- Published on
OpenAI Buys Tens of Thousands of Apple Mac mini/Studio for AI Agent Computer Training
- Authors
- Name
- DeepHorus
It is reported that OpenAI has purchased tens of thousands of Apple Mac mini and Mac Studio computers in recent months to train AI agents capable of operating computers. This large-scale procurement action was disclosed by tech media The Information on August 31. The hardware is not intended for the foundational pretraining of its core large language models but is focused on reinforcement learning tasks. In this training paradigm, AI agents need to repeatedly practice completing operations in real computer environments.
It is understood that OpenAI primarily purchased Mac products equipped with large unified memory. These devices are used to run models and training workloads related to 'computer-operating AI agents.' Apple's self-developed chips' unified memory architecture allows the CPU and GPU to efficiently access the same memory pool, providing an ideal operating environment for AI models. Through such hardware, AI agents can learn to navigate software interfaces, interact with applications, and perform various computer tasks.
This move by OpenAI has directly driven market demand for high-end Mac products. Currently, Mac minis and Mac Studios equipped with large memory have become supply-tight, with some models even out of stock for several consecutive months. This phenomenon is not only due to OpenAI but also reflects the surging demand across the entire AI industry for hardware capable of running large memory workloads locally. AI companies particularly favor such configurations because they provide an efficient and controllable environment for model training.
Apple has not disclosed specific sales figures for individual product models, but its overall Mac business performance is strong. In the quarter ending in June this year, the Mac business achieved a record revenue of $9.4 billion, equivalent to approximately 63.316 billion yuan, a year-on-year increase of about 29%. Apple CEO Tim Cook confirmed on the July earnings conference call that AI demand is a key factor driving this growth, noting that Mac products faced supply constraints during the quarter due to 'very strong' market demand.
To meet market demand, Apple has also updated its product lineup recently. The company launched new Mac mini models featuring the M6 and M5 Pro chips, and released Mac Studio models equipped with the M5 Max and M5 Ultra chips. These new models are expected to further enhance Apple hardware's competitiveness in the AI computing field. Market analysis suggests that Apple, leveraging the energy efficiency advantages of its chip architecture, is establishing a unique position in the AI hardware market.
OpenAI is not the only AI company adopting such a strategy. It is reported that another leading AI laboratory, Anthropic, has also been renting Mac mini computing power through Amazon Web Services. As AI applications expand from the cloud to the edge, demand for local high-performance computing hardware is growing. Apple's self-developed chips' large unified memory architecture is making it a hardware choice attracting significant attention in the AI industry.
This trend marks a diversification in AI training infrastructure. Traditionally, AI training has been extremely dependent on large GPU clusters. OpenAI's case demonstrates that for specific tasks, such as reinforcement learning and agent operations, consumer-grade high-end hardware can also play a critical role. This may prompt more AI developers to evaluate and adopt different hardware combinations to optimize their development processes.
Looking ahead, the integration of AI and specialized hardware will become even closer. Large-scale procurement actions by companies like OpenAI not only shape the supply chain but may also, in turn, influence hardware manufacturers' product design directions. As AI agents move from laboratories to broader applications, demand for hardware platforms capable of running and learning in real physical environments will continue to grow, driving innovation and adaptation across the entire technology ecosystem.
