At TechCrunch's StrictlyVC evening in Los Angeles late last week, two of the most direct investors in AI shared their candid views on navigating a hyper-accelerated market. Carter Reum, co-founder of M13, an early-stage firm with $2.5 billion in assets under management and a track record as seed or Series A investor in 17 unicorns, joined Chang Xu, a partner at Basis Set Ventures, which launched in 2017 as one of the first AI-focused early-stage funds and now manages nearly $1 billion across its fourth fund. In a sun-filled room in El Segundo, their conversation covered pricing deals in the fastest-moving market ever, identifying startups that can survive hyperscaler competition, and the looming impact of a SpaceX IPO on Los Angeles. This discussion has been condensed and edited for clarity.
Is there an AI infrastructure bubble?
Chang Xu: It's both a bubble and not a bubble. It's not a bubble because we've never seen this growth curve. ChatGPT went from zero to $40 billion in revenue within six months—unprecedented growth at that scale. One of our portfolio companies, OpenArt, went from $1 million to $10 million ARR in year one, then $10 million to $70 million in year two, while remaining cash-flow positive most of the time with just 20 people. The bar for what constitutes good growth has fundamentally shifted. With the possibility of compounding accelerant growth, valuations don't seem so crazy because you price that into terminal value. On the other hand, if you apply that math to every deal, no portfolio will work out well. It's a paradoxical time.
Carter Reum: I laugh because we act like this is unprecedented in venture capital, but we've seen similar cycles—cloud computing, the iPhone, even cars in the 1920s. People worried about job losses, and they happened, but life moved on. This cycle is steeper and faster, but the dynamic is the same. What's different now is that past cycles had innovators competing against innovators—Zuck versus Evan, Travis versus John Zimmer. In this cycle, innovators compete with other innovators, but also with the largest, most well-funded innovators the planet has ever seen—the ten biggest tech companies globally. For the first time in history, incumbents actually have the advantage: technology, capital, data, and talent. As quickly as some startups rise, they may fall just as fast. I find it harder to invest in a market like this—but if you get it right, you look like a genius.
How do you price deals when revenue grows faster than ever but sustainability is unclear?
Reum: We always do cocktail napkin math. The other day, we evaluated an AI software company for brands. I asked: how big were the winners last cycle? Will there be more brands globally? Are they willing to pay double or triple for software in this cycle? We couldn't make the numbers check out, so we passed.
Xu: We stay extremely close to what constitutes defensible technical differentiation—the frontier changes every quarter, sometimes every month or even week. Our framework is to invest 'below the AI' and 'above the AI.' Below the AI, infrastructure like databases, version control, and deployment tools are being completely rethought because they must handle new workloads. Above the AI, applications that layer on top of foundation models need unique data moats or distribution advantages. By 2026, we expect this bifurcation to sharpen further: infrastructure plays will consolidate around hyperscalers, while truly differentiated applications will command premium valuations.
via TechCrunch