Andrew Feldman
CEO and founder of Cerebras Systems
“It's 58 times larger than any other chip that had ever been... by going to wafer scale, we could use this fast memory... that's why we're 15 times faster than the fastest GPU. That's why on some problems we're 50, 100, even 1,000 times faster than graphics processing units.”
“There are three areas right now that are limiting vendors and building AI Compute. Number one is HBM... We don't use it. The second part that's limiting is a process inside of TSMC called COAS [CoWoS]... We don't use it. The third thing is... their 3 nanometer factory. We don't use it. We use 5 nanometer.”
“today TSMC has given us as many wafers as we've needed. Business today is constrained by data centers... Data centers right now are everybody's constraint in the entire industry. Powered buildings... that will not change for the next 15 or 18 months, for sure.”
“CUDA was really important in the creating of the AI landscape, but it's not important now and it has no role whatsoever in inference. If you want to move from running a model on GPUs today to running it on us, we can move it in 10 keystrokes.”
“a year ago every major Frontier Lab model had been built on a Cuda foundation and today two of three haven't. So they lost 70% market share... two of the three leading models today use no CUDA. That's a hemorrhaging of share.”
“Anthropic offered a premium service in which they offered tokens twice as fast and charged six times as much, and they sold it out and they couldn't meet the demand. Now, just to give you an idea, we're 15 times faster than they're twice as fast.”
“In December, we signed a deal with OpenAI, north of $20 billion, one of the largest contracts ever signed in Silicon Valley. And then in March, we signed a deal with AWS where they would deploy our systems in their data centers.”
“I think limiting the distribution, the diffusion of our most precious technologies makes sense and I think we have to do it thoughtfully and we have to recognize that means some markets will be foreclosed to us. And I'm okay with that.”
“They cost 30 or 40 billion dollars and take five or six years to build. So that amount of money in that amount of time cuts across administrations. And that's a problem with the politics in the US Is it's hard to make policy that's durable across administrations and across time.”
“this notion somehow that Ben proposed that speed isn't very important in agentic flows is dead wrong. That speed is important in all aspects of productive work and that your ability to get more done in less time is a fundamental advantage that accrues over time.”
“you can jump up right now and run Kimi K2. It's a 1 trillion parameter model. It's an open source model on cerebras where 10 or 15 times faster than others. And what you're paying for is the cost of our power and some cost of the compute that took to calculate it. What you're not paying for was the cost to train it.”
“The open source models, there are no open source models that are as good as the closed source models. Think of it as 3, 4%, 5% different... What is clear is that the closed source is strictly better by a little bit, by how much varies and it's more expensive.”
“You have OpenAI with their coding software, you have Anthropic with their coding software. And you've got companies like Cursor and Cognition that are using open source. We power OpenAI and we power Cognition. You have a battle underway between closed source and open source. And I think that the winners of that battle is yet to be determined.”
“There's one or two hard problems still left beyond putting GPUs in space right now. We're not super good yet at building the clusters in space necessary for the communication between [chips]... we're not good at doing it on the ground, we're really not good at doing it in space... is it one of those problems where the last 10% is 80% of the time? Now, self driving was a problem like that.”
“Historically more money is made after IPO than before... every single study shows that there is more money to be made... the opportunity to make vastly more is after IPO, not before.”
“How big is the market for slow search today? Zero. How big is the market for dial up? It's zero. How long do you wait for a website to resolve before you click away? 3 seconds, 5 seconds? You will not wait for AI. We have to deliver it to you in real time.”
“The hard part here, the hard part is moving data from memory to computer. This is the fundamental problem in AI. We solved it with a way that very few others had even attempted, which was to build a very big chip and to put memory right next to compute... So when OpenAI uses us, we're 15 or 18 times faster than a GPU.”
“We have a $25 billion backlog. And we are not alone in that... All of these players are not chasing, sort of, if you build it, they will come. The demand is booked.”
“All chips prior to us in the processor world followed Moore's Law. And we broke into doubling every 18 months... my view is in the next 18 months, we'll be way over 2x. Now, if you've got a 20 year old architecture like the GPU, it's much harder.”
“There is so much demand right now that there is no silicon that will go unused.”
