Apple Wants Back in the Server Business — With Nvidia's Help
Apple could soon sell the machines that run AI for other companies.
Apple is quietly exploring a return to selling computer servers — the powerful machines that sit in data centers and run AI software. According to a report from The Information, the company is designing a server built around its own 'M8' chips (the same family of processors used in Macs) and has talked with Nvidia about using its networking hardware to connect chips together. The goal: sell these machines to outside companies, not just use them internally.
Why would Apple do this? Because more and more businesses want to run AI on their own equipment instead of renting time from cloud providers like Amazon or Microsoft. This is often called 'on-premise' AI, and it matters for companies with sensitive data — hospitals, banks, law firms — that don't want their information leaving the building. Apple is reportedly focusing on 'inference,' which is simply the act of an AI model answering questions or generating text after it's already been trained.
This would be a notable comeback. Apple last sold servers to the public with its Xserve line, which it discontinued back in 2011. Since then, the server market has been dominated by Dell, HPE, Supermicro, and others. Apple's pitch would likely lean on its reputation for privacy and its tight control over both hardware and software.
But don't hold your breath. The report says neither the Nvidia partnership nor the server itself is finalized, and a potential launch date is listed as 2029 — roughly four years away. Apple could cancel the project before then, and it has killed plenty of products in the past. For now, it's a signal of where the company sees growth: AI infrastructure, not just AI features on your iPhone.
- Apple reportedly wants to sell AI servers to other companies, using its own 'M8' chips and possibly Nvidia networking hardware.
- The focus is 'inference' — running AI models that answer questions — for businesses that want to keep their data on their own machines.
- It's far from certain: no deal is signed, and the report points to a possible 2029 launch that could still be scrapped.
Why It Matters
More private AI options for businesses could mean better data security and more competition in the AI hardware market.