impossible at the entry points to be able to deliver that 3 to 4x type of return at the fund level?

26:52Nate Williams:No, I definitely don't think so. I think this is where we used this word transient earlier. I think there's a reason big companies, tiny startups have a chance against big companies because big companies can innovate, use their capital, turn into light monopolies, get fat and lazy and dysfunctional and say silly things. So at many big companies, they say, that's just a$2 billion opportunity. It's not worth our time. That is a ridiculous statement. That's just a$100 billion opportunity. It's not worth our time. And folks at Google and Anthropic are going to say things like that. That's not even considering like the Uber thing.

27:25Nate Williams:When Uber came out with their first pitch deck, they were going after black cars, which is a tiny market. So either underestimating market sizes as they grow, which big companies are terrible at doing. Look at 80 billion from Mr. Zuck, pissing it away on BR, as well as also following the trends and thinking that a market isn't big enough because they use their denominator. So this is almost like the size of their market cap is actually their Achilles heel, because that's how they measure every incremental dollar, because they're not trying to add 2 billion of market cap. They're trying to add 10 % of market cap.

27:52Nate Williams:So like there's just massive diseconomies of scale that we don't talk about enough. And so I think we're temporarily swung into this model of MAG7 producing 20 % while the rest of S &P is 5%, that these companies are a trillion dollars. While if I had a$5 million exit today, you could convince yourself that doesn't matter. It turns out it does really matter. When I have a$75 million fund three and I own 15 % of it, that's a huge freaking deal. And so I think we're just in this temporary window. The headlines are a little too hypey. We're all a little confused with AI sort of really impacting our lives since December with the Claude stuff coming out at that point.

28:24Nate Williams:So we're in this kind of funny corner, but I do think the idea of a certain size fund, you make 20 bets of about 10%, one or two work, even after dilution, it can return multiples on the fund. That is going to be a perennially good model. And that means that I'm going to say normal, not small, normal M &A and normal IPO. This is 500 million, a billion, 2 billion, 5 billion. It's still going to be really exciting. And there's a way to make a lot of money for LPs and for myself in that process.

28:49Sunil Nagaraj:I tend to agree with that. and I've seen these cycles kind of play out in the same way where you have a lot of euphoria, you have some breakouts. The main risk is when people start to apply that logic to every company and say, well, now the exit is going to be 50 to 100 billion. So I don't care what I pay. I think that is a false sort of pretense to actually operate a venture strategy around. Now you may get lucky and you get one of those 50 billion, 100 billion or even more. But one of the things that we always have to look at beyond what's transit is what has actually happened over the last few years that is actually going to stay past this kind of fever pitch moment.

29:28Sunil Nagaraj:And so venture has changed. I mean, look at the number of funds that have come to market. In 2021, I'll give you a stat, 1 ,300 VC firms raised a fund in 2021. That's up from about 200 per year. The last couple years, we've been averaging about 400 to 500, which is still quite a bit in terms of the not net new funds, but just not net new firms, but funds. In 2021, it was a lot of fun ones and fun twos, though. And that was when we saw the heyday of a lot of capital people starting hanging their own shingle. What do you think is actually, let's assume a year and a half from now, the market starts to stabilize a little bit, some of this frost starts to come back a little bit, or at least retreat because it becomes clear like there are some losers, there was expensive mistakes.

30:15Sunil Nagaraj:Some of these companies looked great, but weren't real long-term durable companies. What do you think stays the same in venture capital versus the things that we are saying that are transient?

30:26Samir Kaji:Yeah, I mean, I'll take a shot at it. Look, several of us on the call have founded companies, right? And the actual mechanics and the throwing motion of starting a company or starting a movement doesn't change. It starts with an idea. It starts with convincing a co-founder. It starts with a small amount of capital and kind of get going. And so as Sunil mentioned, I don't think that changes. Something that I do think is important to double click on is if I look five, seven years ago in the timeline you talked about, Samir, when there was an explosion in challenger firms, that coincided with majority of the tier one firms had upper limits on fund size.

31:04Samir Kaji:It was impossible to get allocation into Sequoia, Kleiner, Greylock, A16Z. In the last couple of years, we've seen much higher hard caps, if any hard caps at all, because you can SMA your way to tens of billions of dollars. And so I don't think that affects the small managers. I actually think that affects the middle managers that are hundreds of millions of dollar funds. So I think that trend will play itself out. On the startup side, Sunil, know, I thought you made a really good point, which is you have to be non-consensus and right. And one of the biggest fears I have, Samir, I mentioned to you offline is we are in this part where there's a, just a big part of venture capital is being conflated with access capital, right?

