Apple May Return to Servers With Its Own AI Chips, Possibly Using Nvidia Technology
Apple is reportedly developing enterprise servers based on its future M8 Ultra processors for AI inference applications. The company may use up to four of these chips in a system potentially launching as early as 2029. Notably, Apple could incorporate NVIDIA's NVLink Fusion technology to connect the processors, signaling an evolving strategy where Nvidia benefits from interconnect solutions even when its GPUs are not used. Nvidia’s involvement highlights a broader shift in AI hardware economics and its efforts to diversify revenue beyond GPU sales by monetizing networking components. This approach allows Apple to leverage its own silicon while still relying on Nvidia for critical infrastructure, potentially capturing up to 10-15% of data-center costs through connectivity solutions. The development could reshape the competitive landscape in AI hardware and server markets as Apple reenters a space it left over a decade ago. While neither plan is finalized, both companies stand to gain from this potential collaboration.
Topics: AI servers Apple M8 Ultra enterprise hardware NVIDIA NVLink Fusion