TechCrunch's StrictlyVC evening in Los Angeles late last week brought together two of the most outspoken investors working in AI today. Carter Liam is the co-founder of M13, an early-stage company with $2.5 billion in assets under management and a seed or Series A investor in 17 unicorn companies, he says. Chang Xu is a partner at Basis Set Ventures. Founded in 2017 as one of the first early-stage funds focused on AI, Basis Set Ventures is currently investing from its fourth fund and has approximately $1 billion in assets under management.
On stage in a sun-drenched room in El Segundo, the two had a funny and enlightening conversation about how to price deals in a market that has never moved so fast, how to find companies that won't be swept away by hyperscalers, and what SpaceX's IPO is going to do for Los Angeles. The conversation has been condensed and edited for clarity.
Is an AI infrastructure bubble occurring?
Zhang Xu: There are bubbles and there are non-bubbles. This is not a bubble, as we have never seen a growth curve like this before. ChatGPT's revenue grew from $1 billion to $40 billion in six months. This is truly unprecedented growth at this scale. Our portfolio company, Open Art, went from $1 million to $10 million in ARR in year one and from $10 million to $70 million in year two. [and it was] With just 20 people, most of the time you'll be cash flow positive. The criteria for good growth has completely changed. If you have the potential to compound accelerated growth, you've factored that into your terminal value, so the valuation doesn't seem all that outlandish. On the other hand, if you price every trade based on that calculation, your portfolio will never do well. So these are paradoxical times.
Carter Liam: I always laugh because we pretend like this is new to the venture capital world, but we've seen this before. In the 1920s with the cloud, iPhones, and cars, people were worried they would lose their jobs, and they did, and life went on. This is steeper and faster, but the dynamics are the same. What's different about this cycle is that in past cycles, innovators competed against each other, like Zack vs. Evan or Travis vs. John Zimmer. In this cycle, innovators compete against innovators, against the largest and best-funded innovators the planet has ever seen, and against the 10 largest technology companies on the planet. And I would argue that for the first time in history, incumbents actually have an advantage in terms of technology, capital, data, and talent. So as soon as some of these companies go up, they could go down. In fact, I think it's difficult to invest in a market like this. But if you get it right, you'll look like a genius.
How do you price a deal if your startup is generating revenue faster than ever before, but you don't know how sustainable it will be?
Rheum: We always do the cocktail napkin calculation. We were recently looking at a business called AI software for brands. I asked: How big were the winners last cycle? Will more brands emerge in the world? Are they willing to pay double or triple for software this cycle? In the end, they couldn't do the math, so no investment was made.
Xu: We are very close to defensible technological differentiation. Because that frontier changes quarterly, maybe monthly, maybe even weekly. The framework we think of is investment below AI and investment above AI. Because underneath AI, the infrastructure that is being reimagined is all built for humans, including databases, version control, and deployment tools. Agents currently use all of these infrastructures, but they require something fundamentally different. Last year, I never thought I would need a new GitHub. I can count on two hands how many really strong teams we'll get this year after becoming GitHub for agents. Beyond AI, when things get really crowded, we always go back to the next thing. What is defensible and what will lead to long-term differentiation?
How do you invest in companies that won't be crushed by OpenAI, Anthropic, or Google?
Lum: We always try to think about where they're going at the beginning and where they're going at the end. It was obvious that they would pursue marketing and obvious locations. Therefore, we have a theory of treating friction as a moat. We love regulated industries. We made just under $1 billion for a company that disrupted 911 call centers with AI. Hyperscalers may get there eventually, but as a multibillion-dollar achievement, they won't get there right away. Healthcare — they will get there, but a lot of regulations are slowing them down.
What keeps us all up at night is that things can change little by little. I could clearly see them coming in my rearview mirror. I say this to all founders: One eye will need a microscope, and the other eye will need a telescope. The microscope is for doing the day-to-day things, what you have to do this week. But the world is changing so rapidly that it's better to leave your telescope outside. You need to be a domino player and a chess player because the board changes all the time.
