Rene Haas on Arm’s AGI CPU: There’s Never a Perfect Time to Do Something Big
Rene Haas is half-propped on a sofa inside his San Jose, California office, a basketball cupped in one hand that partially obscures his face from the camera. When WIRED’s photographer first asked him to strike this pose, Haas couldn’t help but grimace—he already knew exactly what headlines would write. “Everyone’s going to joke that Arm’s CEO sleeps on the job,” he says.
Still, Haas agreed to the shot. He carved out 46 minutes of his packed schedule for our interview before politely ushering us out the door; he had a scheduled call waiting for Masayoshi Son, SoftBank’s CEO and the chairman of Arm’s board.
My meeting with Haas comes just days before the chip giant’s groundbreaking announcement: for the first time in its modern history, Arm will launch its own custom silicon. For a company that built its entire legacy licensing chip architectures to other manufacturers and never touched in-house fabrication itself, this move is one of the biggest gambles in the semiconductor industry’s recent memory. Today, Apple, Tesla, Nvidia, Microsoft, Amazon, Samsung, and Qualcomm all design or sell chips built on Arm’s architecture, whether they license full designs or pay royalties to the firm. By some estimates, there are three Arm-based chips for every person alive on the planet.
Looked at another way, building a first-party chip is actually a return to Arm’s earliest origins. The company traces its roots back to the late 1970s, when two computer architects founded Acorn Computers, which built its own microprocessor around the revolutionary RISC architecture. By the early 1990s, Acorn was struggling, and its then-CEO made the fateful pivot to licensing the firm’s designs to outside companies. Jump ahead to the mid-2010s, and Arm’s notoriously power-efficient mobile chip designs turned it into the most valuable and important chip IP company on the globe.
Arm’s path hasn’t always been smooth. After SoftBank acquired Arm in 2016 and took the formerly public company private, smartphone market growth stalled, forcing Arm to aggressively pursue new lines of business. In 2020, Nvidia attempted to acquire Arm outright, but global regulatory pushback killed the deal. When the acquisition fell apart in 2022, Haas stepped into the CEO role. He led Arm back to public markets, though SoftBank still retains 90% ownership of the company.
Haas first joined Arm in 2013, leaving Nvidia where he had led the computing product business unit. He eventually rose to lead Arm’s IP products group, the company’s profit-driving cash cow. Much like Nvidia CEO Jensen Huang, who often leans on his decades of industry experience to frame current conversations, Haas doesn’t hesitate to reference 1980s geopolitical upheaval when asked if today’s global tensions worry him about Arm’s future—his answer is a clear no. He tells me he’s met former President Donald Trump half a dozen times, but doesn’t lose sleep over the U.S. government interfering in the UK-based company’s operations. Standing tall but far from imposing, the 63-year-old often dresses in heeled Saint Laurent boots, a blazer, and wears a Panerai watch.
Chip industry insiders describe Haas as a master networker who counts many of tech’s biggest names as close friends. The Wall Street Journal once called him a “natural-born diplomat.” But with this long-rumored chip project—one of Silicon Valley’s worst-kept secrets—Arm and Haas risk alienating some of the company’s most loyal long-term partners. Can you stay friends with your biggest clients after decades of working together, only to announce you’re entering their market? Haas is convinced he can pull it off.
This interview has been condensed and lightly edited for clarity.
Lauren Goode: Since you took over as CEO, many observers say Arm has gone through a major cultural shift. Do you agree with that take?
Rene Haas: One thing I’ve learned—I intuitively understood this when I worked for Jensen, but it really clicked when I took over here—is that the CEO sets the entire tone for a company.
My leadership style, the way I approach work, was really shaped when I moved to Silicon Valley 30 years ago. I worked at a handful of startups, then spent years at Nvidia, and the common thread through all of those roles was that I worked for founder-led companies. Back then, I couldn’t exactly put it into words that I thrive in that kind of founder-driven environment. But looking back, that’s what shaped my professional DNA, that’s where I do my best work.
What does that environment look like? It’s about taking risks, chasing bold growth, moving fast in fast-changing markets, being okay with making mistakes, and most of all, being willing to place big bets. When I took over as CEO, that was exactly the kind of culture I wanted to build here.
LG: Was this Arm silicon project already in the works before you became CEO?
RH: No.
LG: How much of this was your idea, versus Masa Son’s? Was this his call?
RH: No, I’m the CEO. Was he fully aware, and deeply involved in discussing the tradeoffs and all the different angles we considered? One hundred percent. But he was more of a brainstorming partner, an ideas person, not a boss telling me “do this, don’t do that.” He doesn’t get into that level of detail. One big upside of SoftBank owning 90% of Arm is that he’s our chairman and our biggest shareholder, so I talk to him all the time. We’re pretty close.
LG: One analyst told me you talk to Masa 10 to 12 times a day. Is that accurate?
