Arm Holdings: How a Cambridge Barn Built the World’s Chip Standard

Arm Holdings is the clearest example in British business of how owning a standard can be worth more than owning a factory. This case study explains how a spin-out from Acorn Computers became the invisible layer beneath the mobile era, why its royalty model is so hard to displace, what SoftBank’s ownership and the 2023 Nasdaq listing changed, and how Arm is trying to repeat the trick in artificial intelligence. For founders and operators, Arm is a masterclass in platform economics, patient capital and the strategic value of neutrality.

What exactly does Arm do?

Arm designs the instruction set architecture (ISA) and processor cores that other companies license to build their own chips. It does not own factories or sell finished silicon. Instead it sells the blueprints and the common language that lets software run across billions of devices from different manufacturers.

There are two products underneath everything. The first is the Arm architecture itself — the ISA that defines how software talks to the processor. The second is a portfolio of ready-made CPU core designs (the Cortex family, and the Neoverse line for servers) that a chipmaker can drop into a design to save years of engineering. A licensee either takes a finished core or, with an architecture licence, designs its own compatible core — which is exactly what Apple does with its A-series and M-series chips.

How did Arm begin in Cambridge?

Arm was founded in November 1990 as Advanced RISC Machines, a joint venture between Acorn Computers, Apple and chipmaker VLSI Technology. Its first office was a converted barn in Swaffham Bulbeck, near Cambridge. The team started with just a dozen engineers and a RISC design originally created for Acorn’s computers.

The pivotal early decision was Apple’s need for a low-power processor for its Newton handheld. Arm’s chips were not the fastest, but they sipped power — a weakness in the desktop era that became a decisive advantage the moment computing went mobile and battery life mattered more than raw clock speed. That single design constraint, efficiency over peak performance, shaped Arm’s destiny. It is a recurring theme across the UK Company Stories hub — British firms that won by being early to a shift others underestimated.

Why is the royalty model so powerful?

Arm charges an upfront licence fee to access a design, then a royalty on every chip a partner ships — typically a small percentage of the chip’s price. Because Arm-based chips ship in the tens of billions each year, tiny per-unit royalties compound into billions in revenue while Arm carries none of the manufacturing cost or inventory risk.

The model has three self-reinforcing strengths. Neutrality: because Arm does not make chips, rival manufacturers can all license from it without arming a competitor. Ecosystem lock-in: decades of software, tools and developer skill are built around the Arm architecture, so switching costs are enormous. And an annuity tail: chips designed today generate royalties for a decade or more as they ship in cars, appliances and industrial gear long after the design work is done.

Why did SoftBank buy Arm in 2016?

In July 2016, SoftBank’s Masayoshi Son acquired Arm for about £24bn (roughly $32bn) — at the time the largest-ever acquisition of a European technology company. Son’s thesis was that Arm would sit at the centre of the coming ‘Internet of Things’ and, later, artificial intelligence, as billions of connected devices all needed efficient processors.

The deal took Arm private and off the London Stock Exchange, a moment many in the UK still view as a national loss of a strategic asset. SoftBank kept Arm’s Cambridge base and pledged to double UK headcount, but the company’s centre of financial gravity shifted. A planned $40bn sale to Nvidia in 2020–2022 collapsed under regulatory pressure in the US, UK and EU, precisely because competitors feared losing Arm’s neutrality.

What changed with the 2023 Nasdaq listing?

In September 2023 SoftBank re-floated Arm on Nasdaq rather than London, valuing it at around $54bn on debut and raising close to $5bn. The choice of New York over the LSE reopened a painful British debate about where the country’s best technology companies choose to list.

Crucially, SoftBank sold only a minority slice and kept roughly 90% control, so Arm trades publicly but is still effectively steered by SoftBank. The listing gave Arm a currency for acquisitions and talent, market discipline, and a valuation that ties its fortunes explicitly to the AI narrative. In April 2026, TSMC fully exited the small stake it had taken at IPO — a reminder that Arm’s share register is unusually concentrated.

How big is Arm now, and where is growth coming from?

