OpenAI has purchased tens of thousands of Mac Minis and Mac Studios over the past several months, according to a report from The Information. Meanwhile, Nvidia's RTX Spark AI-PC chips have sold out to distributors ahead of their scheduled fall 2026 launch.
TL;DR
- OpenAI's acquisition of Apple's AI-ready desktops highlights the growing demand for hardware capable of running large language models locally.
- Nvidia's RTX Spark platform, aimed at AI developers, faces a supply crunch months before its launch, with major PC makers already placing follow-up orders.
What happened
OpenAI has reportedly acquired tens of thousands of Apple Mac Minis and Mac Studios to enhance its reinforcement learning and computer-use agent training efforts, according to The Information. These purchases are separate from the cloud GPU clusters OpenAI typically leases, indicating a preference for Apple's compact desktops for specific tasks.
Rival AI lab Anthropic is reportedly renting Mac Minis through AWS for similar purposes, taking a different approach to acquiring the necessary hardware.
Apple's early launch of refreshed Mac Mini and Mac Studio hardware is attributed to the AI-driven boom in sales, with new silicon variants landing just days before the OpenAI report surfaced.
Nvidia's upcoming RTX Spark platform, which pairs a Grace Arm CPU with Blackwell graphics, has seen its first wave of AI-PC chips sell out to distributors. Major PC makers like Asus, MSI, Dell, HP, Lenovo, and Microsoft have already burned through their initial allocations and are pushing for additional supply.
Why it matters
The demand for hardware capable of running large language models locally is driving AI labs and PC makers to hoard any available hardware, pushing prices out of reach for ordinary gamers and PC builders.
The appeal of Apple's unified memory architecture for AI developers lies in its ability to hold far larger models in memory than a typical discrete graphics card, making it ideal for reinforcement-learning and agentic AI training.
Nvidia's RTX Spark hardware is being pitched squarely at AI developers, with specs highlighting its capability to handle 120-billion-parameter LLMs and local 4K AI video generation. This shift in focus away from gaming could impact the availability and pricing of components for PC gamers and builders.
The global chip squeeze, driven by AI demand, is likely to affect PC builders in different regions at varying times, with Australian and New Zealand builders potentially feeling the impact later than North American and Asian markets.
Key facts
- OpenAI has acquired tens of thousands of Mac Minis and Mac Studios for AI training purposes.
- Anthropic is renting Mac Minis through AWS for similar AI training tasks.
- Apple's early launch of refreshed Mac Mini and Mac Studio hardware is driven by AI demand.
- Nvidia's RTX Spark platform has sold out to distributors ahead of its scheduled fall 2026 launch.
- Major PC makers have already burned through their initial allocations of RTX Spark chips and are pushing for additional supply.
- The flagship N1x configuration of RTX Spark-based PCs is expected to be priced at around $3,400.
- RTX Spark hardware is marketed around 120-billion-parameter LLMs with context of up to one million tokens.
- The global chip squeeze, driven by AI demand, is likely to impact PC builders in different regions at varying times.
Context
The growing demand for hardware capable of running large language models locally is reshaping the compute landscape, with AI labs and PC makers prioritizing the acquisition of AI-capable silicon.
The shift in focus from gaming to AI development is evident in the marketing and specifications of new hardware, such as Nvidia's RTX Spark platform.
The global chip squeeze, driven by AI demand, highlights the need for a more sustainable and equitable distribution of computing resources, ensuring that both enterprise and consumer markets can access the hardware they need.
