Tech news in 3 minutes
How to invest when everything is moving too fast
At a TechCrunch StrictlyVC event in Los Angeles late last week, investors Carter Reum of M13 and Chang Xu of Basis Set Ventures discussed the state of AI investing. Reum, co-founder of the early-stage firm with $2.5 billion in assets under management, and Xu, partner at the AI-focused fund with nearly $1 billion in assets, covered topics including the AI infrastructure bubble, deal pricing, and strategies to avoid being steamrolled by hyperscalers. On the bubble question, Xu described it as both a bubble and not, citing unprecedented growth curves—ChatGPT went from $1 to $40 billion in revenue in six months, and portfolio company OpenArt grew from $1 million to $70 million ARR in two years. Reum noted that while past cycles had innovators competing only with each other, this cycle pits them against the world's largest, best-funded tech incumbents, giving incumbents an advantage in tech, capital, data, and talent. For pricing, Reum uses "cocktail napkin math" to assess whether winners from the last cycle can sustain higher multiples. Xu emphasizes investing "below the AI" (infrastructure built for agents, not humans) and "above the AI" (long-term defensibility). To avoid hyperscalers, Reum favors regulated industries like 911 call centers and healthcare, citing "friction as a moat." Xu distinguishes "velocity markets" (speed of execution) from "depth markets" (hard problems), such as a portfolio company using transgenic chickens for drug manufacturing. On novel ideas, Xu noted that while many founders chase consensus categories like AI for finance or healthcare, the most interesting opportunities are initially unclear—OpenArt's prompt discovery page seemed like a non-business but hit $70 million ARR. Reum believes the current wave is still early innings, with the most exciting opportunities emerging in the second and third ripples of the technological cycle.