OpenAI has been exposed for frantically purchasing Macs, buying tens of thousands at a time.
They accidentally bought out the stock and are trying every means to acquire more.
They don't want MacBook laptops, specifically buying Mac mini and Mac Studio models without screens or keyboards.
So the question arises: What kind of AI business can't be handled by NVIDIA GPUs and Google TPUs, but must use Macs?
Using Tens of Thousands of Macs for Reinforcement Learning, AI Giants Are Doing It Too
According to The Information, OpenAI has purchased tens of thousands of Mac mini and Mac Studio units specifically for reinforcement learning.
Not just OpenAI, Anthropic is also renting Mac minis through Amazon Web Services (AWS) for similar tasks.
These Macs are used to train "computer-use agents," AI systems capable of autonomously operating computers to complete multi-step tasks like editing test code, automatically organizing email inboxes, and summarizing documents.
This frenzy is directly reflected in Apple's financial reports.
In the most recent quarter, Mac sales grew nearly 29% year-over-year, reaching $10.3 billion, outpacing the growth of iPhone, iPad, and all other Apple product lines. Mac became Apple's fastest-growing business.
On June 23rd, Apple Park hosted an event called "Business at the Park." This is uncommon in Apple's history, as the company has traditionally focused on the consumer market and rarely held events specifically for enterprise clients.
Executives from Disney and Ford attended the event, along with Anthropic co-founder Jared Kaplan. Apple's outgoing CEO Tim Cook and incoming CEO John Ternus were also present.
According to one attendee, Apple repeatedly emphasized at the event that its hardware is very suitable for handling AI tasks locally, with the Mac mini being the focus of the entire event.
AI training has long been dominated by NVIDIA GPUs, but Macs have been purchased on a large scale for this specific reinforcement learning niche due to their unified memory architecture.
NVIDIA GPU VRAM and system memory are separate, and data transfer between them creates bottlenecks. Apple's M-series chips use a single, shared memory pool where the CPU and GPU directly access the same memory, providing performance advantages when handling AI workloads.
Furthermore, unlike slim MacBooks, the Mac mini and Mac Studio are equipped with dedicated cooling systems, preventing thermal throttling during prolonged, complex AI tasks. This is crucial for reinforcement learning training sessions that can last for hours or even days.
Apple is also promoting the EXO Labs open-source software project, which can cluster multiple Macs together to run trillion-parameter AI models locally.
Apple's newly released Mac Studio also specifically highlights clustering capabilities, where multiple Mac Studios can be daisy-chained to form a more powerful system for running cutting-edge models.
The timing of this new product launch is also unusual. Apple typically updates its Mac lineup in October or November each year for the holiday season, but this time it was moved up to August.
NVIDIA Has Taken Notice, Apple Scrambles to Respond
The rise of Macs in the local AI space has already caught NVIDIA's attention.
According to an insider who discussed the competitive landscape with NVIDIA executives, NVIDIA views Apple as its biggest competitor in the local AI domain.
Late last year, NVIDIA released the DGX Spark, an AI desktop computer with a design similar to the Mac mini, directly targeting this market.
On Apple's side, they face the practical problem of supply keeping up with demand.
The massive demand for memory chips from AI data centers has led to a historic industry-wide shortage, from which Apple has not been spared.
The higher-spec configurations of Mac mini and Mac Studio, most attractive to AI developers, have been out of stock for months.
Todd Dailey, former AI Products Enterprise Marketing Manager at Apple, revealed that over the past year, due to constrained Mac supply, some enterprises have started looking for alternatives, with NVIDIA's DGX Spark frequently mentioned as an option—and it's currently available.
Dailey left Apple in April this year and is now an independent AI consultant. He also revealed that the Mac's popularity in the enterprise AI market was completely accidental, not part of Apple's proactive planning. Apple does not have a dedicated engineering team for enterprise clients nor any staff focused on developer relations.
The last time Apple sold a server product was the Xserve, discontinued in 2011. The server operating system based on macOS also ceased development in 2022.
However, some have already sensed the opportunity.
Former OpenAI Compute Infrastructure employee Peter Voell founded Mount Thor, a cloud computing company based on Apple hardware, currently still in stealth mode. Its website describes the product as an "Apple hardware-based AI execution environment."
Apple is also pinning its hopes on partners like Mount Thor and webAI to push Macs deeper into the enterprise market.
Apple has finally started using Mac chips to build its own servers. However, these servers are for internal use only, powering the Private Cloud Compute service to handle AI tasks beyond the capabilities of iPhones or Macs.
Some enterprise customers have inquired whether Apple would sell access to these servers, but Apple has so far refused.
Reference links:[1]https://www.theinformation.com/articles/apple-stumbled-ai-hardware-success-mac[2]https://www.apple.com/newsroom/2026/07/apple-reports-third-quarter-results/
This article is from the WeChat public account "QbitAI," author: Focus on Cutting-edge Technology
