The Story of Arm Holdings

A structural look at how a British chip designer became the architectural standard for mobile computing without ever manufacturing a single chip.

Introduction

Arm (ARM) Holdings designs processor architectures that power virtually every smartphone on earth. The company does not manufacture chips. It does not sell chips. It licenses instruction set architectures and processor designs to companies that do. This distinction—intellectual property licensor rather than manufacturer—defines Arm's entire structural position in the semiconductor value chain.

The model is unusual. Most semiconductor companies either design and sell their own chips or manufacture chips designed by others. Arm occupies a third position: designing the blueprint and collecting royalties each time a chip based on that blueprint ships. This creates a business with near-zero marginal cost per unit sold, extraordinary scalability, and deep entanglement with an ecosystem of hundreds of licensees.

Understanding Arm's arc reveals how an intellectual property licensing model—when attached to the right technology at the right structural moment—can create dominance that is self-reinforcing. The more companies build on Arm, the harder it becomes for alternatives to gain traction. The ecosystem itself becomes the moat.

The Long-Term Arc

How did Arm begin as a joint venture?

Arm began in 1990 as Advanced RISC Machines, a joint venture between Acorn Computers, Apple, and VLSI Technology. The original motivation was modest: design a low-power processor for Apple's Newton PDA. The project did not succeed commercially—the Newton famously failed—but the processor architecture that emerged had characteristics that would prove decisive decades later.

The RISC architecture Arm developed prioritized power efficiency over raw performance. This was a deliberate design choice driven by the constraints of battery-powered devices. In the early 1990s, the personal computer industry was obsessed with processing speed. Arm's emphasis on low power consumption seemed like a niche concern. That niche turned out to be the future of computing.

Critically, Arm adopted a licensing model from the beginning. The company lacked the capital to manufacture chips and lacked the scale to compete with Intel or AMD directly. Licensing was partly necessity, partly strategic insight. By making the architecture available to any chipmaker willing to pay, Arm began building an ecosystem rather than a product line.

Why was Arm's architecture built for the mobile moment?

The rise of mobile phones in the late 1990s and early 2000s created the structural moment Arm's architecture was built for. Mobile devices needed processors that could deliver adequate performance while consuming minimal power—exactly the tradeoff Arm had optimized. Intel's x86 architecture, dominant in PCs, was designed for plugged-in machines where power consumption was secondary to speed.

Nokia, Qualcomm, Texas Instruments, and Samsung all adopted Arm-based designs for their mobile processors. Each new licensee strengthened the ecosystem. Software developers wrote code for Arm. Tool vendors optimized compilers for Arm. Each additional participant made the architecture more attractive to the next potential licensee. The feedback loop accelerated.

By the time Apple launched the iPhone in 2007, Arm was already the default architecture for mobile computing. Apple designed its own Arm-based chips rather than licensing a complete Arm processor design—a more advanced form of engagement that generated different revenue for Arm but deepened Apple's commitment to the architecture. The smartphone revolution did not create Arm's dominance; it revealed and amplified a structural position that had been building for years.

What made Arm's ecosystem position so hard to switch away from?

By the mid-2010s, Arm's position had become deeply embedded. Over 200 billion chips based on Arm designs had shipped. The software ecosystem—operating systems, applications, development tools—was built around Arm's instruction set. Switching to an alternative architecture would require rebuilding this entire stack, a cost no individual company could justify when Arm licenses were relatively inexpensive.

SoftBank acquired Arm in 2016 for approximately $32 billion. The acquisition removed Arm from public markets and placed it within SoftBank's portfolio of technology investments. SoftBank's thesis was that Arm's architecture would expand beyond mobile into automotive, IoT, and data center markets—a bet on the proliferation of connected devices requiring efficient processing.

The SoftBank period was structurally significant in ways beyond ownership. Arm increased R&D spending, expanded into new markets, and developed architectures targeting higher-performance applications. The company was being repositioned from a mobile-centric IP licensor to a foundational computing architecture company.

What drove Arm's return to public markets in 2023?

Arm returned to public markets in 2023 through one of the largest technology IPOs in years. The re-listing reflected both SoftBank's need for liquidity and a market narrative around Arm's expanding addressable market. The company's reach now extends well beyond mobile phones into automotive systems, data center processors, and embedded computing.

Amazon's Graviton processors—custom Arm-based chips for AWS data centers—demonstrated that Arm architectures could compete with x86 in server workloads, a market Intel had dominated for decades. Apple's M-series chips showed Arm could deliver leading performance in laptops and desktops. These developments expanded Arm's relevance from a mobile-only architecture to a general-purpose computing standard, increasing the total addressable royalty base.

The structural dynamics of Arm's licensing model mean that each new market penetrated adds royalty streams without requiring proportional capital investment. Arm does not need to build factories or manage supply chains. It needs to ensure its architecture remains competitive and its ecosystem remains healthy. The capital efficiency is remarkable compared to virtually any other participant in the semiconductor value chain.

