Apple weighs return to enterprise servers using Nvidia networking tech
Apple is reportedly discussing Nvidia’s high‑bandwidth networking for a future AI inference server using two or four M8 Ultra processors; the project is exploratory and not expected before 2029.
In this brief: 3 sections 2 min read
Reported source: The Information, cited by Reuters.
Architecture: server could pair Apple M8 Ultra processors with Nvidia networking (NVLink/NVLink Fusion).
Configurations under discussion: two‑chip or four‑chip M8 Ultra setups for inference workloads.
Timing and status: not expected before 2029; could be canceled or proceed without Nvidia technology.
Expands Apple’s custom silicon use from consumer/workstation to enterprise data centers.
Positive for Nvidia’s networking franchise if talks lead to a commercial design‑win.
Long timeline and exploratory status mean near‑term financial impact is likely limited unless Apple confirms program or book orders.
Any company or supplier disclosures, SEC/exchange filings, or Apple investor‑relations commentary would be material.
Earnings‑call capex guidance or supply‑chain procurement notices could indicate program progress.
Nvidia statements characterizing the engagement (networking vs full systems) would change market implications.