Andrew Feldman on Building Cerebras and the Future of Chips | Ep. 57

By Uncapped

Categories: Startup, Product

Summary

To compete against entrenched giants like Nvidia, you can't compete on price or incremental improvements—you need 10-500x better performance through radical innovation. Cerebras' strategy: control everything from chip to software API, building hard technical advantages that competitors can't replicate even if they give products away.

Key Takeaways

  1. The 'Can I and Should I' framework: First identify if a new workload is computationally intensive enough to justify a custom chip, then determine if you can actually build something meaningfully better. This two-question filter prevents wasting resources on marginal improvements.
  2. Competing against Goliath requires 10-500x performance advantage, not incremental gains. Incremental improvements (even 2x cheaper) allow incumbents to bundle, cut costs, or give away products. You need a value proposition so strong they can't compete even with zero margin.
  3. Vertical integration is essential for radical innovation: control chip design, manufacturing, board, system, software, and API. This increases cost and timeline but puts all competitive advantages under your control rather than dependent on third parties.
  4. VCs need naive optimism to fund hardware: experienced investors build scar tissue from hard projects and become risk-averse. Cerebras' success required an investor who didn't fully grasp the technical complexity ahead—sometimes ignorance of difficulty is a feature, not a bug.
  5. Long product cycles require brutal honesty with boards: Cerebras had board meetings for years where the only update was 'still can't make it.' This transparency about hard engineering problems is better than false optimism and builds credibility with stakeholders.

Related topics

Transcript Excerpt

You got board meetings every six weeks and all you've got to say is still can't make it. Right? Is that that's the board meeting? I mean, what else are we going to talk about? >> Still can't make it again and again. >> And they're like, "Should I get in? Can I do anything to help?" >> That's right. They're like, "Can we help?" It's like, "Nope." But we always believed that if we could make it, there would be huge demand for it. >> All right. Super excited to be doing this today. Andrew, thanks a bunch for for being here. So Andrew, you're the, you know, CEO and founder of Cerebrris, which is obviously one of the most important chips companies. And also really happy to have my partner Eric here, who's going to be kind of half host, half guest, um, who, you know, has been working with Andrew…

More from Uncapped