Xu: The frameworks we use are: Is this a depth market or a velocity market? In a velocity market, fast followers are faster than ever. It all depends on execution speed. In depth markets, hard things are still hard. In fact, complex proteins are so expensive to manufacture that some of our portfolio companies are using transgenic chickens as an alternative for pharmaceutical manufacturing. Apparently it's cheaper to have chickens do it for you. It's only today that it takes this long for chickens to hatch. [laughs]. These are depth markets and we invest accordingly.
Nevertheless, are you seeing truly novel ideas right now, or are we mostly seeing new versions of old companies?
Xu: Both. In the consensus categories, agents who applied to finance, agents who applied to health care, etc., there are a lot of really strong founders going after them, and a lot of them end up winning. But the most interesting ideas are the ones where you think, “Hmm, I don't know if that could be a business.” OpenArt, when we first helped them, Dall-E came out shortly after that, Stable Diffusion came out, and they started a search page of prompts that you could enter to get a specific type of generated image. What kind of business is that? I have no idea. They went from $1 million to $70 million in two years and have accelerated since then. The market is so deep that it cannot be understood from the outside. But from the beginning, these were young founders who experimented on the edge of what they thought was exciting and kept iterating until they found the business. If they had started a year later, they would have missed out on that period.
The story of VC is that it's a constant story of bad ideas becoming good again. Four or five years ago, I would have said it was a bad idea to invest in something that would sell to Hollywood. After that, we did a lot of creative AI and generative AI deals. That led to the current wave of companies, first generative images, then video, and now world models, that are doing incredibly well. The world was vaster than previous generations of software sold to Hollywood could have imagined. And then there's Cursor. Everyone said it was just an AI wrapper. 60 billion dollar exit. And researchers — when my husband was getting his PhD at MIT, his salary was barely above the poverty line. Nowadays, researchers are the people everyone follows on Twitter.
Lum: I think it's still in the early stages. The first wave of a technology cycle, even this steep and fast one, is usually the most obvious, has more competition, and is more crowded. The second and third ripples are where things get interesting. Remember when you were a child? If you throw a heavy stone as hard as you can to make it fly over the surface of the water, the heavier the stone becomes and the faster you throw it, the longer the ripples will be. That's the story here. I'm really looking forward to seeing business models and companies that are unimaginable today in two, three, and four years. As a VC, these second and third Ripple bets are the most difficult to get right. But then people think about it less, they pay a more reasonable valuation, and the ROI tends to be much better.
SpaceX's IPO will put a lot of money into the hands of people here in LA, especially its employees. What does that mean for this ecosystem?
Reum: When Anthropic and OpenAI eventually IPO, it's going to be a lot of VCs and institutional investors. Never before has so much money been returned and so widely spread as what is about to happen with SpaceX. If there's someone [in this room] If you have a house, boat, or plane to sell, please take advantage of it. But more importantly, every time a major liquidity event occurs, a second wave occurs. Previous LA cycles produced Riot Games, Tinder, Snap, and more. This is a different order of magnitude.
Three years ago, everyone said San Francisco was dead. It turns out the death toll was a little lower than people expected. I think the same thing applies to people who look down on LA. There are too many smart people here. Technically, of course, but there are also people who understand brands, content, creators, and influence. This first wave is a technological wave, and technical talent is concentrated elsewhere. But what comes after the technological wave? New business models, creative thinking, cultural understanding. I think that's going to be the next wave, and it's likely to be centered around Los Angeles.
Xu: What's interesting is that the next frontier for AI is not computing, but taste. It's about making movies, it's about making videos, it's about making something that resonates emotionally, it's about making something that connects with a particular culture. San Francisco has extraordinary technological talent, and that's exactly what makes it so great at automating and accelerating models. LA really has a sense of style.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.