This week, SoftBank announced that it will be acquiring Ampere Computing, a startup that designs Arm-based server chips, for $6.5 billion. However, this isn’t just another tech acquisition — it’s part of Masayoshi Son's deliberate strategy to control AI's fundamental infrastructure. By securing Ampere's ARM-based technology and its 1,000 semiconductor engineers, SoftBank has solidified its position at every critical layer of the AI stack.
The acquisition aligns with Masa’s long-term investment thesis on artificial intelligence and builds on SoftBank’s 90% stake in ARM Holdings, its acquisition of AI chip designer Graphcore last year, and a $500 billion partnership with OpenAI for the Stargate project. However, SoftBank isn’t just investing in AI — it’s positioning itself as the backbone of the industry, from silicon to software.
By securing ARM, it dominates the fundamental IP behind AI chips. With Graphcore, it gains cutting-edge acceleration hardware. Through OpenAI, it has access to the most advanced AI models. SoftBank is vertically integrating the AI ecosystem, ensuring that the future of artificial intelligence runs on its infrastructure and capturing the means of AI production creating dependencies that competitors cannot easily circumvent.
In a statement about the acquisition, Masa said "Ampere's expertise in semiconductors and high-performance computing will help accelerate this vision, and deepens our commitment to AI innovation in the United States." This statement — albeit quite measured — indicates his ambition to make SoftBank a serious component of global AI development.
By controlling ARM architecture, which increasingly powers everything from mobile devices to data center servers, while simultaneously investing in specialized chips built on that architecture, SoftBank aims to create a self-reinforcing ecosystem. Every Ampere chip drives royalties to ARM, funding further R&D that widens the gap with competitors. Meanwhile, the partnership with OpenAI ensures SoftBank's hardware designs will be optimized for the world's most advanced AI models.
This acquisition reinforces the shift toward ARM in data centers, as hyperscalers like Amazon and Microsoft develop custom ARM chips, challenging x86’s dominance. Intel, once the leader in server CPUs, now faces mounting competition as ARM designs prove better suited for AI and cloud workloads.
Amid this transition, Masa Son’s endgame is clear: while the world fixates on AI breakthroughs, SoftBank is methodically acquiring ownership of the invisible infrastructure that makes those capabilities possible. This isn’t just about investing in AI — it’s about controlling the foundation on which the entire industry is built.
