Key facts
- Apple is reportedly developing an enterprise server using its M-series Ultra chips.
- The server is expected to feature two or four M8 Ultra chips.
- The potential debut for the server is 2029.
- This would be Apple's first enterprise server in nearly 20 years.
- AI companies like OpenAI and Anthropic are using Mac computers for AI workloads.
- Apple is considering integrating Nvidia's NVLink Fusion technology.
Apple is reportedly developing a new enterprise server designed for artificial intelligence workloads, utilizing its high-performance M-series Ultra chips. The project, which could debut in 2029, would mark Apple's return to the server market after a nearly two-decade absence. The server is expected to come in configurations with either two or four of Apple's future M8 Ultra chips. This initiative reportedly received support from new Apple CEO John Ternus when he led the company's hardware engineering efforts a year ago. The move aligns with the increasing adoption of Apple hardware, such as Mac minis and Mac Studios, by AI developers and companies for training AI agents. OpenAI has reportedly purchased tens of thousands of these Mac computers, while Anthropic has been renting Mac minis from AWS for similar purposes. Apple is also exploring the integration of Nvidia's NVLink Fusion technology for its AI server, following discussions between the two companies. However, The Information cautioned that the project could still be canceled or proceed without Nvidia's technology. The potential launch of this AI server could become a significant part of John Ternus' legacy, following his transition to CEO from Tim Cook on September 1. The tech industry, including Apple, faces ongoing memory chip shortages due to the surge in demand for AI data centers, which could pose a supply chain challenge for the new server.