“In the US we need more domestic open source models... Right now it's OSS120B or Chinese models. We run glm, we run Kimmy, we run the Quen set of models and we run OpenAI's models... Sovereignty is a trend.”
“These guys should have as much as they need, they're enormously productive. Over here, we can use maybe an open source model, maybe a cheaper model over here. And now we're sort of running it like a business.”
Andrew Feldman
One-line summary: CEO/founder of Cerebras; built the wafer-scale AI chip. Tracked for the inference-speed thesis, the constraint-routing supply argument (no HBM/CoWoS/3nm), CUDA-moat-erosion claims against Nvidia, and the data-center-as-binding-constraint framing.
What they're known for
Brief factual context — fill in.
Why they matter to stock-market
Why this person's claims are tracked here — fill in.
Said
Speaker-attributed claims extracted from diarized sources. Each bullet mirrors one entry in quotes: frontmatter — keep them in sync.
-
On cerebras, inference-demand-to-wafer-scale-advantage, inference-speed-as-a-pricing-premium:
"It's 58 times larger than any other chip that had ever been... by going to wafer scale, we could use this fast memory... that's why we're 15 times faster than the fastest GPU. That's why on some problems we're 50, 100, even 1,000 times faster than graphics processing units." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On inference-demand-to-wafer-scale-advantage, hbm-supply-bottleneck, cowos-packaging-capacity-crunch, tsmc:
"There are three areas right now that are limiting vendors and building AI Compute. Number one is HBM... We don't use it. The second part that's limiting is a process inside of TSMC called COAS [CoWoS]... We don't use it. The third thing is... their 3 nanometer factory. We don't use it. We use 5 nanometer." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On inference-demand-to-wafer-scale-advantage, ai-capex-to-power-and-materials-cascade, tsmc:
"today TSMC has given us as many wafers as we've needed. Business today is constrained by data centers... Data centers right now are everybody's constraint in the entire industry. Powered buildings... that will not change for the next 15 or 18 months, for sure." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On cuda-moat-erosion-at-inference, cuda-moat-erosion-to-nvda-rerate, nvidia:
"CUDA was really important in the creating of the AI landscape, but it's not important now and it has no role whatsoever in inference. If you want to move from running a model on GPUs today to running it on us, we can move it in 10 keystrokes." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On cuda-moat-erosion-at-inference, cuda-moat-erosion-to-nvda-rerate, nvidia:
"a year ago every major Frontier Lab model had been built on a Cuda foundation and today two of three haven't. So they lost 70% market share... two of the three leading models today use no CUDA. That's a hemorrhaging of share." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On inference-speed-as-a-pricing-premium, inference-demand-to-wafer-scale-advantage:
"Anthropic offered a premium service in which they offered tokens twice as fast and charged six times as much, and they sold it out and they couldn't meet the demand. Now, just to give you an idea, we're 15 times faster than they're twice as fast." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
"In December, we signed a deal with OpenAI, north of $20 billion, one of the largest contracts ever signed in Silicon Valley. And then in March, we signed a deal with AWS where they would deploy our systems in their data centers." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
"I think limiting the distribution, the diffusion of our most precious technologies makes sense and I think we have to do it thoughtfully and we have to recognize that means some markets will be foreclosed to us. And I'm okay with that." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On us-fab-capacity-bottleneck, tsmc:
"They cost 30 or 40 billion dollars and take five or six years to build. So that amount of money in that amount of time cuts across administrations. And that's a problem with the politics in the US Is it's hard to make policy that's durable across administrations and across time." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On inference-speed-as-a-pricing-premium:
"this notion somehow that Ben proposed that speed isn't very important in agentic flows is dead wrong. That speed is important in all aspects of productive work and that your ability to get more done in less time is a fundamental advantage that accrues over time." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On open-vs-closed-source-model-economics, cerebras:
"you can jump up right now and run Kimi K2. It's a 1 trillion parameter model. It's an open source model on cerebras where 10 or 15 times faster than others. And what you're paying for is the cost of our power and some cost of the compute that took to calculate it. What you're not paying for was the cost to train it." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On open-vs-closed-source-model-economics:
"The open source models, there are no open source models that are as good as the closed source models. Think of it as 3, 4%, 5% different... What is clear is that the closed source is strictly better by a little bit, by how much varies and it's more expensive." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On open-vs-closed-source-model-economics:
"You have OpenAI with their coding software, you have Anthropic with their coding software. And you've got companies like Cursor and Cognition that are using open source. We power OpenAI and we power Cognition. You have a battle underway between closed source and open source. And I think that the winners of that battle is yet to be determined." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On terrestrial-power-flat-to-orbital-dc-arbitrage:
"There's one or two hard problems still left beyond putting GPUs in space right now. We're not super good yet at building the clusters in space necessary for the communication between [chips]... we're not good at doing it on the ground, we're really not good at doing it in space... is it one of those problems where the last 10% is 80% of the time? Now, self driving was a problem like that." — 2026-06-06-podcast-all-in-podcast-the-ipo-comeback-why-tech-giants-are-finally (2026-06-06)
-
On ipo-comeback-public-market-value-capture:
"Historically more money is made after IPO than before... every single study shows that there is more money to be made... the opportunity to make vastly more is after IPO, not before." — 2026-06-06-podcast-all-in-podcast-the-ipo-comeback-why-tech-giants-are-finally (2026-06-06)
-
On inference-speed-as-a-pricing-premium:
"How big is the market for slow search today? Zero. How big is the market for dial up? It's zero. How long do you wait for a website to resolve before you click away? 3 seconds, 5 seconds? You will not wait for AI. We have to deliver it to you in real time." — 2026-06-06-podcast-all-in-podcast-the-ipo-comeback-why-tech-giants-are-finally (2026-06-06)
-
On inference-demand-to-wafer-scale-advantage, cerebras:
"The hard part here, the hard part is moving data from memory to computer. This is the fundamental problem in AI. We solved it with a way that very few others had even attempted, which was to build a very big chip and to put memory right next to compute... So when OpenAI uses us, we're 15 or 18 times faster than a GPU." — 2026-06-06-podcast-all-in-podcast-the-ipo-comeback-why-tech-giants-are-finally (2026-06-06)
-
On inference-asic-wave-to-tsm-demand-broadening, ai-capex-to-power-and-materials-cascade:
"We have a $25 billion backlog. And we are not alone in that... All of these players are not chasing, sort of, if you build it, they will come. The demand is booked." — 2026-07-10-podcast-all-in-podcast-open-source-wins-agi-is-here-and-scorsese-s-ai (2026-07-10)
-
On cuda-moat-erosion-at-inference, inference-speed-as-a-pricing-premium:
"All chips prior to us in the processor world followed Moore's Law. And we broke into doubling every 18 months... my view is in the next 18 months, we'll be way over 2x. Now, if you've got a 20 year old architecture like the GPU, it's much harder." — 2026-07-10-podcast-all-in-podcast-open-source-wins-agi-is-here-and-scorsese-s-ai (2026-07-10)
-
On inference-asic-wave-to-tsm-demand-broadening:
"There is so much demand right now that there is no silicon that will go unused." — 2026-07-10-podcast-all-in-podcast-open-source-wins-agi-is-here-and-scorsese-s-ai (2026-07-10)
-
On open-source-share-shift-bullish-for-compute, frontier-lab-vertical-integration-to-sovereign-ai-stack:
"In the US we need more domestic open source models... Right now it's OSS120B or Chinese models. We run glm, we run Kimmy, we run the Quen set of models and we run OpenAI's models... Sovereignty is a trend." — 2026-07-10-podcast-all-in-podcast-open-source-wins-agi-is-here-and-scorsese-s-ai (2026-07-10)
-
On enterprise-token-budgeting:
"These guys should have as much as they need, they're enormously productive. Over here, we can use maybe an open source model, maybe a cheaper model over here. And now we're sort of running it like a business." — 2026-07-10-podcast-all-in-podcast-open-source-wins-agi-is-here-and-scorsese-s-ai (2026-07-10)
Sources
- 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s
- 2026-06-06-podcast-all-in-podcast-the-ipo-comeback-why-tech-giants-are-finally
- 2026-07-10-podcast-all-in-podcast-open-source-wins-agi-is-here-and-scorsese-s-ai
Related
Cross-links — fill in.