Apple retired its Xserve line in 2011 and has spent the decade since letting enterprise servers become someone else’s problem. Now The Information reports the company is planning to get back into the business, and the reason is simple: the AI industry needs more compute power, and Apple believes its ARM-based M processors are the right ones to provide it.
The report comes with a specific timeline and a specific partner. Apple is reportedly borrowing tech from Nvidia to build its new server line, and the product is unlikely to debut until 2029. The M8 Ultra chip is the centerpiece, though Apple only just announced the M6 in August, meaning the M8 line is still years away.
The Return of the Server Game
Apple’s departure from servers was quiet. The Xserve line went away in 2011, and since then the company has mostly let other manufacturers handle the enterprise market. That changed as AI development shifted toward systems built for machine learning, and Apple’s own hardware became a draw for developers building AI models.
The Mac Mini and Mac Studio have proven popular with AI developers, according to the report, and that popularity has led to shortages. A server line built around those same processors could address that demand directly.
The company’s move back into servers is a bet that the AI industry’s appetite for compute power will keep growing. Apple is positioning itself as a supplier to that market, and the M8 Ultra chip is the proposed engine for those servers.
Nvidia’s Role in the Deal
The deal with Nvidia is the key element of the report. Apple is reportedly borrowing tech from Nvidia to build its new server line, and the rumor is that Nvidia’s NVLink Fusion tech could be part of it.
NVLink connects multiple chips together so they can operate as one system. The report suggests Apple’s servers could incorporate this tech, which could lead to further collaboration between the two companies. That collaboration could deepen beyond the initial partnership, though the report notes plans could still change.
Apple might make servers again to cash in on the AI rush.
The two companies have had a frosty relationship behind the scenes, the report notes. That history complicates the picture, but it also explains why the partnership is notable. Apple turning to Nvidia chips to power its revamped Siri servers is a recent shift, and it suggests the relationship has softened in practical terms even if the old tensions remain.
What the M8 Line Actually Is
The M8 Ultra is the chip that would power the new servers. Apple only just announced the M6 in August, so the M8 line is a ways out. The report says the server product, whether it’s called Xserve or something else, will run on two or four M8 Ultra chips depending on the model.
That’s a significant jump in capacity. A server with four M8 Ultra chips would be a serious piece of hardware.
The M8 line is unproven, but the company’s track record with ARM-based processors suggests it has the design capability to deliver.
The Timeline Problem
The 2029 debut is the furthest-out date in the report. That is a long wait, and it raises questions about whether the market will still want what Apple is offering by then.
The AI industry moves fast. Compute demands change quickly, and the landscape between now and 2029 could look very different. Apple is betting that the current trend toward ARM-based processors will hold, and that its M8 Ultra will arrive in time to meet demand.
The NVLink Fusion tech adds another layer of uncertainty. If Apple incorporates it, the design becomes more complex, and the timeline could slip further. Plans could still change, the report notes, and that caution applies to everything else in the report.
Why This Matters
Apple’s return to servers is a statement. The company is entering a market it abandoned a decade ago, and it is doing so with a specific technical approach: ARM-based processors that emphasize efficiency. The M8 Ultra is the proposed engine for these servers, and the NVLink Fusion tech from Nvidia could give them a multi-chip architecture that matches what the AI industry wants.
The frosty history between Apple and Nvidia is worth keeping in mind. The report describes a frosty relationship behind the scenes, but it also notes that Apple turned to Nvidia chips for its revamped Siri servers. That shift suggests the relationship has softened in practical terms, even if the old tensions remain.
The Mac Mini and Mac Studio shortages have been a problem for developers, and a server line could alleviate some of that pressure. If Apple can produce enough of these machines, it could help stabilize the market for AI development resources.
The 2029 debut is the furthest-out date in the report. That is a long wait, and it raises questions about whether the market will still want what Apple is offering by then. The AI industry moves fast, and the landscape between now and 2029 could look very different.
The M8 line is unproven, but the company’s track record with ARM-based processors suggests it has the design capability to deliver. The NVLink Fusion tech adds another layer of uncertainty, and plans could still change.
The report is early, and the details are speculative. Apple has not confirmed any of this, and the 2029 date is a projection, not a promise. But the signal is clear: Apple sees a future in servers, and it is willing to invest heavily to get there. Whether that future arrives on time is another question entirely.
See the a run of 28 images at The Verge.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.