Apple is reportedly working on a server for AI workloads that would be built around its high-performance M-series Ultra chips, a move that could bring the company back into the enterprise server market after nearly two decades away. According to The Information, the project is aimed at a 2029 release and could use either two or four future M8 Ultra chips in each system.
Apple’s reported AI server effort
The potential server would be a notable expansion of Apple’s silicon strategy. Instead of designing the machine around a traditional data center processor, Apple would rely on the same family of chips that powers its Mac desktops, including the high-end Ultra variants used in products such as the Mac Studio and Mac mini.
The Information said the server would come in two configurations and that the project received support from Apple CEO John Ternus when it began about a year ago, at a time when he was still leading the company’s hardware engineering work. If the project reaches the market, it would represent Apple’s first server launch since the company retired the Xserve in January 2011.
Why Apple may be moving now
The timing appears closely tied to a broader shift in the AI industry: developers and companies are increasingly buying Apple hardware to run AI workloads. Apple’s Mac mini and Mac Studio have seen strong sales as AI teams look for compact machines with good performance and relatively attractive power efficiency.
That demand has created an unusual dynamic. Consumer-oriented computers built on Apple silicon are being used in settings that resemble parts of a distributed AI infrastructure, especially for training and experimentation. In that context, a server product built specifically for enterprise use would be a more direct response to the same market pull.
The Information reported that AI companies including OpenAI have purchased “tens of thousands” of Mac minis and Mac Studios to train AI agents through trial-and-error reinforcement learning. Anthropic has also rented Mac minis from Amazon Web Services, according to the report. Those examples help explain why Apple might see a business case for packaging its chips into a dedicated server platform.
M-series Ultra chips at the center
Apple’s M-series Ultra chips are the company’s highest-end desktop-class processors, combining strong performance with the tight hardware-software integration that has become a hallmark of Apple silicon. The reported server would use future M8 Ultra chips, suggesting Apple is thinking well ahead rather than adapting an existing product for immediate release.
Using Ultra chips in a server would be unconventional in a market long dominated by Intel, AMD, and custom AI accelerators. But Apple has already shown that it is willing to push its own silicon into product categories where its architecture can serve a distinct use case, especially when power efficiency and tightly integrated systems matter.
Possible Nvidia connection
Apple is also reportedly weighing a networking approach that would connect its M8 chips using equipment from Nvidia. The Information said Apple and Nvidia have been discussing how Apple might use Nvidia’s NVLink Fusion technology as part of its AI strategy.
That detail is notable because Nvidia’s hardware and software stack has become central to modern AI infrastructure. If Apple were to adopt NVLink Fusion in some form, it could help the company link multiple chips more effectively inside a server or cluster. At the same time, The Information cautioned that the project could still be canceled or move forward without Nvidia’s technology.
In other words, the server concept is not confirmed as a finished product, and the architecture is still being debated. The presence of Nvidia in the conversation, however, suggests Apple may be looking at ways to make its own silicon fit more comfortably into larger AI deployments.
A return to enterprise hardware for Apple
If the reported project advances, it would mark a meaningful return to an area Apple once abandoned. The Xserve, which was discontinued in 2011, was Apple’s last major push into server hardware. Since then, Apple has focused primarily on consumer devices, laptops, desktops, tablets, and phones, with its own chips increasingly defining those products.
A server built for AI workloads would therefore be more than a new product category. It would be a signal that Apple sees an opening in enterprise infrastructure, especially where its silicon already has an unexpected foothold among developers and AI researchers.
It would also fit the company’s long-running emphasis on vertical integration. Apple designs the chips, controls the operating system, and could, in theory, tune the entire stack for specific AI tasks. That kind of control has helped Apple differentiate its consumer hardware, and it may be part of the company’s thinking for server hardware as well.
John Ternus and the project’s significance
The report also links the server effort to John Ternus, who officially became Apple CEO on September 1 after Tim Cook’s 15-year run in the role. The Information says the project gained Ternus’ support when it began, before he assumed the company’s top job.
If the server eventually ships, it could become one of the first major hardware initiatives associated with Ternus’ leadership. That makes the project potentially important not just as a product line, but as a marker of how Apple intends to position itself in the AI era.
Apple has generally moved more cautiously than some competitors in the AI race, at least in terms of public-facing infrastructure and partnerships. A server product would not erase that perception overnight, but it would show that Apple is at least considering a more direct role in AI compute rather than relying only on devices at the edge.
Supply chain pressure could complicate the plan
Even if Apple wants to proceed, the broader semiconductor environment may make execution harder. The same AI boom that is encouraging companies to buy more hardware has also intensified demand for memory chips, creating a shortage across the tech industry.
That shortage has already pushed up prices for a range of products, including gaming devices, flagship smartphones, and multiple Apple products. As data centers race to secure memory for AI systems, consumer electronics makers are being squeezed by tighter supply and higher costs.
For Apple, that matters in two ways. First, it could raise the cost of building a server product designed for AI customers. Second, it could complicate the company’s existing Mac lineup if demand for memory continues to outstrip supply.
What to watch next
For now, the reported Apple server remains a plan rather than a confirmed product. The Information’s account suggests that Apple is exploring multiple paths, including a possible Nvidia-assisted networking approach, while also leaving room for the project to be delayed or dropped entirely.
Still, the report is significant because it connects several trends at once: the rise of Apple silicon in AI workflows, the company’s renewed interest in enterprise infrastructure, and the growing pressure on hardware makers to support the demands of AI development. If Apple does bring the server to market in 2029, it would be one of the clearest signs yet that the company is prepared to use its M-series Ultra chips far beyond the Mac.
Source: Original report
Was this helpful?
Explore more: More Hardware & Devices Tech News Blog
Last Modified: September 17, 2026 at 10:31 pm
0 views
