All-In Podcast Ep. - Arm CEO Rene Haas on AI: Nvidia Lessons, Intel’s Decline and the US-China Chip War

Date: September 30, 2025
Guest: Rene Haas (CEO, Arm)
Hosts: Chamath Palihapitiya, Jason Calacanis, David Sacks & David Friedberg

Episode Overview

In this episode, the All-In hosts sit down with Rene Haas, CEO of Arm, following their blockbuster IPO. The conversation explores Arm's unique industry positioning, the rapid evolution of AI hardware, lessons from Nvidia, Intel’s struggles, China-US dynamics in semiconductors, and the challenges and prospects for onshoring chip manufacturing. Throughout, Haas offers insights on how Arm fits at the heart of global computing and the complex geopolitical and technological forces shaping the industry.

Key Discussion Points & Insights

1. Arm’s Ubiquity and IPO Success

• Arm’s Role: Nearly every chipmaker uses Arm technology, even though Arm doesn’t manufacture physical products.
• IPO Highlight: The blockbusting September IPO valued Arm above $54 billion, the largest in over two years.
  • "It's now $150 billion market cap company." (04:00)
• Masayoshi Son and SoftBank’s Stake: Arm went from private (acquired by SoftBank at $32B) to a public company, and Son refuses to sell, slowly building Arm’s public shareholder base.

2. Nvidia, AI Workloads & Hardware Innovation

• Nvidia as Customer and Former Employer: Haas speaks highly of Nvidia and CEO Jensen Huang, highlighting his vision and bold pivots (01:12–02:48).
  • “Nvidia is a customer of ours... I worked for Nvidia for many, many years … learned so much working there, working for him … vision, speed, fearlessness, taking risk, and ability to pivot very, very fast.” (01:12–01:44)
• Strategic Shifts at Nvidia: Nvidia’s bold pivot from mobile chipsets to SoC (System-on-Chip) and early embrace of AI workloads via GPUs positioned them as leaders when AI took off.
  • “At that time Jensen looked at what was going on with SOCs and ARM based architecture and moved everybody onto the SOC program.” (02:48–03:20)
• AI and the Role of Hardware:
  • “What really drives demand is compute workloads… When a new workload is essentially either identified and or invented, then it comes down to what is the best architecture, processor wise to address that workload.” (04:14)
• Nvidia’s Grace Blackwell Chip: Uses 72 Arm CPUs—Arm remains a lynchpin even at the highest end of AI training chips.

3. Arm as "Arms Dealer" to the AI World

• Flexible Business Model: Arm enables custom and standard chip solutions for many AI and data center players—including Nvidia, Google (TPUs), Cerebras, Grok, Tesla, etc.
  • “We are now increasingly that microprocessor that connects to these accelerators... Could we do something ourselves, Custom? It's possible. Could we also supply the intellectual property... We're doing that today.” (06:46)
• Future Intentions: Hints at moving “a bit further” into the market, but doesn’t confirm plans to build and sell chips directly.
  • “I hinted in the last conference call that we're looking at going a little bit further than we do today.” (07:41)

4. AI Hardware Market Divergence: Training vs. Inference

• Potential Bifurcation: Training and inference chips may split, with most companies developing custom inference chips—training remains Nvidia’s stronghold, but innovation is happening.
  • “Could we see in the next few years a divergence in the market between training and inference?... Yes.” (07:49)
• Third Model—Hybrid Chips: The future could have “professor-student” models (big models training/teaching smaller ones) at the silicon level.
  • “You have a third bucket where training distills down to simpler training chips… It's almost like the professor teaching a student who can also be a student teacher who can do a little bit of both.” (08:27)

5. Physical AI & Robotics as the Next Frontier

• Gigantic Market: Physical AI in robotics could surpass data center AI due to sheer unit volume.
  • "Physical AI is going to be a gigantic market... The robots themselves will have tens of chips, hundreds of chips... the numbers are going to be well beyond what we see today." (09:23)
• Current Practice: Many robotics systems run on recycled automotive chips; Haas sees a wave of chips designed specifically for physical AI.

6. US-China Chip War & Export Controls

• Arm’s Role: As early-stage IP and design providers, Arm gets a unique window into global innovation but doesn’t manufacture or sell directly into geopolitically sensitive supply chains.
  • “We are early in the value chain... we probably see what people are doing earlier than anybody else…” (10:24)
• On Export Controls:
  • “The China ecosystem on software looks a lot like the west, which for us is obviously great... it's great if the global ecosystem remains open." (10:24–11:31)
• Regulation Risks:
  • “If you shut off supply ... what will happen? ... they will find a way around the problem. And once that happens, you've now created two parallel universes.” (20:21)

7. Intel’s Failure to Adapt

• Falling Behind: Intel missed the mobile revolution and failed to invest in leading-edge EUV (Extreme Ultraviolet) lithography. Now fab (foundry) leadership belongs to TSMC.
  • “If you miss a few [product cycles], time will punish you for that … Once you fall behind in chips, it’s very… difficult to catch up because the cycle gets on top of you.” (11:44)
• Ecosystem Flywheel: Once leaders like TSMC attract the key companies, their lead becomes self-reinforcing.