31:49Samir Kaji:Like we've got the Hollywood vibes in venture capital and we've got celebrities investing and we've got like triple layer SPVs. And I think there are economic reasons why that happens, which is the fact that there's a ton of companies that are generating exceptional wealth, similar to when Goldman Sachs went public in the 90s, there's going to be venture capital firms that go public. So I think the reasons, as Guy said, are economical, but I think the knock-on effects are going to be quite terrible. Because what ends up happening at the end of any boom cycle when there's a bust, there's a sense of alienation.

32:25Samir Kaji:And so if a lot of us study financial markets, you can be risk on forever. When you're risk off, you don't get to risk on very fast. And so if we go risk off, we could be risk off for three, four years. So let me just play that forward as somebody who likes to say, as I built my firm, I made some errors in terms of judgment, right? Building out a set of capabilities for founders that founders didn't necessarily need. They don't need OKRs. They don't need book clubs. They don't need exact coaching. You know what they told me, Samir, I need you to help me with downstream capital formation. And I need you to be effectively like a part-time CRO for the company.

33:04Samir Kaji:And so I think over time through trial and error, most of us VCs kind of get it together. And then the key, and I'm facing this right now, so hold me accountable. I can't go into much detail, but I have a new EIR in my team who's coming from a very large three,$4 billion company. I think that process of how you can work with somebody to create magic is so special. That's what that's why I'm not a growth investor. I want to be there when it actually starts when the magic starts.

33:33Sunil Nagaraj:Can we double click on something you just said you said and I think the old adage certainly in venture was on consensus and right if you look at the quadrants you're either right or wrong and yeah it's either consensus or non-consensus and no one wants to be in that quadrant where you were wrong and non-consensus. That's obviously like the terrible place because then you bet against the crowd and you were wrong. And then everybody thinks you're stupid or you're not going to be able to raise more capital and actually continue to build a firm. But I want to at least posit something and see what you guys think.

34:07Sunil Nagaraj:I think you can be consensus and right and still do well in certain cases. So I do think if you're a bigger firm, you can afford to just focus on consensus where you could put the most amount of capital behind a few sort of companies that are on this hyperscaler mode. And I also think it works in bull markets. I think in bear markets, it's a little bit tougher. So I'd love to get your guys' thoughts on at least bring a little bit nuance to this consensus versus non-consensus.

34:36Guy Perelmuter:I think there's very little doubt in all of our minds that something that approaches, not perfectly, but approaches consensus is growth money, right? As you see growth firms piling onto a relatively small cohort of names, that's their consensus, right? That's inevitably or almost inevitably companies that have been de-risked at some point by folks like us that come in early, that at some point were very non-consensus, very non-obvious, because hindsight is 2020. Now the idea that SpaceX will be worth, call it a trillion dollars, give or take a couple of billion dollars, a hundred billion dollars.

35:22Guy Perelmuter:Maybe two trillion. Exactly. All that feels very natural, very obvious. And to your point, very consensus-like. And because there are, again, relatively few names, not only a few companies that are reaching that kind of stage and size and gravitas, but also few firms, relatively few firms that can raise the billions of dollars necessary to keep that going. So it's almost like a circular argument where those few firms are going to subsidize or are going to invest in those very few names. Whereas in a very traditional barbell approach, there will be a ton of other firms looking to think about what is the second or the third derivative after this movement that will be the thing we're going to be talking about in, I don't know, four or five, six years.

36:18Samir Kaji:Samir, let me put a point on that real quickly because I think you nailed it. If everything was consensus, we wouldn't have roadblocks, right? Like if we're so consensus and it has to be up and to the right, triple, double. Within three years, you raise a$50 million seed, you raise$100 million Series A. There's nothing wrong with that. What I'm saying is one of the fears that I have is if you're heuristic to underwrite the next round is that it happened within 12 or 18 months. If you look backwards, there's a variety of companies that are very successful publicly traded companies that don't fit that profile.

36:57Samir Kaji:And so to think about what we do in the craft of venture, you have to be willing to break your own rules. I'll just get, I won't call out the VC firm, but they were very specific to me that they said, if you bring me a company that's raising a series A and that company was not formed in 2024 or 2025, it's almost always an automatic pass. And so I asked the question, why happens if they took time to measure thrice cut once, low burn, et cetera? we were focusing on clock speed and the best founders can iterate super quickly. There's nothing wrong with that, but there are plenty of anti-portfolio cases where that company that was light on capital.