Structural Patterns

IP Licensing as Business Model — By licensing designs rather than manufacturing, Arm achieves near-zero marginal cost per unit and avoids the capital intensity that constrains chip manufacturers. Revenue scales with industry volume, not with Arm's own production capacity.

Ecosystem as Moat — Hundreds of licensees, billions of deployed devices, and a mature software ecosystem create switching costs that no individual competitor can overcome. The ecosystem reinforces itself—each new participant makes the architecture more valuable to all others.

Architectural Neutrality — Arm licenses to competitors. Qualcomm, Apple, Samsung, and MediaTek all use Arm architectures while competing fiercely in chip markets. Arm's neutrality—enabling all without favoring any—is structurally essential to maintaining ecosystem breadth.

Power Efficiency as Structural Advantage — The original design emphasis on low power consumption aligned with the most important trend in computing: the shift from plugged-in desktops to battery-powered mobile devices. A design choice made in 1990 determined competitive position decades later.

Royalty Model Compounding — Arm earns royalties on every chip shipped using its architecture. As the number of connected devices grows, royalty volume grows without proportional effort. The model compounds with industry expansion.

Market Expansion Without Capital Expansion — Entering automotive, IoT, and data center markets requires R&D investment in new architectures but no factories, no supply chains, and no manufacturing risk. The capital-light model extends naturally to new domains.

Key Turning Points

The decision to adopt a licensing model in 1990 was born partly from constraint—Arm lacked manufacturing capital—but it established the structural foundation for everything that followed. Had Arm manufactured its own chips, it would have become one competitor among many rather than the architectural standard that all competitors share. The licensing model transformed a potential weakness into the company's defining advantage.

Nokia's adoption of Arm for mobile phones in the late 1990s triggered the ecosystem flywheel. Once the leading mobile manufacturer committed to Arm, other handset makers followed. Software developers targeted Arm. The feedback loop between hardware adoption and software support created momentum that alternatives could not match. By the time smartphones arrived, Arm's position was structural rather than merely competitive.

Apple's decision to build custom Arm-based chips for Mac computers—announced in 2020—broke the assumption that Arm was limited to low-power mobile applications. When the world's most prominent computer maker chose Arm over Intel's x86 for its flagship laptops and desktops, it validated Arm as a general-purpose computing architecture. This single decision expanded Arm's addressable market from mobile devices to the entirety of computing.

Risks and Fragilities

RISC-V represents the most direct architectural threat. As an open-source instruction set architecture, RISC-V offers chip designers an alternative that requires no licensing fees. While RISC-V currently lacks the mature ecosystem that Arm has built over decades, ecosystem gaps narrow over time. If RISC-V achieves critical mass in any major market segment, it could erode Arm's position in that segment permanently. The threat is structural rather than imminent—but structural threats are the ones that matter most.

Revenue concentration in a small number of large licensees creates dependency. A handful of companies—Apple, Qualcomm, Samsung, MediaTek—generate a disproportionate share of Arm's royalty revenue. Any shift by a major licensee toward alternative architectures would have outsized impact. Arm's pricing power is also constrained by the need to keep licensing attractive enough that licensees do not invest in alternatives.

Geopolitical fragmentation threatens the universality that makes Arm's model work. If semiconductor supply chains bifurcate along geopolitical lines, Arm's ability to serve all markets equally could be constrained. Export controls, technology restrictions, and national semiconductor strategies could create barriers that fragment the ecosystem Arm depends on. A fragmented ecosystem is a weaker ecosystem.

What Investors Can Learn

Licensing models can create extraordinary capital efficiency — Companies that earn revenue from intellectual property rather than physical production can scale without proportional capital investment. Arm's model generates royalties on billions of chips without owning a single factory.

Ecosystems are stronger than products — Arm's moat is not any single technology but the ecosystem built around its architecture. Software compatibility, developer familiarity, and tool availability create switching costs that no product feature can replicate.

Design choices compound over decades — Arm's 1990 decision to optimize for power efficiency aligned with a computing trend that would not become dominant for fifteen years. Structural advantages often originate in decisions whose significance is not apparent when made.

Neutrality enables breadth — By licensing to all competitors equally, Arm maximized ecosystem participation. Companies that favor specific partners may optimize short-term relationships but narrow their structural reach.

Open-source alternatives are structural risks to licensing models — When the value proposition is an architecture, an open-source alternative—even an immature one—represents a fundamentally different kind of threat than a competing commercial product.

The Story of Arm Holdings

Introduction

The Long-Term Arc

How did Arm begin as a joint venture?

Why was Arm's architecture built for the mobile moment?

What made Arm's ecosystem position so hard to switch away from?

What drove Arm's return to public markets in 2023?

Structural Patterns

Key Turning Points

Risks and Fragilities

What Investors Can Learn

Inside CompanyGraph

Related

The Story of Ashtead Group

The Story of ASML

How to Find Beaten-Down Stocks With Strong Fundamentals