The AI Slowdown: Market Crash or Buying Opportunity? Here's What We Would Buy

In this episode of **Being Exponential**, Luke Lango breaks down the growing debate surrounding a potential **slowdown in frontier AI development**—and why the market's initial reaction may be missing the bigger picture. We discuss the latest developments surrounding **Hugging Face and frontier AI labs**, growing concerns over increasingly powerful models, warnings about catastrophic risks to humanity, and the push for stronger **testing, oversight, and regulation** before the next generation of AI systems is released. But does slowing the pace of AI capabilities actually mean slowing the **AI investment boom**? Luke explains why increased safety requirements and regulation could actually require **more compute and infrastructure**, why building trust could make the AI growth cycle more sustainable through the end of the decade, and how a more measured development cycle could potentially reduce the risk of an eventual dot-com-style overbuild. We also examine the bear case: diminishing returns from frontier models, regulatory restrictions, weaker returns on massive AI CapEx, and the possibility that slower capability gains eventually translate into slower infrastructure spending. Finally, Luke shares his **major takeaways for investors today**, what he's watching next, and what the frontier slowdown could mean for semiconductors, memory, networking, data centers, and the broader AI trade.