Cerebras CEO Andrew Feldman Talks IPO

Summary

Podcast Summary: Bloomberg Talks Episode: Cerebras CEO Andrew Feldman Talks IPO
Date: May 14, 2026
Host: Bloomberg
Guest: Andrew Feldman, CEO of Cerebras

---

Episode Overview

This episode of Bloomberg Talks features a high-profile interview with Andrew Feldman, CEO of Cerebras, on the day of the company's historic IPO. Feldman discusses the significance of Cerebras’ debut, its rapid growth in the AI hardware space, business strategy, key partnerships (notably with OpenAI), and the challenges and outlook for scaling production. The conversation offers insights into Cerebras’ technology, vertical integration, and Feldman's vision for future industry leadership.

---

Key Discussion Points & Insights

1. IPO Performance & Emotional Impact

• Cerebras priced its IPO at $185 per share, substantially above guidance; shares were indicated to open at $350.
  • Feldman’s reaction: “Pretty good day, huh?” (00:56, Feldman)
  • Significance: The biggest semiconductor IPO ever, one of the largest tech IPOs in history.
  • Quote:

    “We couldn't be more proud. This is the culmination of a decade of work, of countless late nights and long weekends.” (01:20, Feldman)

2. Cerebras’ Position in the AI Ecosystem

• Market Perception: Feldman articulates how Cerebras is now a major player, with an “extraordinary demand right now for fast inference,” and claims performance is “15 times faster than the nearest competitor.” (02:18–02:50)
  • Bloomberg probes the credibility of this market position and the importance of partnerships like OpenAI.

3. Major Partnerships and Revenue Timing

• OpenAI Engagement:
  • Recent deal for $20B+ with OpenAI for 750 megawatts of compute capacity.
  • Other large-scale deals and “dozens” in the $10–$50 million range.
  • On revenue outlook:

    “We are extremely confident that they will be an enormous channel for us and a partner in delivering our technology… globally.” (03:31, Feldman)

  • Revenue recognition depends on operationalization and contract finalization.

4. Retail Investor Inclusion

• Listener Question (05:00): Why wasn’t there more focus on retail investors?
  • IPO was 25x oversubscribed; difficult allocation decisions were made.

  • “We thought we did it with integrity and I think we are very comfortable with… the ending allocation.” (05:08, Feldman)

5. Business Model: Full Stack, Vertical Integration

• Unique Approach: Cerebras builds entire supercomputers, not just chips.
  • Compares to Nvidia (which focuses on GPUs, with third parties like Dell or Supermicro assembling systems) and highlights different margin structures.
  • Current gross margin sits around 40–41%, with Feldman confident about improvements as scale is achieved.

  • “As we grow, we will have more leverage in the supply chain and our cost of goods will come down… I think the demand for fast inference is overwhelming.” (06:09, Feldman)

6. Technology Differentiation: Wafer-Scale Engine

• Unique Chip Design:
  • Cerebras builds chips the “size of a dinner plate,” compared to typical “postage stamp” chips.

  • “Every previous effort had failed to build a chip of this size… Not only had everybody failed until we succeeded, but several tried to copy us and have since failed as well.” (07:17, Feldman)

  • Why not just sell chips? True performance gain comes from optimizing the full system, akin to building “the entire Porsche, not just the engine.”

7. IPO Proceeds & Expansion Plans

• Use of Funds / Scalability:
  • $5.5 billion will go toward scaling capacity and onboarding new customers.

  • “There’s tremendous demand for what we’re doing. I think we can be aggressive on that front.” (09:10, Feldman)

---

Notable Quotes & Memorable Moments

• On the IPO Achievement:

  “This is the culmination of a decade of work, of countless late nights and long weekends. We are enormously proud and excited and ready to get back to work and start working on the next… great thing.”
  (01:19, Feldman)

• On Technology Leadership:

  “We’re the fastest, not by a little bit, but by more than an order of magnitude. You know, we’re 15 times faster than the next nearest competitor.”
  (02:32, Feldman)

• On Industry Challenges:

  “For the entire 70-year history of the computer industry, every previous effort had failed to build a chip of this size... Part of what we were capable of doing and able to do was that we were able to use not just the chip, but also our expertise in packaging and in system design...”
  (07:17, Feldman)

