Arm

Arm licenses CPU architectures and IP used in nearly every mobile device and increasingly in data-center and AI chips. IPO in Sept 2023 valued at ~$50B; royalty model scales with AI chip proliferation. Nvidia acquisition blocked by regulators.

Arm's architecture underpins custom AI chips from Amazon (Graviton), Google, and Microsoft, challenging x86 in the data center.

arm.com EDA & IPUKpublic · ARM

At a glance

Total raised

$4.9B

last round ipo

Headcount

—

no public estimate on record

Headquarters

Cambridge

Founded

1990

36 years active

Status

Public

Ticker · ARM

Latest valuation

$54.5B

IPO · 2023

For subscribers

Subscribe to see the full profile

Including Funding rounds, Bull / Bear thesis, Stock + earnings, Roster changes, Patents, News, and Open roles.

Founding

50% off

/month

91 of 100 spots left

Full

$10

/month

Available once all 100 Founding Member spots are claimed.

Get full access

Already subscribed? Sign in →

Funding history

Total raised

$4.9B

Rounds tracked

Latest stage

IPO

Latest valuation

$54.5B

Date	Stage	Amount	Valuation	Lead investors
Sep 14, 2023	IPO	$4.9B	$54.5B	—

Stock performance

How they make money

Content licensingPattern

They run an intellectual-property licensing model: chipmakers pay Arm an upfront license fee to access its processor architectures and then a per-chip royalty on every unit shipped that uses Arm designs. Arm does not manufacture or sell chips itself — it licenses blueprints. Customers include nearly every major semiconductor company: Qualcomm, Apple, Samsung, MediaTek, and increasingly data-center players such as Amazon, Google, and Microsoft building custom server chips (Graviton, Axion, etc.). Revenue is recognized from the license fee at the start of the agreement and from royalties as chips are manufactured and sold, typically reported one to two quarters in arrears. The model's economics are structurally attractive because the marginal cost of licensing an additional chip design is near zero — once the IP is created, each additional royalty dollar flows at very high incremental margin. R&D spend to develop new architectures (Armv9, CSS, Neoverse) is the fixed-cost base that supports the entire revenue stream. The royalty stream is effectively a tax on global ARM-based chip production, giving Arm upside from any end market that adopts its architecture — mobile, IoT, automotive, or data-center AI silicon — without taking on manufacturing risk.

Generated by deepseek-v4-flash · June 2, 2026Analytical summary · not investment advice

AI bull / bear

Bull case

Arm is now at the center of a three-front war against x86 dominance. Nvidia's RTX Spark superchip — a 20-core Grace Arm CPU mated to a Blackwell GPU — landed flagship design wins from Microsoft's Surface Laptop Ultra, with 128GB unified memory and 6,144 CUDA cores purpose-built for agentic AI workloads. This is Nvidia explicitly choosing Arm over x86 for the next generation of Windows PCs, and it arrives just as Qualcomm's Windows-on-Arm exclusivity expired, opening the floodgates for multiple chip vendors. In the data center, Amazon's Graviton, Google's Axion, and Microsoft's Cobalt are already displacing Xeons at hyperscale. Arm's own homegrown, AI-tuned server CPU signals it's not just an IP licensor but a platform-defining force. The royalty model means every one of these chips — from Nvidia's PC SoCs to custom server silicon — pays Arm per unit. AI inference moving to the edge and PC creates a new royalty base measured in hundreds of millions of devices annually, layered atop the mobile royalty stream that already makes Arm ubiquitous.

