Skip to research
stock3dby Fynca
Explore in 3D

COMPANY RESEARCH / ARM

Arm Holdings

Processor architecture & robotics IP

Published by FyncaResearch checked How we research

HUMANOID ROBOT

Where it fits in the robot

Arm’s processor architectures span real-time motor control through AI planning, enabling chips used in robotics and physical AI.

What the evidence establishes

Humanoid evidence

Arm explicitly maps licensed processor architecture to humanoids. IP licensing is distinct from manufacturing the robot’s processor chips.

An explicit humanoid product, development program, application, or partnership. This does not establish revenue, market share, or production volumes.

Component roleProcessor architecture & robotics IP
EvidenceArm explicitly maps licensed processor architecture to humanoids. IP licensing is distinct from manufacturing the robot’s processor chips.
Primary sourceArm Holdings — supporting disclosure
Company or official disclosure; vendor claims are attributed to their source.
Checked. This is the research review date, not the source publication or quote time.

Keep in perspective

Primarily an intellectual-property and compute-platform business; it is not a complete robot or actuator manufacturer.

Platform business exposure. This describes the breadth or role of the business; it is not an investment rating. A robotics product or development announcement does not by itself establish material humanoid revenue, market share, or future returns.

Connections to the assembly

These links explain the technology relationship. The model is illustrative and does not represent an actual manufacturer’s bill of materials.

AI & control processor

The main processor executes perception and planning. Real-time control can run on separate processing cores or devices.

Scope of the company links: AI processors and embedded real-time controllers serve different workloads. Arm licenses processor IP; it does not manufacture chips.

SMART GLASSES

Where it fits in smart glasses

Licenses the CPU and NPU IP inside glasses chips. Arm says Meta’s Ray-Ban glasses are Arm-based and that the Neural Band is built on Arm CPUs and Ethos-U55 NPUs (Arm newsroom, Dec 2025). Cortex-M cores are documented in Bestechnic BES2800BP, NXP i.MX RT600 (Rokid) and the main MCUs of Brilliant Labs Halo and Frame. Arm makes no chips.

What the evidence establishes

Smart-glasses evidence

Disclosed: Arm’s newsroom says Meta’s Ray-Ban glasses are Arm-based and the Neural Band runs on Arm CPUs and Ethos-U55 NPUs; Cortex-M cores are documented in several glasses chips.

An explicit smart-glasses product, design-in, supply role, program, application page or partnership, including wrist or ring input sold with glasses. For a device maker, it means the company’s own glasses, not a supplier relationship. This does not establish revenue, market share, volumes or future sourcing.

Component roleCPU and NPU IP inside glasses chips
EvidenceDisclosed: Arm’s newsroom says Meta’s Ray-Ban glasses are Arm-based and the Neural Band runs on Arm CPUs and Ethos-U55 NPUs; Cortex-M cores are documented in several glasses chips.
How we knowDisclosed The company’s own filing, release, annual report, product page or investor record describes the smart-glasses role. Where customers are unnamed, the listing says so.
Primary sourceArm Holdings — supporting disclosure
Company or official disclosure; vendor claims are attributed to their source.
Checked. This is the research review date, not the source publication or quote time.

Keep in perspective

Arm licenses IP and makes no chips. Its claim about the Neural Band’s NPU is not reconciled with the teardown’s chip identification.

Platform business exposure. This describes the breadth or role of the business; it is not an investment rating. A smart-glasses product, design-in or partnership does not by itself establish material glasses revenue, market share, or future returns.

Connections to the assembly

These links explain the technology relationship. The model is illustrative and does not represent an actual manufacturer’s bill of materials.

AR SoC

An AR-class SoC such as Snapdragon AR1 Gen 1 runs the camera ISP, on-device AI, audio and display timing. With no fan and skin contact, sustained power stays around 1-2 W.

Scope of the company links: Snapdragon AR1 Gen 1 is teardown-verified in Meta Ray-Ban Display (TechInsights) and Oakley Meta HSTN (iFixit) and named in Samsung’s and Rokid’s releases. Bestechnic and Actions sell dedicated glasses SoCs and report glasses mass production in their 2025-26 filings. Arm licenses the CPU/NPU IP (Arm calls Meta’s Ray-Ban glasses Arm-based); it does not make chips.

AI SERVER RACK

Where it fits in an AI server rack

The 36 Grace CPUs in a GB300 NVL72 carry 2,592 Arm Neoverse V2 cores (NVIDIA). On Mar 24, 2026 Arm launched its first Arm-designed data-center chip, the AGI CPU (up to 136 Neoverse V3 cores, TSMC 3 nm), with Meta as lead partner and OpenAI, Cloudflare, SAP and SK Telecom among first customers; broader availability is expected in H2 2026.

What the evidence establishes

AI-rack evidence

Partner-disclosed: NVIDIA’s GB300 NVL72 page counts 2,592 Arm Neoverse V2 cores, and Arm’s newsroom announces its own AGI CPU silicon.

An explicit AI-server or AI-rack product, design-in, supply role, program, reference design or partnership, such as a part named for rack-scale AI systems or a product sold for AI data centers. For a rack builder, it means the company’s own AI servers or racks. This does not establish revenue, market share, volumes or future sourcing.

Component roleNeoverse CPU cores in Grace; Arm AGI CPU silicon
EvidencePartner-disclosed: NVIDIA’s GB300 NVL72 page counts 2,592 Arm Neoverse V2 cores, and Arm’s newsroom announces its own AGI CPU silicon.
How we knowPartner-disclosed A named partner, customer or platform owner’s document names the company. The scope is that document’s product or program, not a general supply relationship.
Primary sourceArm Holdings — supporting disclosure
Company or official disclosure; vendor claims are attributed to their source.
Checked. This is the research review date, not the source publication or quote time.

Keep in perspective

For Grace, Arm licenses the cores and earns royalties, not chip revenue; terms are undisclosed. The AGI CPU is a general data-center CPU launched in 2026, and its use in NVL72-class racks is not established.

Platform business exposure. This describes the breadth or role of the business; it is not an investment rating. An AI-server product, design-in or partnership does not by itself establish material AI-rack revenue, market share, or future returns.

Connections to the assembly

These links explain the technology relationship. The model is illustrative and does not represent an actual manufacturer’s bill of materials.

Host CPU

An Arm-based or x86 processor that runs the operating system, stages data and coordinates the GPUs. In rack-scale designs it shares a coherent high-bandwidth link with its GPUs.

Scope of the company links: NVIDIA’s Grace is the host CPU of GB200/GB300 NVL72: a GB200 rack has 36 Grace CPUs with 2,592 Arm Neoverse V2 cores (NVIDIA spec page). Arm licenses the Neoverse cores; in March 2026 it also launched its own AGI CPU silicon. x86 alternatives are AMD EPYC (18 Venice CPUs in AMD’s Helios rack) and Intel Xeon hosts in 8-GPU HGX-class servers. Montage’s Jintide x86 CPUs have no documented AI-rack role and are not mapped.

Related research

Companies with overlapping sector exposure. Their products and evidence may differ; this is not a list of equivalent investments.

Suggest a correction · Research and market-data disclosures