Apple Announces Enterprise AI Server for 2029, Tells Enterprises to "Just Hold On"
The M8 Ultra-based system will be Apple's answer to a market that, by 2029, will have been asking the question for seven years.
CUPERTINO, Calif. — Apple confirmed Monday that it is developing an enterprise AI server built around a future M8 Ultra chip, with a launch targeted for 2029, and asked corporate customers currently evaluating AI infrastructure to "sit tight, maybe don't sign anything, we're coming."
"We take our time because we get it right," said an Apple executive in a statement. "When the Apple AI Server arrives, it will be beautifully designed, deeply integrated, and available in two finishes. It will also be the year 2029. We ask for your patience and, ideally, your budget."
The system is expected to use a high-bandwidth interconnect that Apple is reportedly exploring in partnership with Nvidia, a company Apple has not partnered with since a dispute over laptop graphics chips in 2008, which sources say both companies "have mostly moved on from" and which Apple "has not forgotten."
Enterprise buyers reacted with the mix of admiration and bewilderment that Apple's timelines traditionally provoke. "We're deploying 4,000 GPUs this quarter," said the CIO of a Fortune 500 insurer. "Apple has asked us to imagine a world three years from now in which we did not. I have imagined it. In that world I no longer have this job."
Apple said the server will run a version of macOS, will not require a keyboard or mouse but will ship with both, and will be repairable only at Apple-authorized data centers. Pricing was not announced, though the company noted the stand will be sold separately.
Analysts said the announcement was consistent with Apple's broader AI strategy, which one described as "arriving late with a nicer box," and noted that the company's original Siri overhaul, first promised for 2025, is now also targeted for 2029, "so at least everything will be on the same truck."