No Priors

The Story Behind Cerebras’ $63 Billion IPO with Founder and CEO Andrew Feldman

Companies in Silicon Valley from Nvidia to AMD are racing to fuel the AI revolution with postage stamp-sized AI chips. Meanwhile, a chip the size of a dinner plate just fueled a $63 billion IPO for Cerebras. Elad Gil and Sarah Guo sit down with Cerebras founder and CEO Andrew Feldman to discuss the company’s journey to making one of the largest tech go-publics in history. Andrew details the multi-year journey of pioneering wafer-scale AI computing, including surviving a brutal period of being ahead of market demand. He also explains the engineering breakthroughs that led to delivering inference speeds at 20x that of standard GPUs. Andrew then shares how a remarkable $20 billion deal with OpenAI came together in only four weeks. Plus, Andrew’s thoughts on why architecting the future of AI requires the fortitude to be a “professional David” against the Goliaths of tech. Sign up for new podcasts every week. Email feedback to show@no-priors.com Follow us on Twitter: @NoPriorsPod | @Saranormous | @EladGil | @andrewdfeldman | @Cerebras Chapters: 00:00 – Cold Open 00:41 – Andrew Feldman Introduction 00:48 – Cerebras’ Evolution 02:17 – Wafer-Scale Bet Pays Off 06:38 – Challenges and Breakthroughs 08:37 – Crossing the Market Chasm 10:38 – Scaling Software and Hardware 12:03 – Relevance of AI-Generated Coding 13:31 – Leadership and Hiring Culture 17:16 – When to Quit vs. Persist 19:40 – Why Cerebras Went Public 22:57 – The OpenAI Deal 25:54 – Open Source and Post-Trained Workloads 27:37 – How Speed Opens Up New Business 30:07 – Conclusion

AFAndrew FeldmanguestEGElad GilhostSGSarah Guohost

May 21, 2026·30m·Watch on YouTube ↗

📖 CHAPTERS

1. 1
   0:00 – 0:41

   Why fast AI changes entire business models (cold open)

   Feldman frames Cerebras’ core thesis: speed doesn’t just improve existing workflows, it enables fundamentally new products and markets. He compares fast AI to how broadband transformed Netflix from DVD delivery into a studio.

2. 2
   0:41 – 0:48

   Cerebras today: AI computers and a surge in inference demand

   Elad introduces Cerebras and its ~$63B public-market valuation, then Feldman summarizes what the company builds. He emphasizes Cerebras’ inference performance advantage—far beyond GPUs—and how real-world daily usage made latency critical.

3. 3
   0:48 – 2:17

   From contrarian ‘weird architecture’ to wafer-scale conviction

   The hosts revisit early skepticism that Cerebras’ architecture was “wrong.” Feldman argues radical performance requires radical architecture, and explains the foundational wafer-scale bet.

4. 4
   2:17 – 6:38

   Technical crucible (2017–2019): failing, burning cash, then yielding

   Feldman recounts the hardest period: years where the wafer-scale system wouldn’t work while spending ~$8M/month. The breakthrough came in summer 2019 when the team achieved yield and saw the system run for the first time.

5. 5
   6:38 – 8:37

   The market lag: being years ahead while “nobody cared”

   Even after technical success, Cerebras faced a multi-year period with limited market pull. Feldman explains why speed didn’t matter when AI was still novelty and not used daily, and why that changed around 2025.

6. 6
   8:37 – 10:38

   Crossing the chasm via supercomputing and a sovereign anchor customer

   Cerebras followed a familiar hardware adoption path: start with supercomputing and specialized high-compute industries, then try to reach mainstream volume. A pivotal partnership and large order from G42 enabled scale testing and supply-chain transformation.

7. 7
   10:38 – 12:03

   Scaling constraints in hardware: manufacturing and the long road to compilers

   Feldman contrasts software’s rapid scaling with hardware’s real-world bottlenecks. He also highlights the decade-long effort required to build the compiler and mature the software stack that makes the hardware usable.

8. 8
   12:03 – 13:31

   AI-generated coding inside Cerebras: from novelty to real leverage

   Sarah asks how much AI coding tools matter internally; Feldman says usage has exploded. He notes that the biggest gains come from engineers who learn to “govern agents” with QA and workflow changes, while others adopt more slowly.

9. 9
   13:31 – 17:16

   Leadership at ~850 people: preserving a fearless culture and hiring bar

   With rapid growth and a large backlog, Feldman focuses on cultural risk: companies often become timid as they scale. He argues against “butt in seat” hiring and describes prioritizing extraordinary ambition over safe execution.

10. 10
    17:16 – 19:40

    Founder psychology and motivating teams through long periods without validation

    Feldman discusses the loneliness of leadership and the need to love the journey, not just outcomes. He frames Cerebras as a repeated “David vs Goliath” story and argues intrinsic motivation is essential for decade-long bets.

11. 11
    19:40 – 22:57

    Quit vs persist: avoiding the ‘slippery slope’ with explicit hypotheses

    Elad probes when founders should give up; Feldman proposes a disciplined framework. If your original win-hypotheses test negative, it’s time to stop—while guarding against endless “one more test” rationalizations.

12. 12
    22:57 – 25:54

    Why Cerebras went public: cost of capital, legitimacy, and a pure-play story

    Feldman explains IPO motivation as swapping specialist private investors for broader public capital at lower cost, with stricter governance. He notes a new era where a few companies can raise public-scale private rounds, but argues public-company legitimacy still matters—especially for enterprise relationships.

13. 13
    25:54 – 27:37

    The OpenAI deal: fast inference becomes urgent and execution speeds up

    Feldman recounts how discussions with Sam Altman in mid-2025 led to trials that showcased Cerebras’ speed advantage. The parties moved from term sheet to master agreement in weeks, illustrating how the AI market has redefined operational tempo.

14. 14
    27:37 – 30:07

    Open source and post-training workloads: ecosystem pressure and creativity

    Cerebras credits open source for sustaining momentum when closed models were expensive and for pushing frontier players to improve. Feldman highlights the vibrancy created by diverse ideas and techniques, including strong advances from Chinese model makers.

15. 15
    30:07 – 30:33

    What ultra-fast inference unlocks next: reorganization of work and new markets (wrap-up)

    Feldman returns to the core theme: speed enables rethinking workflows and creating entirely new products, similar to how cloud enabled SaaS and new operating models. The hosts close with optimism about productivity leaps beyond today’s obvious replacements.