Sep 3, 2026 · 37m · no-priors
Redefining Chip Architecture with Arm CEO Rene Haas
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Arm CEO Rene Haas explores the crucial orchestration role of CPUs in AI, Arm's strategic evolution from IP licensing to physical silicon, and the physical supply chain and power constraints governing technological scaling.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 24.7% of the talking time here. How this is scored →
The hosts as informed peer 5.3 Guest teaching 4.3 Guest disagreement 0.3 The hosts pushing back 0.1
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 27:20 Haas rejects claims that data centers do not create local jobs
Haas emphatically pushes back against the common characterization of data centers as empty automated warehouses, calling BS and pointing out the massive peripheral engineering and trades ecosystem.
Hardest push from the hosts ▶ 2:37 Elad presses on the rationale behind Arm breaking its pure licensing modelElad challenges Arm's fundamental strategic pivot, asking why a company that succeeded for decades on pure IP licensing decided to enter physical chip manufacturing.
Biggest teaching moment ▶ 35:00 Haas reframes the AI compute narrative around CPU orchestrationHaas educates on basic computer architecture, explaining that while the market fixates on GPU token factories, CPUs remain indispensable as the logistics trucks orchestrating token routing.
The host holds their own ▶ 6:55 Elad cites formal verification at Amazon and new EDA toolingElad demonstrates strong industry knowledge by referencing specific corporate EDA practices and formal verification adoption before querying Haas about internal implementation.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Arm's Position in the Global Semiconductor Supply Chain | 5 | 5 | 0 | 1 | Sarah sets the stage asking about Arm's place in the supply chain, and Elad demonstrates domain familiarity by asking why Arm pivoted from pure IP licensing to designing physical CPUs. Haas explains the strategic shift from components to compute subsystems and custom silicon like Meta's agentic CPU. | |
| Ecosystem Dynamics and Customer Reception of Arm's Hardware | 6 | 3 | 0 | 0 | Sarah highlights Arm's historic 98.5% gross margin as an IP business, prompting Haas to detail the operational complexities and talent required when transitioning into physical fabless manufacturing. | |
| Accelerating Chip Verification and Design Cycles with AI | 6 | 4 | 0 | 0 | Elad demonstrates technical knowledge of EDA and formal verification tools. Haas explains that while RTL generation remains early, AI verification and debugging dramatically compress 24-36 month development timelines. | |
| Fine-Tuning Models on Proprietary IP and the Future of RTL Generation | 5 | 4 | 0 | 0 | Elad asks about model fine-tuning and future design cycle compression down to 6-12 months. Haas tempers the timeline, predicting that automated GDSII generation from ideas for simpler designs is still 5 to 10 years away. | |
| Hyperscaler Custom Silicon and Enduring Supply Chain Constraints | 6 | 5 | 0 | 0 | Sarah and Elad list cascading hardware bottlenecks from advanced packaging to memory and fab allocations. Haas elaborates that infrastructure and physical data center construction are emerging as the next primary constraint. | |
| Capital Strategy and SoftBank's Neo Cloud Platform | 5 | 4 | 0 | 0 | Sarah asks how startups should navigate extreme capital intensity. Haas details how SoftBank acts as a strategic backstop and introduces SoftBank Neo Cloud as an alternative channel for emerging silicon companies. | |
| SoftBank Portfolio Synergies and Masayoshi Son's Vision | 4 | 4 | 0 | 0 | Elad and Sarah inquire about Haas's broader mandate across SoftBank Group International, and Haas discusses coordinating strategy across robotics, energy, and AI infrastructure with Masayoshi Son. | |
| The Evolution of Generalized and Reprogrammable Robotics | 4 | 4 | 1 | 0 | Sarah compares robotics development timelines to LLMs, and Haas contrasts purpose-built legacy robotics with modern reprogrammable generalized robotics targeting human labor tasks. | |
| Robotic Form Factors and Arm's Ubiquitous Footprint | 5 | 4 | 0 | 0 | Elad outlines differing hypotheses on humanoid versus specialized robotic form factors. Haas argues both will coexist and highlights that Arm microprocessors power the brains, sensing, and edge controllers. | |
| Domestic Semiconductor Sovereignty and Geopolitical Policy | 5 | 5 | 1 | 0 | Sarah brings up domestic manufacturing policies and export controls. Haas gives an impassioned defense of domestic fab sovereignty, comparing the data center and chip supply chain ecosystem to post-WWII Detroit auto manufacturing. | |
| Countering Data Center Backlash and Communicating AI Value | 6 | 4 | 1 | 0 | Elad and Sarah bring up coordinated local backlash, water misinformation, and union dynamics surrounding data center permits. Haas validates the hosts' points, stressing that technology leadership has no downside and requires clear communication. | |
| Arm's Core Strategic Positioning in the AI Era | 6 | 5 | 1 | 0 | Sarah breaks down the dual CPU demand drivers across rack infrastructure and software agents. Haas contextualizes this, emphasizing that while accelerators generate tokens, CPUs orchestrate all data movement and dominate low-power edge devices. |
Statements from this episode (23)
Haas: Arm now directly sources wafers and substrates for its own chip
“We also introduced our first product last March, the one you just mentioned at Hotchips, the Army GI CPU. So now we're in that soup ourselves from the standpoint of we're also having to figure out how to buy substrates and buy wafers and buy memory, et cetera.”
