Arm Holdings plc, the global leader in processor architecture design, announces a major strategic shift: moving from licensing its chip designs to manufacturing and selling its own processors.
Strategic Changes and Ambitions
Chief Financial Officer Jason Child disclosed the plan on 17 August. Arm is actively seeking acquisition targets, including smaller firms similar to DreamBig, and does not rule out large mergers or purchases. Over the past two decades the company has acquired roughly twenty small private enterprises. When evaluating deals, Arm focuses on whether the target can expand its total addressable market or enhance internal capabilities for research, development, and production.
Challenges of Chip Manufacturing
Child noted that producing finished chips is substantially more complex than licensing designs. Arm must secure its own foundry capacity and memory supply. In the context of the global memory shortage, new entrants will have to queue for manufacturing resources. Arm aims to increase its share in chip fabrication and memory supply within two years, starting from zero in this area.
Role in AI Development
Memory remains a critical bottleneck for artificial‑intelligence infrastructure. Additionally, constraints on chip‑fabrication capacity and power availability limit the construction of large AI data centers, as previously highlighted by TSMC. Arm processors will be paired with NVIDIA GPUs in AI servers. The company forecasts that customer demand for its AI‑focused chips will exceed US$2 billion in fiscal year 2027‑2028.
Historic Step: First Self‑Produced Chip
Arm plans to launch its AGI CPU in March 2026 – the first chip designed and manufactured in‑house after 36 years of existence (founded in 1990). This processor targets inference workloads in AI data centers. Early adopters include Meta, OpenAI, Cloudflare, and SAP.
Xpert Take
Arm’s transition to in‑house manufacturing marks a fundamental shift in its business model, moving from a pure IP licensor to a vertically integrated semiconductor player. This move could intensify competition with established foundries and IDMs, particularly in the AI‑chip segment where supply‑chain resilience is paramount. Success will hinge on Arm’s ability to secure reliable fab partnerships, manage memory‑supply risks, and deliver performance‑competitive silicon that complements rather than cannibalizes its existing ecosystem. If executed well, the strategy may unlock new revenue streams and strengthen Arm’s position amid growing demand for specialized AI hardware.