37:37Samir Kaji:I'm thinking of a company like Viva in the emergence portfolio. I think they went public with like$7 million of venture capital raise. I would love a company like that. So it shouldn't be Samir. You mentioned the great food at the beyond summit. It shouldn't be just because there's a buffet and you can have unlimited food. I don't need 4 ,000 calories for breakfast. Like I should be smart enough to know like I eat enough food and then I go take my meetings.

38:02Sunil Nagaraj:It's a really good point. And it actually sparked a couple of things in my head that I'd like to ask you guys about. So number one is when we think about consensus, there's different definitions of what consensus could mean. It could mean traction, like companies growing 5, 10, 15, 20x per year. And in today's world, of course, I think there's this little bit of a false positive of companies going from 5 to 100. And then you kind of realize, okay, how are they actually computing AR? Is it durable? Is it experimental? Putting that aside for a second, that has created a subculture, though, whereby now companies raise a lot of money, grow crazy, or it's not interesting.

38:43Sunil Nagaraj:I've had investors tell us like, oh, you only go three or four X year over year. That's not interesting. Where's that 10 to 100, 100 to 500, which seems a little bit silly to feel like you're having these conversations. But this is what is happening every single day. And you guys see it. So consensus right now is like super high growth, raise a lot of capital. Consensus is not just about traction in some cases. It's also where do the people spin out of? Like, again, I go back to that DeepMind OpenAI. I go to, hey, this is deep tech is consensus right now. So like now you're attracting higher valuations and you're putting on these companies on this hyper growth path, which may not be compatible with the actual businesses.

39:24Sunil Nagaraj:How is that affecting you guys on the ground when you see these companies? And now a lot of things start to become consensus and then get priced as consensus.

39:34Nate Williams:Impacts in my Ubiquiti portfolio, 40 companies, we always invest at pre-seed or seed, but then I help my babies graduate to series A. It's gotten a lot harder to raise Series A's. And I think part of it is the consensus capital. Part of it is the turnover in the industry. The average Series A person hearing one of my portfolio company's pitches has been in the industry one or two years. And so they're over-rotated on the blog post in the last one or two years. And so I even wrote my own post about this supernova thing being sort of BS, but it sort of fixated everybody on like, if you're not growing 10x a year, I can't take it to my boss, right?

40:04Nate Williams:That's what the average associate would say, which really puts a damper on everything. and I'm not even asking for accommodations for deep tech at this point. Even if you're a SaaS company doing this, you're just not getting the same level of attention. Now I'd like to think with everything, it's a cycle. I've done this pendulum motion a few times. It'll eventually come back. But for the moment, I believe there's a tainted impression of sort of what that growth looks like. Either it's not real. It's just people have read about it being 10X and they fixated or it's 10X for a while, but then it comes crashing back down.

40:30Nate Williams:I just don't, I think AI has trained some structural stuff. I think you can get more done with less. You can prototype more stuff, but I don't think that the new hurdles are the new sort of baseline norm. It used to be triple that sort of thing. I don't think it's 10x10x. I don't think even though as much as some of the big firms post would like you to believe, I think it makes for good headlines. You have a lot of folks trying to ride the buzz to get more attention, but it also does. When I wrote that post about the supernova thing being BS, I got maybe 50 emails from friends saying, thank you for saying what we've all been thinking.

40:58Nate Williams:It's sort of the quiet thing, but nobody posted publicly about it.

41:01Sunil Nagaraj:Any thoughts on that? Because I do think it's, impactful when you have so much capital going into these companies. It does change what you can invest in. And in many cases, as a seed manager, you still have to play in that non-consensus, like what is true. And by the way, you've invested in companies, Sunil, like Halter, for example. I wouldn't say sensors on cows is the most consensus type of thing. Or even if you look back, Uber, companies like Airbnb, these were not consensus companies. So going back to that Nate, how do you, when you're investing right now, and there's a huge part of the universe right now where there are consensus because of the profile of the people or the fact that it is a deep tech company.

41:45Sunil Nagaraj:And if it's like space or AI frontier labs, they kind of pass you in terms of what actually makes sense for your models. So how do you adapt to that?

41:54Samir Kaji:Yeah. I mean, I think similar to, I was knocking Hollywood for a second, but I would actually derive a couple of learnings from Hollywood, which is the best people in Hollywood think of the George Lucas's of the world and the Clive Davis, they always reinvent themselves. And so I think Samir, you have to constantly be thinking about what your value proposition either as a solo GP or as a firm is to these founders. And so what I end up thinking on this is like part of the job that I do at union is effectively reading signals from series A and series B investors, right? That could be folks who sit at Google Ventures.

42:31Samir Kaji:It could be people at Lux like Peter or Josh Wolf. It could be Leo at Eclipse. I need to understand how they're thinking about the world. And I also need to have kind of the human Rolodex, which I do have, of the types of opportunities I think fits their taste. Because this is still a taste business, right? There's not an AI that's basically selecting Series A investments for these companies. And so what I've learned over the past eight years being a full-time investor through my EIR at Kleiner Perkins is I have a job to do to downstream capital to avail them of opportunities before they're non-consensus.

43:06Samir Kaji:I'll just give you two examples. This isn't talked in my book, but you and I had talked about, I have a company called Trustpoint AI, which is one of the leaders in risk decisioning for construction. They manage the financial process of a draw request in construction and actually creates a score that's quite helpful to private credit. The investors that when we first talked about on Sandhill, that was not an area they were spending much time in. And so it was easy pass. Now that company has PMF doing great. Another company, Urban Sky, came out of Techstar space, they have a stratospheric micro balloon, they raised their Series A from Lara Hippo, all toasted to Series B.

43:42Samir Kaji:So that's an example where we actually went out of the coverage of It's got to be Lux or Eclipse or Data Collective, et cetera, and said, hey, these people are smart. There are people on Altos who have military backgrounds who understand these types of companies, and they turn out Altos is an amazing investor. So I do think what it requires is a lot more legwork and advocacy. What I try to do is make that personal. I hope when I reach out to an investor that they see that my outreach, I'm not sending a bunch of like Mixmax messages, a Samir to GPs, 30 people for one portco. I'm making a specific case.

44:20Samir Kaji:Shaheen, you are the GP at Lux who loves cars. You have a 911 pictured in your office. You have beautiful kids. This is a company for you. And I think that bespoke, I learned that from John Doerr, who for all his success, that is one of the most maniacal, sleeves rolled up, tactical GPs I've ever seen. Calling in closing like director of engineering when he's already had Google, right? Already had big hits. And so I think that's the business. And that's the part that I get really excited about is doing the little things.

44:52Guy Perelmuter:I think, Nate, there's some very kind of symmetrical happenstance now, because to your point, I think we're trending towards a venture market, especially in deep tech, where the power of this equation is going to flip somehow. in the sense that I've been a firm believer that the best founders choose their investors, right? Even at pre-seed or seed, these founders say, okay, that guy from Software Beyond the Screen, that Sunil guy, that's the guy I want to talk to. That's the guy I want to pitch my startup to because he has the chops, the experience, the knowledge, yada, yada. And then once Sunil or Nate or any one of those very early stage great investors take that founder under their wing, coach them, feed them, help them build their company.

45:55Guy Perelmuter:And when that company is ready for other stage of its history, then to Nate's point, it's up to us to cherry pick which funds are now able to do the work. Who's going to get it? Who's going to understand the founder, the company, the opportunity that may or may not be consensus at that point in time? But I think that we are heading towards a venture world where the power is going to be not on the hands of the people that have the money, but it's going to be on the hands of the people that, of course, they have money, they have the capital, but they also have the credentials, the intangibles, the network, the coaching, the mentorship, the skills that are not on paper, but that every single founder talks about.

46:45Guy Perelmuter:and that's how a lot of founders find their way into our firms because other founders will tell them, you know what, Nate is the guy you want to talk to for that particular problem or Sunil is the guy that you want to talk to that particular problem.

46:58Sunil Nagaraj:I think the other thing you probably have to think about is who are the ones with the right aligned incentives with the company? Because what we are seeing right now is companies being pushed on almost this homogeneous path to take a lot of capital, try to get that 10x growth, which can actually result in a lot of value destruction. So we talk about some of the expensive mistakes that will happen and some of these companies that we'll read about in two years, three years, that were high flyers that didn't make it. And some of it is because they took on too much risk. They imbued on too much operational risk by the amount of capital.

47:31Sunil Nagaraj:And then what ends up happening is those companies get orphaned and then they go into a point where they sell for almost nothing or they get acquihired. And I think that's going to happen. And we've talked a lot around a number of different things in terms of the venture market, deep tech. I'd like to end with asking you all kind of the same thing, because one of the maybe not so great effects of AI is like we see a lot of AI slop out there. And this is permeated through all the social media networks. Everyone has opinion that they write about venture capital. What is the thing that you see parroted the most in whether it's LinkedIn or Twitter or whatever about venture capital that you fundamentally disagree with and wish it wasn't parroted as much as it is?

