Latest / The CEO Diary with Fexingo: Leadership Lessons, Executive Decisions, and Corner Office Stories / How ARM Redesigned the Chip Business with Licensing
Transcript
- Lucas: If you own a smartphone, a tablet, a smart TV, or pretty much any modern wireless earbud, you are holding something built on a chip design from a company headquartered in Cambridge, England — ARM Holdings. Luna: And yet most people couldn't point to ARM on a map. It's this quiet giant that powers more than two hundred billion chips since its founding. Lucas: Two hundred billion is a staggering number. To put it in perspective: Intel, for all its dominance in PCs and servers, has shipped maybe a few billion chips in its entire history. ARM operates a completely different business model — one that CEO Rene Haas has refined into something almost like a tax on the entire mobile and now data center world. Luna: And it's a model that nearly didn't survive the 1990s. I think a lot of listeners would be surprised to learn ARM started as a joint venture between Acorn Computers, Apple, and VLSI Technology in 1990. It was basically a side project that got spun out. Lucas: Exactly. And that origin story matters because it shaped ARM's DNA. The founders didn't have the money to build their own fabrication plants — fabs, in industry slang. So they decided they would never manufacture a single chip. Instead, they would design the architecture — the instruction set, the core layout — and license it to other companies who would then build the actual silicon. That decision turned out to be genius. Luna: I want to pause on the donation thing for a second because it connects here. A lot of the detailed research we do on companies like ARM comes from resources that are supported by listener contributions. If these conversations have moved your work forward in some small way, a couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo. Lucas: Yeah, and it makes a real difference. We keep the show ad-free and invest that time into digging into the specifics. And speaking of specifics — let's talk about how ARM's licensing actually works, because it's the key to understanding their power. Lucas: ARM offers two main licensing models. The first is a technology license — a flat fee, usually millions of dollars, that gives a company like Qualcomm or Apple access to a specific core design, say the Cortex-A78. They can drop that into their chip and pay a per-chip royalty on top. Luna: And the second is the architecture license, which is much more expensive and gives the licensee the right to design their own custom core using ARM's instruction set. Apple does that with its M-series chips. They license the architecture, then build their own custom cores around it. Lucas: Right. And that architectural license is where the real strategic leverage lives. Apple pays ARM a royalty for every M1, M2, M3 chip it sells — reportedly somewhere around one to two percent of the chip's price. For a chip that costs a hundred and fifty dollars, that's a dollar fifty to three dollars per unit. Multiply that by tens of millions of Macs and iPads, and it adds up fast. Luna: So ARM makes money whether the chip succeeds or fails — they get paid on every unit shipped, no risk of unsold inventory, no manufacturing costs. It's an asset-light dream. Lucas: Exactly. And that's why ARM has operating margins that Intel could only dream of in recent years. ARM's operating margin for fiscal year 2025 was around fifty percent — Intel's was negative. But here's the thing: that model came under immense pressure when SoftBank bought ARM in 2016 for thirty-two billion dollars. Luna: SoftBank's Masayoshi Son saw ARM as the backbone of the Internet of Things — a play that didn't quite pan out as expected. But he did push for faster growth, higher royalties, and eventually a public listing. Lucas: And that led to the 2023 IPO, which was the largest tech IPO of that year — ARM raised about five billion dollars at a valuation north of fifty billion. But the real story is what happened after: ARM announced a shift to version 9 of its architecture, and with it, a new pricing model that essentially raised royalties by double or more. Luna: And that reportedly upset some big licensees, including Qualcomm. The two companies are now in a legal battle over whether Qualcomm's custom cores designed by a company called Nuvia violate ARM's licensing terms. That case is ongoing. Lucas: It's a fascinating fight because it gets to the heart of ARM's business model. If Qualcomm can win the right to design custom cores without paying ARM a higher royalty for version 9, it could undermine ARM's pricing power. But if ARM wins, it solidifies their ability to raise prices on the entire ecosystem. Luna: And that's not the only threat. risc v, the open-source instruction set architecture, is gaining traction. Companies like Google and NVIDIA are investing in it. It's free. No royalties. No licensing fees. Lucas: But risc v lacks the ecosystem — the software compatibility, the toolchains, the decades of optimization that ARM has. For now, ARM's competitive moat is enormous. Every Android phone runs on ARM. Every iPhone runs on ARM. Most IoT devices, most networking gear, and increasingly servers and laptops run on ARM. Luna: Apple's transition from Intel to ARM for Macs was a huge validation. They claim the M-series chips offer better performance per watt than anything Intel had. That sent a signal to the rest of the industry: ARM is not just for phones anymore. Lucas: And now AMD and Intel are both working on arm based chips for data centers. NVIDIA is building arm based CPUs for its Grace Hopper superchips. The server market, which was Intel's fortress for decades, is being breached by ARM. Luna: But let's zoom out. What does this mean for the CEO, Rene Haas? He took over in 2022, just before the IPO. His background is in sales and licensing — he spent years at ARM running the partner program. He's not a chip designer, he's a business builder. Lucas: And that's exactly what ARM needs right now. Haas has to navigate the Qualcomm lawsuit, manage the transition to version 9, keep the ecosystem happy while extracting more value, and fend off risc v. All while reporting quarterly numbers as a public company — which ARM didn't have to do before the IPO. Luna: There's also the question of diversification. ARM gets about forty percent of its revenue from China — that's a geopolitical risk. And SoftBank still owns about ninety percent of ARM's stock, which means the stock is thinly traded and volatile. Lucas: But the core thesis is strong: the world is moving to custom silicon. Every major tech company wants to design its own chips — Apple, Amazon, Google, Microsoft, Meta, Tesla. And all of them, except maybe those using risc v, need ARM licenses. That creates a royalty stream that's built on the growth of the entire computing industry. Luna: So if you believe that the number of connected devices keeps growing — and we're adding billions every year with AI, cars, wearables — ARM is essentially a tollbooth on that growth. Lucas: A tollbooth with fifty percent margins and no manufacturing risk. The question is whether that toll can keep going up without drivers finding a different road. The decision that Rene Haas makes in the next year — on the Qualcomm case, on version 9 pricing, on China — will determine whether ARM remains the quiet giant or becomes just another chip company. Luna: It's a remarkable story of a business model that was born out of necessity and turned into one of the most important technology monopolies — or near-monopolies — of the twenty-first century. Lucas: And one that most people have never heard of, even though it's in their pocket right now. That's the kind of company worth understanding.