TechCrunch's StrictlyVC evening in Los Angeles brought together two straight-talking AI investors: Carter Reum, co-founder of M13 with $2.5 billion in assets under management, and Chang Xu, partner at Basis Set Ventures, an early-stage fund with nearly $1 billion. In a sunlit room in El Segundo, they discussed how to price deals in a market that has never moved this fast, how to find companies that won't get steamrolled by hyperscalers, and what the SpaceX IPO means for L.A. Here are key insights from the conversation, edited for clarity.
Bubble or No Bubble? The AI Growth Paradox
Chang Xu described a paradoxical time: on one hand, it's not a bubble because growth curves are unprecedented — ChatGPT went from $1 to $40 billion in revenue in six months. Their portfolio company Open Art grew from $1M to $10M ARR in year one, then $10M to $70M in year two, cash-flow positive with just 20 people. The bar for good growth has completely changed. With compounding accelerant growth, valuations don't seem crazy if you price in terminal value. But if you price every deal that way, your portfolio will suffer. So it's both a bubble and not a bubble.
Sponsored Protocol
Incumbents' Advantage in Tech History
Carter Reum noted that while past cycles (cloud, iPhone, automobile) showed similar dynamics, this cycle is different because innovators compete not only with each other but also against the ten largest tech companies, which for the first time have an advantage in technology, capital, data, and talent. This makes investing harder, but if you get it right, you look like a genius. Reum advocates "cocktail napkin math" — e.g., how many brands will exist, and will they pay double or triple for software? If the math doesn't work, pass.
Investing Below and Above AI for Defensible Differentiation
Chang Xu stressed staying close to defensible technical differentiation, as the frontier changes every quarter or month. Their framework: invest "below AI" (infrastructure like databases, version control, deployment tools, which must be rethought for agents) and "above AI" (what is defensible with long-term differentiation). For instance, no one expected a new GitHub for agents, but now many strong teams are building it.
Sponsored Protocol
Avoiding Being Crushed by OpenAI, Anthropic, and Google
Depth Markets vs. Velocity Markets
Reum advises thinking about where hyperscalers go first and last. Friction as a moat works: regulated industries like 911 call centers (a near-billion exit) or healthcare offer protection. Chang Xu adds the distinction between depth markets (hard things stay hard, like drug manufacturing with transgenic chickens) and velocity markets (fast followers are faster than ever, where speed of execution is key).
Truly Novel Ideas or Updated Versions of Old Companies?
Xu admits both exist. Consensus categories (agents for finance, healthcare) attract strong founders, but the most interesting ideas initially seem impossible as businesses. Open Art started as a prompt discovery page for generative images — no one saw it as a business. Within two years it hit $70M. Venture capital is a story of bad ideas becoming good again.
Sponsored Protocol
First Ripples and Next Waves
Reum compares the tech cycle to a rock skipping on water: the heavier and faster the rock, the longer the ripples. The first wave is obvious and crowded, but the second and third ripples promise interesting returns. In two, three, or four years, entirely new businesses will emerge. For deeper insights, check out Qualcomm's acquisition of Modular and the transfer learning guide. Read the original TechCrunch article for full context.
Source: https://techcrunch.com/2026/06/23/how-to-invest-when-everything-is-moving-too-fast
