Apple is reportedly shopping for AI chip startups, and the reason why tells you a lot about where the company’s AI ambitions currently stand. According to The Information, cited by Reuters, Apple has been holding discussions with investment bankers and semiconductor startups about potential acquisitions aimed at strengthening the server-side hardware that powers Apple Intelligence and Siri. Apple hasn’t confirmed any of this publicly.
The context behind the move is the interesting part. Apple’s AI servers currently rely on M2 Ultra chips, and those chips have reportedly struggled with some large-scale AI inference workloads. A dedicated next-generation AI server processor internally codenamed “Baltra” was in the works, but that project has reportedly been delayed. Facing a gap in capability and timeline, Apple needed a solution and acquisitions offer a faster path to specialized AI chip architectures and engineering talent than building everything from scratch.
The Siri connection makes this even more notable. Earlier this year, Apple was reportedly testing Google Gemini models as part of its next-generation Siri development. The problem was Apple’s own server hardware couldn’t efficiently run those large AI models, so the company temporarily leaned on Nvidia GPUs operating inside Google’s cloud infrastructure to keep development moving. That’s a significant admission of infrastructure limitations for a company that has built its identity around controlling its own silicon stack.
An acquisition strategy fits Apple’s history. The company has consistently used smaller, targeted purchases to absorb technology and talent rather than making splashy mega-deals. The A-series and M-series chip dominance Apple enjoys today was partly built on acquisitions like PA Semi and Intrinsity years ago. The same playbook applied to AI server silicon could accelerate Apple’s path to competitive cloud AI infrastructure.
This also sits alongside Apple’s recently announced $30 billion-plus multiyear agreement with Broadcom to expand production of custom silicon components in the United States, targeting more than 15 billion US-made chips. Together, these moves paint a picture of Apple getting serious about owning more of its AI infrastructure rather than depending on Nvidia and third-party cloud providers long term.
No acquisition targets have been named and nothing is confirmed. But the direction is clear: Apple is building toward AI infrastructure independence, and it’s willing to buy its way there faster if the right opportunity presents itself.
Discover more from Phoonomo
Subscribe to get the latest posts sent to your email.
