Editorial illustration for Apple Reportedly Building AI Servers With M8 Ultra Chips, Nvidia NVLink
Apple Builds AI Servers With M8 Ultra, Nvidia NVLink
Apple wants a seat at the AI data center table, and it's building the hardware to get there. According to The Information, citing people familiar with the project, Apple is developing an enterprise server aimed squarely at AI developers, businesses, and governments, not consumers. The server would come in two configurations, one with two M8 Ultra chips and another with four, both designed for AI inference workloads rather than training.
The timeline is long. Reports suggest Apple wouldn't ship this before 2029, and the project could still fall apart or take a different technical path before then. One open question is how Apple links those chips together inside a rack. The company is reportedly weighing Nvidia's NVLink Fusion, a connection technology Nvidia built for its own silicon but has since opened up to outside hardware makers.
The push reportedly has backing at the top. John Ternus, Apple's new CEO, supported the effort roughly a year ago while he was still running hardware, betting on demand in the AI compute market. That bet looks less speculative now that firms like OpenAI and Anthropic are buying Mac Minis and Mac Studios in bulk.
Apple is developing an enterprise server with its own chips, targeting AI developers, businesses, and governments. The server would ship in two versions with either two or four M8 Ultra chips, according to The Information, citing people familiar with the project. It's built for AI inference, running already-trained models.
Why this matters
A 2029 target is a lifetime away in chip years, and The Information's own sourcing flags this could still get scrapped. So treat this as a signal of intent, not a product roadmap. What's notable is the NVLink Fusion detail: Apple, a company that builds its silicon empire on tight vertical control, apparently needs Nvidia's plumbing to make its own M8 Ultra chips talk to each other fast enough for data-center work. That's a tell about how far ahead Nvidia's interconnect advantage still sits, even for a company with Apple's cash and chip design talent.
For developers and founders, the practical takeaway is patience. Two or four M8 Ultra chips aimed at inference, not training, suggests Apple wants a foothold in serving already-built models to enterprises and governments, not competing for the frontier-training market Nvidia and AMD fight over. If it ships, it's another inference option years out. If it doesn't, it's a reminder that entering AI hardware means depending on Nvidia's ecosystem even to compete with it.
Related Reading
Open Source D&B rebuilds 642 million‑business database after AI agents hit limits
Open Source OpenAI's GPT Transcribe Cuts Error Rate to 3.31%, Improving on GPT-4o
Open Source Study: 6.7% of Deepfake Requests on Hugging Face Targeted Children
Related Topic NVIDIA and Google Cloud let developers scale AI from prototype to production
Market Trends
Related Topic NVIDIA NeMo powers telco reasoning model for autonomous network workflows
Industry Applications
From Archives EU President Warns AI Agent Escape Just a Preview of Coming Hacks
From Archives Mozilla Report: Open AI Models Should Be Default for Most Work
From Archives NVIDIA's Vera Rubin NVL72 Leads MLPerf Inference Debut
Research & Benchmarks
From Archives Nvidia CEO Jensen Huang: Innovation and Safety Aren't a 'False Choice
Business & Startups
Common Questions Answered
What is Apple's M8 Ultra server designed for, and who is it targeting?
Apple's M8 Ultra server is designed specifically for AI inference workloads rather than training, and it targets enterprise customers including AI developers, businesses, and governments. The server comes in two configurations: one with two M8 Ultra chips and another with four chips, positioning it as an enterprise product rather than a consumer offering.
Why is Apple reportedly using Nvidia NVLink in its M8 Ultra chips?
According to the article, Apple needs Nvidia's NVLink interconnect technology to enable its M8 Ultra chips to communicate with each other at the speeds required for data-center operations. This reveals that despite Apple's reputation for vertical integration and control of its own silicon, it still requires Nvidia's advanced interconnect advantage for enterprise-scale AI infrastructure.
When is Apple expected to ship its AI server with M8 Ultra chips?
Apple is reportedly targeting a 2029 release date for its enterprise AI server, which is several years away. However, the article notes this timeline should be treated as a signal of intent rather than a confirmed product roadmap, as the project could still be scrapped before launch.
How does Apple's M8 Ultra server differentiate itself from competitors in the AI data center market?
Apple's M8 Ultra server uses Apple's proprietary chip architecture rather than relying on standard processors, allowing the company to enter the AI data center market with custom silicon. The server specifically focuses on AI inference workloads, positioning it as a specialized solution for running already-trained models rather than a general-purpose data center processor.
Further Reading
- Apple Is Developing Enterprise Server for AI Age, Report Says - Bloomberg
- Apple weighs return to server market with M8 Ultra AI machines, talks Nvidia networking - MacDailyNews
- Apple planning to sell AI servers powered by M8 Ultra chips, says report - 9to5Mac
- Apple weighs Nvidia technology for potential server market return, Information reports - Reuters
- Apple plans to sell AI servers powered by M8 Ultra chips, says report - iPhone in Canada