Arm reported full-year fiscal 2026 revenue of about $4.92bn, up 23% year on year, with fourth-quarter revenue of roughly $1.49bn (up 20%). Growth is being driven by higher royalty rates on its newer Armv9 architecture and by demand for its Compute Subsystems, which package more of a chip design together for customers.

The strategic pivot is into the data centre. Arm’s Neoverse cores already power custom server chips such as Amazon’s Graviton and Nvidia’s Grace, and the company has signalled a move into selling more complete AI CPU platforms itself — securing more than $2bn in customer commitments for fiscal 2027–2028 and targeting up to $15bn in annual AI CPU revenue by 2031. For a company built on smartphones, AI infrastructure is the next battleground, a theme explored across the British AI & Deep Tech stories in the UK Company Stories hub.

What can founders learn from Arm?

Arm’s central lesson is that owning the standard beats owning the factory. By staying capital-light and neutral, it extracted value from the entire mobile industry without ever competing with the customers who paid it. A constraint that looked like a weakness — low power instead of high performance — became the foundation of dominance when the market shifted.

The second lesson is about patience and platform compounding. Arm’s royalty annuities and ecosystem took decades to build and are now almost impossible to unwind. That is the kind of durable advantage worth reading alongside the other British technology case studies in the UK Company Stories hub, from enterprise software to cybersecurity.

How does Arm’s model compare with Intel’s?

Arm and Intel represent two opposite philosophies of the chip industry. Intel is an ‘integrated device manufacturer’: it designs its chips and builds them in its own multibillion-dollar fabs, capturing the whole margin but carrying enormous fixed costs and capital risk. Arm designs nothing physical and owns no fabs, capturing a thin slice of value from a vast number of chips built by others.

For decades Intel’s integrated model dominated PCs and servers, where performance mattered most. But the mobile revolution rewarded Arm’s efficiency-first, capital-light approach, and its ecosystem of competing licensees out-innovated a single vertically integrated giant. Intel’s later struggles to enter mobile — and Arm’s push into laptops and servers — show how a business model, not just a product, can decide who wins a platform shift. It is a contrast worth holding in mind when reading the hardware and engineering stories in the UK Company Stories hub, where British firms have repeatedly had to choose between owning production and owning design.

Why did the Nvidia acquisition collapse?

Between 2020 and 2022, SoftBank agreed to sell Arm to Nvidia in a deal that ballooned to about $40bn. It never completed. Regulators in the US, UK and European Union, together with a chorus of Arm licensees, opposed it on competition grounds.

The objection went to the heart of Arm’s value: its neutrality. If Nvidia — itself a major chip designer and Arm customer — owned Arm, rivals feared they would lose access to a supposedly impartial platform, or that Nvidia would steer the roadmap to its own advantage. The collapse validated the strategic logic of the whole business. Arm is worth most precisely because it takes no side, a subtle point founders building marketplaces and platforms should absorb: the moment a neutral layer picks a winner, it starts to lose the trust that made it valuable.

What does Arm mean for the UK economy?

Arm is the anchor of the ‘Silicon Fen’ technology cluster around Cambridge and one of the few British firms that sets a global technology standard rather than merely adopting one. It employs thousands of high-skill engineers in the UK and trains talent that seeds the wider ecosystem, including many of the AI and deep-tech companies profiled elsewhere in the UK Company Stories hub.

Its ownership and listing history also crystallise a national policy anxiety: Britain is world-class at inventing foundational technology but has repeatedly seen it acquired or listed abroad. Whether the UK can build and retain the next Arm — rather than sell it early — is one of the defining questions for British industrial strategy, and a backdrop to almost every story in this collection.

How does Arm support software developers?

Arm invests heavily in the tools, compilers and documentation that let millions of developers write code once and have it run across the vast Arm ecosystem. This developer support is a moat in itself: the more software that is optimised for Arm, the more attractive the architecture becomes to the next chipmaker.

That flywheel — more chips attract more software, which attracts more chips — is why challengers cannot simply match Arm on engineering. They must also recreate decades of accumulated software and skills, a barrier explored across the technology stories in the UK Company Stories hub.

Frequently Asked Questions