Arm Holdings Reports Record Q4 Revenue as Data Centre Demand Doubles for New AGI Chip

Arm Holdings has reported a record fourth quarter of 2026, posting revenue of $1.49 billion. This figure represents a 20 per cent year-on-year increase, marking the company's highest quarterly revenue to date and exceeding the midpoint of its guidance. The financial performance was underpinned by significant growth in both licensing and royalty streams, with licensing revenue rising 29 per cent to $819 million and royalty revenue increasing 11 per cent to $671 million.

The primary driver of this expansion is the unprecedented demand for the company's newly launched Arm AGI CPU, designed specifically for agentic AI workloads. Customer orders for this silicon have doubled to over $2 billion across fiscal years 2027 and 2028, a stark contrast to the initial $1 billion forecast provided at the time of the product launch. Executives confirmed that this surge in demand is being met by a strategic shift in focus, with the data centre segment on track to become Arm's largest business area.

This transition is supported by a growing number of major hyperscalers adopting Arm-based silicon for their AI infrastructure. Key industry players, including AWS, Google, and NVIDIA, are increasingly utilising Arm chips as head nodes alongside their accelerator-based systems. Google recently announced plans to replace x86 host processors with custom Arm Axion CPUs in its TPU systems, while AWS continues to scale its custom silicon strategy with Arm-based Graviton alongside its Trainium and Nitro offerings.

Despite the doubled demand, management has chosen to maintain its current revenue guidance for the quarter, citing ongoing supply chain constraints. Jason Child, Chief Financial Officer, noted that while the company has line of sight to more than $2 billion of demand, the first production chip revenues are expected to land in the fourth quarter of the current fiscal year. The company is currently working around the clock to secure the necessary supply chain capacity, including access to wafers and memory, to support the increased volume.

Looking ahead, Arm executives outlined a vision where the ratio of CPUs to GPUs shifts not in chip count, but in core count. Rene Haas, CEO, indicated that future Arm AGI CPUs could reach 256 or 512 cores to handle agentic AI workloads efficiently. This strategy aims to capture the majority of the $100 billion-plus total addressable market by 2030, with Arm claiming a current 50 per cent market share in the data centre CPU space.

The company's long-term strategy relies on two growth vectors: expanding royalties through IP and compute subsystems, and adding silicon as a new revenue stream. With over 350 billion chips shipped to date and a robust ecosystem of more than 50 leading companies supporting the platform, Arm positions itself at the centre of the next generation of AI workloads, moving from cloud infrastructure to the edge and the physical world.