In this interview, the two speakers discuss how artificial intelligence is fundamentally reshaping venture capital and startup dynamics. Speaker A argues that the old "zero‑to‑one" VC model that powered post‑2008 growth is collapsing because AI now lets companies reach $100 million ARR with only a few dozen employees, dramatically increasing capital efficiency. Speaker B expands on this, tracing the historical inflection points from on‑premise software to SaaS and now to AI‑native applications, emphasizing that AI makes both software and traditionally hard‑tech sectors (robotics, biotech, space) far cheaper and faster to develop. They highlight Garage Capital’s unique advantage: a team of engineers from the University of Waterloo who can technically diligence deep‑tech startups better than finance‑focused funds. The conversation then turns to the shortening of business cycles—AI‑driven rapid deployment means incumbents can be displaced quickly, and founders must be agile to survive. Founder‑market fit is presented as a critical lens: founders with deep domain experience are more likely to succeed. The interview also touches on enterprise reactions, where companies are automating roles with AI to cut headcount, and ends with a speculative question about the future of money in a world of abundant automation and production.
Chapters
My thesis is that venture capital as we know it, the paradigm is collapsing on itself, that these moonshot zero-to-one companies that dominated the, the post-2008 to 2021 era, that model is no longer the default mode. Now with AI and new leverageable technologies, the venture capital landscape is going to change. How do you see that happening?
There's been these incredible technological inflection points over the past 20 to 30 years. Venture capital has always been rocket fuel to accelerate the deployment of your solution so that you have no limit on supply, so that you can go and find as much demand as possible. And that's what makes a technology business so great. You can essentially have unlimited supply. So let's go back to the days of software on-premise, a server in the corner, right? The business model was on-premise, you pay a fee for the this tower and then we're going to come in and do maintenance on it. Then Amazon comes along and is like, wow, we've built this amazing server farm to help people buy books on the internet. Let's make that available to all these startup companies.
SaaS.
Boom. We're renting a server. We're going to use venture capital, rent the server. We are going to get software into everybody's hands faster than it's ever happened before. Massive inflection point, right? Venture capital booms, flares into helping companies deploy software really, really, really quickly. And when you're no longer managing all these spread servers around the world, you can move way faster. Mobile. Now we can get these solutions in everybody's pockets. This is massive for distribution, right? Again, we have unlimited supply. Let's go find some more distribution. The way I think about AI is as follows. There is a layer of applications that are now completely redundant because AI exists, right? And a lot of people might say, well, that means that software companies are no longer viable. What I see in the venture capital world and the companies that we fund is that companies are getting to $100 million in ARR with 20 to 30 employees, whereas before AI it took 1,000 and it was hyperscaling, dumping money into the market. And these businesses are obviously profitable. So there is an accelerated curve, which is why if you are not building natively with and for AI in your software startup, you are going to get left behind. So those opportunities technically for venture capitalists are better than ever because these companies are getting way more yield out of the infinite supply of their software because demand is also exploding. People are becoming more used to using these technologies. But there's another thing that's happening, which is really, really, really cool. We are seeing businesses being able to deploy really hard technologies. I'm talking robotics, things that were previously very scary for venture capitalists to fund because they're very expensive. They're error-prone. Hardware deteriorates. You're supply constrained.
Hard tech's always been a tough space.
Hard tech's always been a tough space, but hard tech is now becoming an easy space. And you hear on the All In podcast, Friedberg was talking about this is going to be the year of the autonomous AI-powered machine. I'm like, that's a robot. We've been investing in robots for 10 years. In fact, the best gains we've had as a fund have been in the technical space and robots. So this is why I think we're special at Garage Capital, if I may toot our horn just for a moment.
Beep beep, let's go.
We are 3 Fundamental first principle engineers trained at the University of Waterloo. We wear our iron rings. We know how to technically diligence a company. Mike McCauley built a robotics business. Devin and I are still building Vidyard. We're deep in AI, so we can diligence a technical asset better than most funds whose managers come from finance backgrounds, et cetera. And what we're seeing in this inflection point in technology adoption is that AI is giving engineers the ability to test more models, more outcomes, more outputs. And so they can get to market far faster and they can bring robotics and these complex technologies to market far faster. And so we're looking at biotech, we're looking at the commercialization of space. We're looking at some really cool stuff on this planet as well. It's fundamentally unlocked by venture capital. So I see this as the golden era of when the promise of technology making our lives better truly comes to fruition thanks to AI. And that's what gets me so fired up about Garage.
So many directions to go. There's so much there to unpack. I think the first question is, do you think this shortens business cycles in general? Because when I hear that it's, it's faster, like the flash to bang of deployment and demand generation is faster, that also means that incumbents can be unseated faster. And so you may be a novel solution today and somebody could, with a no-code AI platform, build a better software tomorrow and then have access to more rapid distributions. The lifecycle of a company in general may be shortened. Is that fair?
Yeah. So we look at companies, those companies, it's very high risk, low deployment strategies. If we find a great founder who's building something and we're like, okay, we don't know where the LLM research and what AI or DeepSeq, where that is going to end and where the application is going to be, application layer is going to begin. We're betting on that founder to find it in that specific space, in that specific vertical. And so there is a lot of question marks around productivity software that we are either skeptical of or deeply thoughtful and want to back founders that we know are agile enough to not get disrupted by that next generation. The kids in the garage. That's why we're called Garage Capital Fund. Kids in the garage are always the disruptors, right? Since the beginning and origins of Silicon Valley. That's why we like the hardware bets. And that's why we really like founders that have very unique experiences in a certain field of interest that go and build solutions there. Because just like my earlier story, when we're in co-op, we're in a business, we're seeing and learning all of the inefficiencies of it. That's what we call founder-market fit. And the deeper that experience is, the better we are aligned to solving that problem, the more likely we are to invest in that organization and invest in those founders.
So right now, my mind is drawing a parallel between your 2008 experience walking into your companies and seeing 50,000 people laid off overnight, saying, how can I protect from the future?
Future.
How are you thinking about that now? Let's start from a founder lens and then let's answer the question as a, from an investment, from an investor lens. So as a founder, how do I play in this new paradigm?
Yeah, it's interesting, right? Because there is a lot of momentum behind the way things used to be done. And you've got executives that are trained for 20 years on playbooks that, as far as I'm concerned, are either obsolete or becoming obsolete as we speak. And so this is like a really important distinction because startups have never been full of those executives. They've always been a couple buddies from an undergrad program getting together to build a solution, to build their company in their garage, in their garage, based on their view of the world. And they're going to take the shortest path to the outcome. Whereas this trained executive with this playbook that's dusty, 20 years old. They don't even remember how that playbook got designed from first principles. So in this moment, right, we see enterprises floundering, floundering, floundering, floundering. So as a founder who sells software to the enterprise, we closed a deal a couple of weeks ago, contract value, $750,000. That was funded through headcount budget, not from IT budget. So enterprises are trying to react because what they're worried about, and they should be, is these startups. These 10 to 20 people that are AI native, that are using agents instead of people to complete tasks where these enterprises have these massive bands of people. And Klarna talks about this— unverified, I don't know if it's true— but when they have somebody leave a team to get out of this, to disrupt themselves, to get out of this way of how things used to be done, they ask that team to go automate that person's role using AI. And if they're able to do it, they redistribute the savings to the team that solved that problem. Wow. And they don't backfill that individual. So they're basically saying you can make way more money if we have way less people. It's a bit of Lord of the Flies at this moment in the enterprise, but it's a powerful statement to what's happening in the market in general.
All right. Last question, and this is getting a little bit futurist, but I would love how you think about— this show is called Money Stories. What is the future of money? Like if we are racing toward a future where abundance is the default state where we can automate most work with machines and technology, where we can have abundant food production. We are racing toward a future where human work is no longer valuable. How do you see the future of money?
It's interesting. I think the big constraint is still energy, and I don't know that we fully comprehend that. Now, of course, we've seen models get deployed that are far more energy efficient. Than we thought AI needed to be. When you think of this future state of robotic intervention, think of this future state not being gas and diesel powered, but being battery powered. We're limited by the technologies in batteries. We haven't cracked the code on solid state. We haven't cracked the code on nuclear fusion. Until those things happen, I don't believe we can truly live in abundance. And so that's why we look at investments in micro nuclear reactors. That are looking to distributed energy supply so that difficult locations, let's say civilizations on other planets or in the middle of Saudi Arabia, have access to clean, sustainable energy so that they can start stepping ourselves towards that. But again, when you do the math, it doesn't QED as of yet. Now, 100 years, maybe 20 years, maybe. It's really about solving this energy problem in order to unlock that true state of abundance.