RH: Some days, that’s right. He loves talking on the phone, and I’m up at all hours anyway. He knows my personal schedule really well—he knows when I wake up, when I work out, when I go to sleep. He’s really respectful of that, actually.
LG: What made you decide this is the right time for Arm to build its own chip?
RH: Somewhere along the line, we evolved from just an IP company to a full compute platform company. Let me explain that: when you think about the role a CPU plays in any tech ecosystem, there’s such deep interdependence between the CPU hardware and the entire software ecosystem, whether that’s Windows, macOS, iOS, Android, or Linux. I don’t think we ever really acknowledged that, or fully understood how important that link was, but when I took over, I knew this was our identity, and it was something we had to lean into and advance.
So why build our own chip? When you’re a compute platform company, there are times when the whole ecosystem benefits from you building a physical product yourself. We’ve seen this play out before: Microsoft built the Surface laptop to move the entire Windows ecosystem forward, and HP, Dell, and Lenovo still make plenty of Windows laptops. Google built the Pixel phone, and Samsung still makes wildly popular Android phones. These first-party products are only a small sliver of their overall businesses, but they lift the entire ecosystem up. If Gemini is optimized for Android on Pixel, that makes Samsung’s Android phones better too. If Microsoft builds new tools or features for Surface that improve Windows, that helps HP, Lenovo, and Dell as well.
That’s the point people have always missed about us. They say “oh, Arm is just an IP company.” But the IP we sell is a full compute platform, and at the end of the day, platform companies have to do things to move their entire ecosystem forward.
LG: What’s the new chip called?
RH: It’s the Arm AGI CPU.
LG: AGI as in artificial general intelligence?
RH: That’s right. Do you like the name?
LG: I don’t mind it. But I wonder if AGI is a term that will fall out of favor eventually, as the industry coalesces around other terms for advanced superintelligence.
RH: I’m betting it won’t.
LG: Who is your target customer for this chip?
RH: Our first customer is Meta. But we also have SK Hynix, Cisco, SAP, Cloudflare—we have multiple customers lined up already. And this is a chip built specifically for data centers.
LG: What makes this chip stand out from others on the market?
RH: There are a few things, but first, it’s incredibly power-efficient. This company was built on making chips that ran on batteries for mobile phones, so efficiency is baked into our DNA. And right now, everyone is talking about how much energy AI growth requires. Any step you can take to build the world’s most power-efficient server CPU is a huge win. This will be the most power-efficient data center CPU on the market.
The other big advantage is that it’s optimized for running agentic AI. There’s a common myth going around that with the rise of AI, GPUs and accelerators are all that matter. But when you look at how AI agents run inside data centers, that’s work only a CPU can handle. GPUs aren’t going away, but you need far more CPUs to run all those agent workloads.
LG: Meta just signed two huge deals with Nvidia and AMD for both CPUs and GPUs, including Nvidia’s new Vera CPUs that are part of the larger Vera Rubin superchip.
RH: That’s right. And those Vera CPUs are built on Arm.
LG: Pardon my language, but won’t launching your own CPU anger a customer like Nvidia?
RH: Let’s go back to the platform point I made earlier— a rising tide lifts all boats.
LG: That’s a very convenient answer.
RH: But it’s not wrong. The more software that gets written for Arm, the more optimizations that are made for the Arm architecture— all of that helps anyone who builds on Arm. The AI and data center opportunity is massive. The amount of money up for grabs here is just enormous. To answer your question: will it anger Nvidia? I’d guess it will anger Intel and AMD a lot more than it will anger Nvidia.
LG: Why is that?
RH: Because we’re taking market share from them. We’re going after the x86 market that Intel has dominated for decades.
LG: So you don’t think this will upset your friend Jensen, but AMD and Intel will have a problem with this move?
RH: I used “anger” as tongue-in-cheek. Building this chip is good for the entire Arm ecosystem, and that makes it good for Jensen too. If you’ve got Nvidia’s Vera chip, which is a great product, and you’ve got the Arm AGI CPU, which is also a great product— that’s only bad news for Intel and AMD, that’s all I’m saying.
LG: Will Nvidia actually buy your CPU? Amazon executives appeared at your launch event to praise your chip— will Amazon, which already builds its own Arm-based chips, purchase yours?
RH: I’d love it if Jensen bought some, but the Vera CPU is already tightly integrated with Nvidia’s own full stack, and they have their own proprietary networking tech to connect all their chips. And Amazon will almost certainly keep using their own in-house Graviton chips.
But we didn’t build this chip for Amazon. We built it because the market for Arm-based data center CPUs is huge, and it’s massively underserved right now. Look at our early customers: Meta is going to use our chip in air-cooled racks, they don’t need Nvidia’s NVLink. Then there’s Cloudflare— they can’t buy Amazon’s Graviton, and they don’t have the resources to build their own chip in-house. This fills a gap.
LG: Who is fabricating your chip? What about co-packaging and networking?