Bear case

Arm's competitive moat is narrower than it appears. Nvidia's RTX Spark is an Nvidia-controlled platform — Arm supplies the CPU cores, but Nvidia captures the system value, the GPU margin, and the software stack. The same dynamic plays out with every hyperscaler custom chip: Arm trades architectural lock-in for a royalty per chip that is tiny compared to what Intel or AMD would have charged for a full SoC. If Arm successfully pushes licensing fees higher to capture more value from these wins, it risks pushing Nvidia, Apple, Amazon, and Google further down the road of designing custom CPU cores under their existing Architecture License Agreements — or even migrating to RISC-V, which is free and gaining toolchain maturity. Qualcomm's messy legal history with Arm over licensing terms previews this tension. Meanwhile, Nvidia's RTX Spark threatens to marginalize Arm's own platform ambitions: if the PC market consolidates around Nvidia's superchip, Arm becomes a second-tier supplier of CPU cores inside a system whose architecture is dictated by another company's GPU roadmap.

Generated by deepseek-v4-flash · June 2, 2026Refreshes quarterly · see disclosures

In the news

Watch

Leadership

Rene Haas

CEO

Jason Child

EVP & CFO

—

Richard Grisenthwaite

EVP & Chief Architect

—

Will Abbey

EVP & Chief Commercial Officer

—

Products

Arm Cortex CPU Cores
Licensable processor IP spanning the application-class Cortex-A (smartphones, servers, infotainment), real-time Cortex-R, and microcontroller-class Cortex-M lines. Cortex-A designs power nearly every smartphone and increasingly data-center and AI-host CPUs, while Cortex-M dominates embedded and IoT. Customers integrate these cores into their own SoCs and pay Arm license and per-chip royalty fees, the foundation of Arm's business.
Arm Compute Subsystems (CSS) & Neoverse
Neoverse is Arm's server- and infrastructure-class CPU IP behind custom data-center chips such as AWS Graviton, Microsoft Cobalt, Google Axion, and Nvidia Grace. Compute Subsystems (CSS) deliver pre-integrated, validated multi-core configurations that shorten time-to-silicon for partners. Together they anchor Arm's push into hyperscale data centers and AI infrastructure, moving Arm up the value chain from cores to subsystems.
Arm Mali / Immortalis GPUs
Graphics and compute GPU IP for mobile and embedded SoCs, with the flagship Immortalis line adding hardware ray tracing for premium smartphones. Mali GPUs handle gaming, UI rendering, and on-device ML acceleration across a vast installed base of devices. As licensable IP, they complement Arm's CPU cores to give partners a full mobile compute platform.
Armv9 Architecture
The instruction-set architecture that licensees build chips around, with Armv9 adding Scalable Vector Extension 2 (SVE2) for AI/ML and DSP workloads plus enhanced security features like the Confidential Compute Architecture. The architecture license is Arm's highest tier, letting partners design custom cores. Armv9 underpins the company's roadmap across mobile, automotive, and data-center silicon.

Most recent patents

459 patents on file, but none with both an extractable figure and an abstract on Google Patents yet.

Arm

At a glance

Subscribe to see the full profile

Funding history

Stock performance

How they make money

AI bull / bear

In the news

Nvidia's RTX Spark could caplitalize where Qualcomm's Arm-based efforts have not — following the expiration of Qualcomm's Windows on Arm deal, Nvidia stands poised to pick up the slack

NVIDA Introduces RTX Spark: An Arm SoC for Windows PCs

We went hands-on with Qualcomm's new '$300 and up' ARM laptop platform with mystery eight-core CPU — active-cooled Snapdragon C laptop surfaces in Acer Aspire Go 15

Microsoft Surface Laptop Ultra weilds Nvidia's RTX Spark superchip with 128GB of RAM, 20 Arm CPU cores, and a Blackwell GPU — 15-inch mini-LED PixelSense Ultra display rounds out the powerful package

Nvidia unveils RTX Spark Superchip for laptops and desktop PCs at Computex 2026 – new platform promises to turn Windows into an agentic AI OS with Arm CPU, Blackwell GPU, and 128GB unified memory

Watch

Leadership

Products

Arm Cortex CPU Cores

Arm Compute Subsystems (CSS) & Neoverse

Arm Mali / Immortalis GPUs

Armv9 Architecture

Most recent patents

Compare Arm with