• On Strategic Vision:

  “You don’t just get 15 or 18 or 20 times faster than the competition because you built a good chip. That’s one of a collection of different things that enable that sort of performance.”
  (07:43, Feldman)

---

Timestamps for Key Segments

• 00:32 – 01:19: IPO pricing, initial market reaction, feelings and context
• 01:50 – 03:07: Cerebras’ position in AI, major OpenAI deal, competitive benchmarks
• 03:07 – 04:21: Revenue recognition and details with enterprise cloud partners
• 05:00 – 05:34: Retail investor allocation decision
• 05:34 – 06:54: Business model discussion: vertical integration, margin outlook
• 06:54 – 08:56: Technical details: wafer-scale chip, system architecture, analogy to Porsche
• 08:56 – 09:25: IPO proceeds usage and capacity growth plans

---

Concluding Insights

Feldman’s interview underscores Cerebras' ambition to become the go-to infrastructure provider for fast AI inference globally, touting technological breakthroughs, aggressive scaling plans, and a vertically integrated strategy. The IPO marks both a milestone and a springboard, with Feldman projecting confidence in the face of immense demand and complex execution challenges.

---

Best for listeners interested in:

• Cutting-edge AI hardware
• IPO strategies and market dynamics
• The practical business of scaling next-gen technology

(Episode skips advertisements and non-content sections, focusing solely on the interview’s main content.)

Transcript

[A]

Indiana University strengthens tomorrow's workforce with practical real world experience. IU grads make a difference in your community, serving as teachers, nurses and engineers who rise to tomorrow's challenges and meet them. Learn more at iu.edu impact

[B]

Bloomberg Audio Studios Podcasts Radio news.

[A]

Welcome back to Bloomberg Tech to our TV viewers and radio listeners around the world. We're joined by Andrew Feldman, Cerebra CEO. Cerebras Price this IPO Andrew $185 a share above the top end of the market range. I'm looking at the Bloomberg terminal shares indicated to open $350 a share. Your reaction to that?

[B]

Pretty good day, huh?

[A]

How you feel? I mean, did you see this coming? There's some mechanics to it, right? But you know, I know you, I know you like to, let's be honest, you like to talk about what's happening in AI. The big focus on inference, cerebral his place in it. But this is the biggest IPO of the year so far. You know, what does it mean to you and your employees and your backers?

[B]

It's the biggest tech, one of the biggest tech IPOs in history and it's the biggest semi IPO in history. We couldn't be more proud. This is the culmination of a decade of work, of countless late nights and long weekends. We are enormously proud and excited and ready to get back to work and start working on the next. On the next great thing.

[A]

Again, the market is indicating shares to open at $350 a share. Let's see where we're at when trading starts. But you're being priced in as a major player in this field. Now what evidence would you point to that you are in fact a major player? A lot of focus is on the concentration in the relationship with Open Air. Tell me about some of the other Frontier Labs or other names that you have some concrete talks with.

[B]

Sure. In the last months we announced a deal with OpenAI that's north of $20 billion for 750 megawatts of compute. We also announced a major engagement with us where our equipment would be deployed in their data centers. I think those are obviously the largest, but there are dozens of others that are in what used to be a big deal in the 10 to $50 million category. And so there's just an extraordinary demand right now for fast inference. We're the fastest, not by a little bit, but by more than an order of magnitude. You know, we're 15 times faster than the next nearest competitor. And as AI has become useful, everybody wants it to be fast.

[A]

Nobody wants to wait on that, you know, 21x on, on, you know, performance and also like a lot of emphasis on the dollar per token. Right. That, that is the metric that the field cares about. But that's on paper. You know, you just talked about the relationship as an engagement. When does the word engagement end and real revenues start to show up for you from those relationships?

[B]

Sure. We, we signed a binding term sheet with us as described in our RS1 and we're working through the master agreement. I think in dealing with organizations of that size takes little time to dot all the ISE and cross the T's. But we are extremely confident that they will be an enormous channel for us and a partner in delivering our technology to, to large enterprises and medium sized enterprises around the world. I think they are one of the preferred cloud providers for just about every enterprise on earth. And so an opportunity to have your solution embedded in their offering as their bedrock offering. That's a huge win for us.