8. Industrial Policy: Lessons from China and the Role of Government

• Supply Chain Security: Governments should invest not just in chipmakers (like Intel), but critical upstream players (ASML, Carl Zeiss, rare earth refiners).
  • “The issue is in the refinement and actually building the factories that can refine the materials. Again, that's a decades level of investment.” (13:36)
• China’s Long-View:
  • "One of the things that I was very impressed with … is the industrial policy that sits inside the central government that will last respectfully an election cycle." (13:36)
• US Approach: Broader industrial cooperation between universities, corporations, and private equity is needed to develop critical infrastructure.
  • “You need universities, but you need corporations to get behind this as well as financing private equity... Because this is a huge capital investment.” (15:02)

9. Challenges in Onshoring Chip Manufacturing (U.S.)

• TSMC in Arizona: Building world-class fabs in the U.S. is not just about money—it's about workforce mindset, training, and culture.
  • “TSMC is a 24/7 operation … we’ve lost the muscle memory … on how to do that.” (16:15–17:54)
• Solution: Invest in university programs for chip design and manufacturing operations excellence.
  • “I was at Carnegie Mellon… They now have microelectronics classes for chip design. That was gone a number of years ago.” (18:09)

10. Global Ecosystem, Talent & Multicultural Dynamics

• Arm Origins: From a barn in Cambridge as a joint Apple—VLSI venture for the Newton, to global expansion.
• Leadership Style: Haas (first non-UK CEO) aims to blend UK scientific depth with Silicon Valley speed.
  • "Keep the great scientists ... but inject a bit of Silicon Valley aggressiveness and twist to moving faster." (21:56)
• Talent Needs: The demand for engineers is higher than ever; AI can’t replace chip designers yet.
  • “We need far more investment... For engineers, AI for development, AI for creation, AI for science, that's still a hard problem to solve.” (23:06)

11. US-China Relationship & The Future of AI

• Collaboration vs. Competition: Haas is cautiously optimistic about cooperation.
  • "I'm going to be an optimist here, Jason, and say, I think, yes. I think that China views some of the things around AI ... their minds are in the right space." (24:14)
• Endgame: Not a nuclear arms race, but a need for dialogue between technical powers for mutual benefit.

Notable Quotes & Memorable Moments

• On AI hardware’s evolution:
  • “Every one of these workloads requires a CPU to not only run the computer, but help the accelerator run. And that's where Nvidia is a customer today.” (04:41 – Rene Haas)
• Describing Nvidia’s pivotal moment:
  • “...a supposed roadmap review turned into, we're changing the strategy. We're abolishing this product line. We're going to move 2,000 engineers off Project X onto Project Y.” (02:13 – Rene Haas)
• On Intel’s missed opportunities:
  • “If you miss a few [product cycles], time is very, very… you will be punished for that. And I think Intel has unfortunately been punished on a few areas.” (11:44 – Rene Haas)
• On the risk of overregulating chips:
  • “The world works really well when it's flat and there isn't constraints relative to who you sell to or how ecosystems get built…” (20:21 – Rene Haas)
• On the U.S. chip manufacturing challenge:
  • “We've lost the muscle memory inside the United States … It is a mindset. TSMC is a 24/7 operation…” (16:15 – Rene Haas)

Key Timestamps by Topic

• 00:52 | Welcome, opening joking banter on nicotine pouches and Jensen Huang
• 01:12–02:48 | Lessons from working for Nvidia & strategic pivots under Jensen Huang
• 03:27–04:14 | SoftBank, Arm's IPO & the company’s unique role in the AI ecosystem
• 04:14–06:13 | GPU, CPU, and AI workload fundamentals; Arm’s role in Nvidia’s highest end chips
• 06:13–07:38 | Arm’s business flexibility—“arms dealer” to the AI revolution
• 07:49–09:09 | The potential bifurcation between training and inference chips
• 09:23–10:02 | Robotics and the rise of “physical AI” as a major new chip market
• 10:02–11:31 | Arm’s vantage point, export controls, and China’s technology ecosystem
• 11:44–13:03 | Intel’s decline: missed mobile and EUV, TSMC’s rise
• 13:36–15:36 | Government industrial policy: rare earths, ASML, US strategic investments
• 15:36–18:09 | U.S. fabs: workforce, culture, and regaining “muscle memory”
• 18:35–21:36 | Export controls, the risk of parallel tech universes
• 21:56–23:00 | Building a multicultural, global company and engineering talent crisis
• 23:43–24:52 | Geopolitics—US, China, and the hopes for cooperation

Conclusion

Rene Haas offers a clear and candid look into the heart of the semiconductor industry, its leadership drama, and the global chessboard of chip politics. He stands as both realist and optimist, convinced that innovation flourishes when ecosystems remain open and global—even as the world teeters on the edge of fragmentation. The episode is essential for anyone wanting to understand the real forces shaping the next decade of computing.