Arm moves beyond licensing to build its own AI chips
Arm reported quarterly revenue of $1.5 billion and raised its forecast for its new data centre processor to $2 billion in 2027 and 2028, marking a significant expansion beyond its traditional business of licensing chip designs to other manufacturers.
The UK-based chip designer has historically made money by licensing its processor architectures to companies like Apple, Qualcomm, and Nvidia. Those firms then design and manufacture their own products. The shift into building complete processors represents a fundamental change in how Arm generates revenue.
Chief executive Rene Haas said demand for the new data centre CPU had "exceeded expectations, reinforcing Arm as the compute platform for the AI era."
Why CPUs matter now
Graphics processing units dominated the early stages of AI deployment. Central processing units are now playing an increasingly important role in running inference workloads and managing system-level tasks in data centres.
Arm believes CPU demand for AI applications will grow substantially in the coming years. The company's projections suggest a significant market opportunity.
The licensing business still performs
Arm's traditional licensing revenue exceeded analyst forecasts in the latest quarter. Royalties-historically a core income stream tied to device shipments-came in slightly below expectations.
The company's newer designs aimed at AI-enabled devices drove strong licensing performance despite pressures in the broader smartphone market, which faces supply constraints from memory shortages driven by AI data centre demand.
A strategy shift with built-in risks
Parent company SoftBank has been seeking a stronger presence across the AI computing stack. Project Izanagi, an initiative to create alternatives to established players in AI hardware, reflects this ambition to compete more directly in high-performance computing.
By designing and selling its own chips, Arm is now competing directly with customers that have historically relied on its IP. Large technology firms developing their own AI infrastructure could view Arm as both a partner and a rival.
Balancing these relationships while expanding into product development will be critical as the strategy evolves.
What this means for product teams
For product development professionals, Arm's transition signals a broader industry shift. The company is moving from a pure licensing model to a diversified business combining architecture design with complete silicon solutions.
This strategy reflects growing demand for processors optimized for specific AI workloads. Teams building AI infrastructure or considering processor options may find themselves evaluating Arm's own chips alongside designs from traditional customers.
Understanding AI for Product Development can help product teams navigate these changing market dynamics and evaluate new hardware options as they emerge.
Your membership also unlocks:
