Apple is looking to acquire AI chip companies to cut its dependence on Nvidia for heavy AI tasks, according to The Information. While Apple processes some AI work in its own data centers with its own chips, it relies on Nvidia hardware in Google Cloud for demanding workloads such as the Gemini model that powers Siri. Apple's own server chip, code‑named Baltra, was expected to ship this year but has been delayed.

Apple’s acquisition history has centered on deals in the hundreds of millions, but January’s purchase of Q.ai for $2 billion—second only to the $3 billion Beats deal—shows a shift toward larger spends. CFO Kevan Parekh said Apple will no longer target net‑cash neutral, giving it more flexibility to fund big deals. The company is also eyeing smaller AI firms that can shrink models for on‑device use.

Apple’s chip‑design roots trace back to its 2008 acquisition of PA Semi for $278 million. With John Ternus set to replace Tim Cook and Johny Srouji’s expanded hardware role, dealmaking may accelerate. Apple’s internal team focuses on mobile chips; it is developing server chips such as M5 Ultra and M7 Ultra, the latter slated to support up to 1.5 TB memory by 2029, aiming to rival Nvidia’s Blackwell. A partnership with Broadcom continues through 2031.