RH: TSMC, all of it. We’re also building a reference server design, not on our own— we’re working with partners like Super Micro and Foxconn. These days, launching just a chip isn’t enough. So we worked with ecosystem partners to deliver a full turnkey solution for a complete server rack.
LG: Now you’re going to have to deal with all the new questions around volume, yield, gross margins that you never had to worry about as a pure IP company.
RH: That’s right. When I joined Arm from Nvidia in 2013, I remember thinking “Wow, there’s no scrap, no yield issues, no returns, no complicated forecasting. You just hit send and the royalty checks come in. What an amazing business this is.”
So there’s a part of me that goes, “Am I crazy for getting back into this?” But yeah, all of this is stuff we’re going to have to manage now, and I know how to do it. I know what it takes, and we’ve already built out the team internally, from operations to go-to-market, everything we need to support this new line of business.
LG: How many employees are working on this project?
RH: We’ve added around 2,000 engineers to the group that does backend design, implementation, and subsystem work. Not all of them are only on this chip— we leverage a lot of the work we do on our general compute subsystems for this project. I don’t really want to pin it down to a specific number like 500 or 700 people, because it builds on a lot of work we’ve already done across the company.
LG: A lot of industry veterans say that when you launch your first chip, it usually takes two to three generations before you see the impact you’re hoping for. Do you agree with that?
RH: Yeah, that’s generally true. But since I became CEO, we’ve been building what we call compute subsystems. These are essentially all the core building blocks of a compute engine that we can hand off to customers to build their own systems-on-chips. In some cases, we’ve already done 85% of the work for them. That’s a huge amount of pre-existing development.
The compute subsystem that the Arm AGI CPU is built on has already shipped in silicon from other Arm partners. So our confidence that this chip will work is extremely high. It’s true that new chips often take a few iterations, that’s usually the case when you’re entering a brand new market building something that’s never been made or tested before, and you don’t know how the market will receive it.
For this one? We’re extremely confident the first generation out of the gate will be very, very good, and ready for high volume production.
LG: Lord, give me the confidence of a chip CEO. Alright, let’s do a quick rapid-fire round. I’ll say a name or term, you give me a quick response. First up: Intel.
RH: Historic.
LG: RISC-V.
RH: Nascent.
LG: Still nascent? It launched back in 2014.
RH: I’m giving you one word, so nascent it stays.
LG: I don’t know if you read the WIRED article my colleague wrote last year about the history of RISC and the founding of RISC-V?
RH: Yeah, I read it.
LG: And a lot of people in that community weren’t very kind about Arm— they called you guys assholes.
RH: [Gestures to his chief of staff Saumil Shah] Him or me? We could say the same about them.
LG: Back to rapid-fire: Sam Altman.
RH: Brilliant.
LG: What makes him brilliant in your view?
RH: I know him personally. He plays the very long game. He thinks really big. He’s incredibly visionary when it comes to how he thinks about problems and technology. Yeah, I think he’s brilliant.
LG: Between political divides, geopolitical tension, a rocky macroeconomic climate, and the ongoing global memory shortage— this is a pretty chaotic time to launch a new chip. How much has that changed your plans as you get ready to bring this chip to market?
RH: It hasn’t changed anything at all. This is a long-term journey for us. The Arm AGI CPU is just the first product, this is only the beginning.
My first job out of college was at Texas Instruments in 1984. That was a boom year for semiconductors, but the early 1980s were one of the worst recessions we’d ever seen. Then in 1985, Reagan started his second term and cut defense spending, which was a huge part of TI’s business back then.
What did I learn from that? There’s never a “perfect” time to do anything big. Seriously. There’s never a good time, and there’s never a bad time. You just have to do it.
LG: I read you’re a huge fan of legendary basketball coach Phil Jackson. Why is that?
RH: What I love about Phil Jackson is that he took two already incredibly talented teams and turned them into champions. Michael Jordan wasn’t a champion before Phil—
LG: Jordan won a title at North Carolina.
RH: Yeah, he won a title at UNC, but he wasn’t the best player on that team. And Kobe Bryant wasn’t a champion before Phil Jackson either. Phil turned those superteams into champions.
LG: Have you ever met Phil Jackson?
RH: No, I haven’t.
LG: How has that not happened yet?
RH: [Nodding toward his chief of staff] Someone here isn’t doing their job right.
LG: What irritates you as a leader? And don’t say “stupid questions.”
RH: People who are happy with the status quo.
LG: Can you give an example?
RH: “We’ve always done it this way. I don’t think we can do any better. I can’t see any other way to improve this.” That’s the opposite of what we need here.
LG: It’s the opposite of taking risks. What inspires you as a leader?
RH: Making a real difference. Doing things that change the world.
LG: Everyone in Silicon Valley says that. What does that actually mean to you?
RH: Yeah, it’s a cliché, so let me distill it down to Arm. What inspires me right now is what we’re doing with this chip. It’s a huge, transformative change for the entire company.
