OpenAI Hoards Tens of Thousands of Macs for Agent Training as Apple Emerges as AI Infrastructure Player
机器之心 · wechat · 2026-08-31
According to The Information, OpenAI has purchased tens of thousands of Mac mini and Mac Studio units in recent months to build training grounds for computer-using agents. These devices leverage the native macOS environment for task execution and trajectory collection to facilitate reinforcement learning. Simultaneously, Anthropic is renting significant Mac compute power via AWS. Apple's M-series chips, featuring unified memory architecture (up to 512GB) and multi-machine clustering, are well-suited for local LLM inference and agent execution. Consequently, Mac mini supplies have tightened, driving significant revenue growth in Apple's Mac business. Nvidia is also entering the fray with its RTX Spark series, competing with Apple for desktop AI compute dominance. The surge in memory demand has driven up chip prices, shifting the AI compute battle from cloud data centers to desktop terminals.
More from Infra
- MiniCPM-V 4.5 MoE reportedly hits 13+ tokens/s on iPhone with just 3GB memory — heyshrutimishra · 2026-10-16
- Solo-built Vulkan inference engine LaForge hits ComfyUI-level SD1.5 speed: 512px in 0.8s, 402MB container — Ichidan256 · 2026-10-16
- Trading 2 RTX 5090s for 4 R9700s: does doubled VRAM pay off? — poofph · 2026-10-16
- Running PolyStrata + Cline coding agent on a $0-fancy laptop at ~90 tok/s — rpwoerk · 2026-10-16
- TensorFold 1.0.12 runs 120B vision model on 1-2 DGX Sparks, adds Mac image input — HankYeomans · 2026-10-16
- vLLM brings frontier model local deployment to NVIDIA RTX Spark PCs — vllm_project · 2026-10-16