48:15Sunil Nagaraj:And I'll start with Guy and then I'll go to Nate and Sunil.

48:18Guy Perelmuter:I think this idea that venture capitalists are the ones responsible for the success or failure of a company is kind of usually blown out of proportion. I think the good ones, they absolutely will have a big role in helping, building, and kind of guiding their respective companies. But I think that the headlines that capture the LinkedIn's of the world or the newsreels of the world are much more simple and very repeatable. And I will give you guys an example that we all remember, even though it was very brief. but during a relatively short period of time, the world was fixated on virtual reality.

49:01Guy Perelmuter:That was going to be the next big thing, right? Then there was this even smaller period of time when blockchain was absolutely going to change the world, right? Flip it upside down. And now it feels that, yeah, now we hit something that will actually change the world, will actually make everybody's lives different. but at the same time, this is not because a handful of GPs had their hand on it. This again, this is a trend. These are multiple currents that were trending towards a single point, a convergence of factors. And I think it's a little naive. And I think a lot of investors fall for that narrative that, oh, this is because that investor or that other investor has this touch and they were able to transform this trend into reality.

49:53Guy Perelmuter:They just wished that to be true and it became true. So that's something that I think, again, going back to my earlier comment, I think that could have a very nasty spillover effect when, not if, when the losers are clear and the money is effectively lost.

50:11Samir Kaji:I think that's a really good point. Yeah, mine's a little bit different. I really, there's two things I really don't like. I think the first one is correlated to the second. I think the first one is really that the best tech always wins, which is absolute bullshit, right? The best tech doesn't always win. That just does not happen. And because of number two, which is distribution is really going to be the kingmaker for the next 10 years. And so we have just been, I got here in the Valley from grad school in 2005. I worked at Intel, I was working on speeds and feeds of semiconductors and many core.

50:47Samir Kaji:The thing about that time is the Googleization of the only job on the field or the only position that mattered, Samir, was engineering. It was engineering, then it was product, and then it was finance, then it was marketing, et cetera. And I think we're about to see a big role reversal. Even that trope that we see where you get a very technical founder, it's like, I think I need that sales guy. I'm going to get the sales guy at series A or the sales guy or gal. And all of a sudden they raised$20 million in a mango seed. They have no traction. And so that's something that I see. I try to attack it all the time when it's ever propagated that of course, deep tech founders are always, almost always hacker hacker teams.

51:33Samir Kaji:In the 1990s, Boston 95, 128 corridor, hacker hacker teams who are PhDs with a GP who is a PhD. The problem is the world is so different now. And if you don't have that Rolodex to Schneider, to Siemens, to Bosch, Verizon, AT &T, Best Buy, Apple, that's really hard to uncork success for a founder. So I would say it's not always the best tech, it's distribution and work with people who understand that distribution.

52:03Nate Williams:both of those i agree with them i'll add one final one here this notion of you're not going to get replaced by ai you'll get replaced by someone that's using ai i think no i think some people will get replaced by ai i think a lot of vcs mark and recent among them will say yeah other jobs will get replaced but not vc it's special i don't think so i think like a lot of the stuff that i do a lot of the stuff that people around ubiquity do will get replaced by uh i want to be compassionate and delicate and thoughtful about it and retrain and things like that but i think that that quote is kind of just wrong.

52:34Nate Williams:I think we need to take on the challenge ahead, like eyes open as opposed to it. So at the moment, again, I mostly think of the agent stuff is like six months old since the Sonnet upgrade, Opus upgrade in December. We are all need to rethink kind of how we're approaching things at every level of it. It's not, it's not that you're going to get replaced by someone using AI. I think that's parroted way too often.

52:56Sunil Nagaraj:Well, there's a lot of things we could probably talk about that gets parroted out there. And these are, this is a great way to end. And we appreciate you guys coming on and recreating some of the parts of the conversation that we had over breakfast. But thanks again. And really appreciate you guys coming on.

53:11Guy Perelmuter:Thank you.

53:11Sunil Nagaraj:Awesome, guys. Thank you. Thanks for listening to another episode of Venture Unlocked. I hope you enjoyed the conversation with Nate, Sunil, and Guy. If you'd like to get Venture Unlocked content straight to your inbox, go to VentureUnlocked.substack.com and sign up or head over to Apple Podcasts or Spotify and subscribe. Thanks again for listening.