^{Photo: Bloomberg News}

Arm Holdings saw its stock jump roughly 6% in after-hours trading after unveiling an ambitious long-term growth strategy centered around artificial intelligence and its first-ever in-house chip. The announcement marks a significant shift for the UK-based chip designer, which is now moving beyond its traditional licensing model into direct silicon sales.

At a high-profile event in San Francisco, CEO Rene Haas laid out a bold financial roadmap, projecting total annual revenue of $25 billion by 2031—more than six times the company’s 2025 revenue, which stood just above $4 billion. A substantial portion of that growth, approximately $15 billion, is expected to come from a newly introduced processor known as the AGI CPU.

The AGI CPU represents a strategic turning point for Arm. Designed specifically for AI inference workloads, the chip is optimized for the next wave of computing driven by agentic AI systems. These systems require significantly higher processing efficiency and scalability, creating new demand for advanced central processing units.

According to Haas, demand for CPUs is entering a new growth cycle, fueled by AI applications that go beyond traditional cloud computing. He suggested that CPU demand tied to AI could increase as much as fourfold in the coming years, potentially exceeding even current projections.

The initial customer for the AGI CPU is Meta, signaling strong early validation from one of the world’s largest AI infrastructure players. The chip was developed at Arm’s advanced facility in Austin, Texas, highlighting the company’s increasing investment in end-to-end product development.

This move positions Arm in direct competition with some of its own ecosystem partners, a notable departure from its long-standing role as a neutral technology provider. For decades, Arm has operated as the “Switzerland” of the semiconductor industry, licensing its architecture to chipmakers and collecting royalties on billions of devices—from smartphones to servers.

Now, by producing its own chips, Arm is entering a more competitive and potentially more lucrative segment of the market. CFO Jason Child indicated that the new chip line is expected to generate gross margins of around 50%, significantly expanding the company’s profit potential.

The broader semiconductor landscape has already been shifting toward Arm-based designs. Since launching its Neoverse platform in 2018, Arm has steadily gained traction in data centers, challenging traditional x86 dominance led by Intel and AMD. Major cloud providers including Amazon, Google, and Microsoft have already adopted Arm-based architectures for their custom AI chips.

The introduction of a fully integrated Arm-designed processor could further accelerate this trend, particularly for companies that lack the resources to build their own silicon from scratch. Arm executives have emphasized that the new chip will be competitively priced, with industry estimates suggesting costs could run into the thousands of dollars per unit.

Importantly, Arm has reassured existing partners that it will not force them to adopt this new model. Instead, the company aims to offer additional flexibility—allowing customers to choose between licensing designs or purchasing complete chip solutions.

From an investor perspective, this strategic pivot introduces a new layer of complexity. Analysts note that Arm is no longer just a licensing and royalty business but is evolving into a hybrid model with direct revenue streams from hardware sales. This shift could unlock significant upside but may also require time for markets to fully understand and price in the new growth trajectory.

Ultimately, Arm’s latest move reflects a broader transformation in the semiconductor industry, where control over both architecture and hardware is becoming increasingly valuable in the AI era. With strong early demand signals and aggressive financial targets, the company is positioning itself at the center of one of the most important technology shifts of the decade.