OpenAI buys tens of thousands of Macs to train AI, Apple unexpectedly steps into NVIDIA's territory OpenAI has purchased tens of thousands of screenless, keyboard-less Mac minis and Mac Studios to develop reinforcement learning training and 'computer operation' AI agents. Anthropic is also renting a large amount of Mac computing power through AWS. The core advantage lies in Apple chips' unified memory architecture, which can directly load quantized models with tens of billions or even hundreds of billions of parameters, without the need to shuttle data back and forth between VRAM and RAM like NVIDIA GPUs. This wave of enterprise demand has directly boosted financial reports: Mac's latest quarterly revenue surged nearly 29% year-over-year to $10.3 billion, making it Apple's fastest-growing hardware product line. NVIDIA internally now views Apple as the biggest competitor in the local AI market, and its recently launched DGX Spark even resembles the Mac mini in design. However, the supply chain is becoming a bottleneck—demand for AI chips has caused an industry-wide memory shortage. High-spec Macs have been out of stock for months, and some companies are turning to NVIDIA products with more readily available inventory.