[A]

We're live on Bloomberg Television and radio. This is a Bloomberg Tech takeover and we're speaking to Andrew Feldman, the CEO of Cerebrus whose company just priced the IPO at $185 a share and is indicated right now. I'm looking on the Bloomberg terminal, maybe to start trade at $350 a share. Andrew, I think you know this about me. Whenever we have a big moment to have a conversation like this, I always go to the Bloomberg Tech audience. What do they want to know? And actually the first question is, what was your attitude toward retail investors? Why did you not do more for the retail investor in this big moment for your company?

[B]

We were more than 25x oversubscribed. There was a lot of hard decisions that needed to be made and nobody got what they wanted. And you know, we did our best. It's about all you can do when, when there's 25 times more demand for your stock at the institutional level and at the, at the retail level than there are shares to be sold. So we're really proud of the way we chose to do it. We thought we did it with integrity and I think we are very comfortable with, with the ending allocation.

[A]

Full stack, fully vertically integrated. And what I mean by that is you literally build the supercomputer top to tail, right. So Nvidia will do the tray, not just gpu, but then Dell Supermicro will assemble it. Dell's margins, low teens. Nvidia's margins mid-70s. Your margins 40, 41%. And what I'm trying to help the audience understand is why that strategy of owning everything top to tail is going to pay off in the long run. And I would say the other question I get for you is what's the future outlook for margins based on your plans?

[B]

Well, I think a couple of things. I think we have obviously opportunities at, to improve our cost structure. You know, we did half a billion last year in sales. That means we put 250 million in the supply chain. Obviously that's not an efficient spot. As we grow, we will have more leverage in the supply chain and our cost of goods will come down. I think we have an opportunity to, to increase prices. I think the demand for, for fast inference is overwhelming this minute. And so I think in the, in the long run we'll be really proud of our, of our gross margins and where they will sort of wash out as we hit scale.

[A]

I've got a question for you from, from a terminal client via ib. Thank you for the question. Everyone's tuned in right now, Andrew. They want, they want, they want the detail. And so one way that people look at it is that, you know, this is custom silicon, but Cerebras isn't, isn't a chip per say. It's like wafer level, right? Is there a reason why you can't just sell that versus the whole server?

[B]

Yeah, that's a good question. For the entire 70 year history of the computer industry, every previous effort had failed to build a chip of this size. So for your audience, this chip is the size of a dinner plate, while traditional chips are the size of a postage stamp. And not only had everybody failed until we succeeded, but several tried to copy us and have since failed as well. Part of what we were capable of doing and able to do was that we were able to use not just the chip, but also our expertise in packaging and in system design to solve some of the problems and build a truly compelling solution. Right? You don't just get 15 or 18 or 20 times faster than the competition because you built a good chip. That's one of a collection of different things that enable that sort of performance. You can build a great chip and the system vendor, the odm can nibble away at your performance by not delivering the right amount of power or the right amount of IO. Right? There's, there's a reason why Nvidia then sought to control the IO, the NV link, because they didn't want others to nibble away at their performance. And so by, by building the system, we were able to optimize all parts. I mean, you could ask the same thing about Porsche. You know, why don't you sell just engines business, right? Andrew Turns out that that's an i11 is a, is a beautiful car because of the engine and everything else they put in it.

[A]

Andrew, the team's telling me we got to go. But I got to ask you, $5.5 billion. What do you use the proceeds for? And actually, how flexible can you now be in allocating capacity to new customers? We just have 30 seconds.

[B]

Oh, I think we will use it to increase capacity. We are excited to bring many new customers on board. There's tremendous demand for what we're doing. I think we can be aggressive on that front.

[A]

Andrew Feldman, Cerebra CEO Pricing the IPO185 indicated to open $350 a share. Thank you for your time.

[C]

A business gift should do more than check a box. It should reflect your brand and show someone they're appreciated, recognized and truly seen. 4imprint offers thousands of high quality customizable products like premium apparel, drinkware, tech and more, making it easy to create a GIF that feels meaningful and on brand. And with 4imprint's expert support and their 360 degree guarantee, you can be 4imprint certain your order will arrive exactly as intended. Explore gifting with [e-mail] 4imprint4 certain.