Haas: Meta Partnered with Arm to Build Its Agentic CPU
“So Meta was that first example. They wanted a general purpose agentic CPU. There wasn't anybody out who could give, give it to them. They came to us and said Hey, why don't we do this together? And that's how we got into it.”
Haas: Nvidia, Amazon, Microsoft, and Google Backed Arm's Physical Chip Strategy
“So whether it was Nvidia, Amazon, Microsoft, Google, All people who build ARM-based server chips, they were all on board, and I think the ultimate proof point was when we announced the product last March, we had Jensen, we had Ronnie Boker, we had Amin, we had…”
Haas: Arm has no intention to build semiconductor fabrication plants
“We don't have a fab and we have no intention to build a fab, but we fit in that ecosystem.”
Haas: Verification and validation take more time in chip design than RTL
“Chip design can take anywhere from 24 to 36 months, depending on the complexity of the chip, et cetera. The actual design of the architecture, the RTL generation, if you will, the mapping of the architecture, Is not the largest amount of time. The largest amou…”
Haas: 80% to 90% of Arm engineers use AI daily
“I would say we probably have 80 to 90% of engineers today in Sidearm who use it on a daily basis.”
Haas: Arm Likely Has Richest IP Portfolio and Documentation in Chip Industry
“We probably have the richest IP portfolio, both in terms of not only the IP, and this is the killer, the documentation, the test benches, you know, how, how you build the IP.”
Haas: Chip IP Without Documentation and Testability Is Untrainable for AI
“And if it's unusable and untestable, it's actually untrainable. And if it's untrainable, it's not usable for AI.”
Haas: AI Could Automate Simple Chip Design to GDSII in 5+ Years
“I don't know if it's in two to three years away, but five plus years, can you go from idea to a GDS two file being the file that you actually send to the fab to go get built. For certain designs, quite possible. So it takes that whole design piece out of the w…”
Haas: AI Won't Push-Button Design Frontier Chips Outperforming Leading Architectures
“Now, if you go into the tool and say, design me something that's 10% faster than Vera Rubin, 20% cheaper, and 30% more efficient on this model, You're not gonna get press a button and have it happen right away.”
Arm CEO Rene Haas: AI chip supply constraints will last 3–5+ years
“That's not going to stop in 12 months. It's not going to stop in 24 months. I think we're going to be in this constrained environment for Three to five years at least.”
Haas: AI chip startups need supply chain access, not just great design
“So long as the transformer is the unit of energy relative to how you generate AI training and AI inference by design, it is a, it is, it's very compute intensive, it's very memory intensive. So if you think about that, that's going to drive a lot of demand on …”
Haas: Physical data center construction will be next AI compute bottleneck
“I think building out the data centers is, is going to be a bottleneck.”
Haas: AI compute capacity is not even close to oversupplying demand
“The bubble in terms of, are we over oversupply to demand? Not even close. And I think again, that's because the demand is insatiable.”
Haas: SoftBank announced SoftBank Neo to become a neocloud provider
“We just announced, we being SoftBank SoftBank Neo which is our intent to become a Neo cloud.”
Haas helps direct SoftBank portfolio companies Ampere, Graphcore, and Stack AV
“So, yes, I, I'm leading the direction of Ampere and Graphcore and another company called Stack AV that's doing things around Autonomous, but maybe a better way to think about it, Elad, is that I'm kind of in the room on a lot of discussions that Masa's having …”
Rene Haas: Robots will easily replace human labor in construction, service, and security
“Where a lot of things will ultimately be done by robots construction, infrastructure, service, security. You know, right now you see a lot of stuff on Instagram or TikTok of Olympic races with robots, et cetera. I don't think anyone's gonna have any interest I…”
Haas: Most humanoid robot brains currently run on Arm architecture
“Today, whether it's NVIDIA or some of the work that Qualcomm does, most of the brains that you see in the humanoids, those are all, all running on ARM today.”
Haas: Distribution, delivery, and factories will be the first automated sectors
“Around factory automation and delivery and distribution, that will be one of the very first to be automated. No doubt.”
Haas: The US-China tech race is an infinite game with no final winner
“You know, my personal view is that it's an infinite game, I believe, first in terms of the race, that there's not going to be a winner, the race is going to be over, but you could get to a situation where a lot of the critical technologies are not U.S. Based, …”
Haas rejects claims that automated data centers fail to create jobs
“People look at data centers and say, oh, it's a big Costco box, and there's two parking, two cars in the parking lot, and all of that is being driven automatically, so there's no jobs. I call BS on that because if you think about whether it's around energy, Li…”
Haas: Demand for CPUs Will Remain Huge Alongside AI Accelerators
“So until something's invented that says the CPU has gone away and we're now doing it through some other mechanism, which has yet been defined or invented. The CPU is gonna be doing just fine and there's gonna be a lot of demand for it, a ton of demand, in addi…”
Haas: Edge AI Will Be a Sweet Spot for Arm Architecture
“And in fact, as you get to the smaller footprints, where more and more AI is going to take place, that's going to be a sweet spot for Arm, because the CPU's table stakes anyway, you have to have it to do all the things that are required in the edge device. But…”