Arm to Manufacture Server CPUs, With Meta as First Major Customer

Arm, traditionally known for licensing its chip blueprints to companies like Apple, Nvidia, and Qualcomm, is making a radical shift by producing its own server CPUs in-house. The UK-based company, owned by SoftBank, plans to unveil its first self-made chip as early as this summer, with Meta as one of its first major customers.

A Major Shift in Arm’s Business Model

• From licensing to manufacturing – Arm has historically designed chip architectures but has never produced its own full processors.

• Targeting data centers – The upcoming chip will be a server CPU designed for large-scale AI and cloud computing workloads.

• Meta as an early adopter – The social media giant is moving away from Intel and AMD in favor of Arm-based server chips to power its AI workloads.

• Production outsourced – While Arm will design the chip, manufacturing will be handled by a third-party such as TSMC.

Meta’s finance chief, Susan Li, stated that the company is expanding its custom silicon efforts to AI training workloads, aiming to enhance efficiency and performance by optimizing chips for its specific computing requirements.

SoftBank’s Larger AI Strategy

The move aligns with SoftBank founder Masayoshi Son’s ambitious Stargate initiative, which aims to build a $500 billion AI infrastructure in partnership with OpenAI, Oracle, Abu Dhabi’s MGX, Microsoft, and Nvidia. Arm is a key technology provider for this project, further solidifying its role in the AI revolution.

Additionally, SoftBank is in talks to acquire Ampere, an Oracle-backed chip designer specializing in Arm-based server chips, in a deal worth up to $6.5 billion. This acquisition could accelerate Arm’s entry into server chip manufacturing.

Challenges & Industry Disruption

Arm’s decision to produce its own chips puts it in direct competition with some of its biggest customers, including:

• Qualcomm – Already engaged in a legal dispute with Arm over licensing terms.

• Nvidia – The world’s most valuable chipmaker, which has relied on Arm’s architecture but could now see Arm as a rival.

Despite the risks, Arm’s valuation has more than doubled to $160 billion since going public in 2023, driven by its strong partnerships with Nvidia, Amazon, and AI leaders like OpenAI and Meta.

Looking Ahead: A New Era for Arm & AI Computing

Arm’s shift from licensing to manufacturing represents a bold expansion that could reshape the semiconductor industry. With AI workloads driving demand for power-efficient chips, Arm’s low-power, high-performance designs make it an attractive alternative to traditional x86 processors from Intel and AMD.

If successful, Arm’s first in-house CPU could cement its position as a leader in AI-driven data center computing while fueling SoftBank’s grand AI ambitions. However, this move also risks disrupting relationships with long-time partners, setting the stage for new competition in the AI chip market.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.