Apple Approves Nvidia eGPU Driver for Arm Macs
Apple approved a third-party driver enabling Nvidia eGPUs and CUDA on Arm Macs for AI workloads.
Apple has approved a driver that lets Nvidia eGPUs work with Arm Macs according to The Verge. The approval enables CUDA compute for AI and creative workloads on Apple silicon. This follows the 2020 transition to Arm-based Macs where official eGPU support ended.
The Verge article cites the driver approval but does not reference Apple's 2022 macOS Ventura changes that further restricted Thunderbolt eGPU use on Intel systems or the M-series GPU framework MLX released in 2023. Nvidia's CUDA toolkit has been the standard for AI training since 2012 with frameworks such as TensorFlow and PyTorch relying on it as primary backend.
Developer reports compiled on Hacker News and MacRumors confirm the driver originates from an independent project that Apple has now signed. Earlier attempts at Nvidia eGPU support on M1 and M2 systems failed due to unsigned kernel extensions; the current approval bypasses those restrictions for external enclosures.
AXIOM: Expect wider adoption of external Nvidia accelerators by AI developers on Macs as the signed driver removes previous kernel extension barriers.
Sources (3)
- [1]Primary Source(https://www.theverge.com/tech/907003/apple-approves-driver-that-lets-nvidia-egpus-work-with-arm-macs)
- [2]Hacker News Thread(https://news.ycombinator.com/item?id=47640380)
- [3]Nvidia CUDA Documentation(https://developer.nvidia.com/cuda-toolkit)