Apple’s smallest desktop has become an unexpected strain on Cupertino’s hardware lineup. Higher-memory versions of the Mac Mini and Mac Studio are growing harder to find as demand from developers and power users outpaces a supply strategy designed for what was once a niche product. A device long considered a minor part of Apple’s portfolio is now closely tied to the rise of local AI workloads.
According to Consumer Intelligence Research Partners, the Mac Mini represented approximately 3% of Apple’s US Mac unit sales last year. That share appears to be shifting rapidly. In recent months, the compact desktop has become a preferred platform for running persistent local AI agents.
At the center of that shift is memory capacity. Running LLMs locally can require tens of gigabytes of RAM, especially for customized or always-on agents. The Mac Mini, Apple’s compact desktop built on Apple Silicon, has emerged as a relatively accessible option for handling those workloads. Keeping inference on-device also avoids usage caps and token-based pricing, encouraging more developers to adopt local setups.
The pressure is now showing in Apple’s online store. Higher-memory configurations, including M4 Mac Mini models with 32GB of RAM and M4 Pro versions with 64GB, are currently unavailable on Apple’s website. Other models list delivery estimates ranging from several weeks to as long as 12 weeks. A similar trend is affecting the Mac Studio, where premium configurations are no longer immediately available.
Notably, the MacBook Pro remains widely available, including configurations with up to 128GB of RAM, and has shorter delivery estimates. Lower-memory systems across the Mac lineup are also shipping without significant delays, indicating the bottleneck is focused on specific high-capacity SKUs rather than the lineup as a whole.
Apple has not provided a public explanation, but the factors behind the issue are becoming clearer. One appears to be a straightforward misread of demand.
“Apple was caught up by the number of people buying Minis for Clawdbot [aka OpenClaw], which would have been impossible to predict a few months ago,” Francisco Jeronimo, vice president at IDC, told The Wall Street Journal.
That surge highlights how quickly open-source AI tools have matured. Projects that allow local agent deployment with minimal setup have lowered the barrier to entry, turning machines like the Mac Mini into lightweight infrastructure rather than standard endpoints. Similar trends have appeared across the wider PC market in recent weeks, with vendors and component suppliers seeing renewed demand for high-memory systems built for local inference. The supply limitations may also stem from Apple’s cautious inventory management strategy.
“The lead times on supply are longer than one might think,” said CIRP co-founder Michael Levin. The Mac Mini remains a niche product in Apple’s portfolio, and overproduction carries its own risks. As Levin put it, “Apple also doesn’t want demand to wane suddenly and have a year or more of inventory sitting around.”
Another consideration is Apple’s product cycle. Its desktop Mac lineup appears due for updates, particularly as M5 chips begin rolling out across other categories. Analysts note that inventory tightening often precedes hardware refreshes, though the current pattern is inconsistent.
Kieren Jessop noted that a conventional prelaunch inventory reduction would be expected to impact all configurations, not solely high-memory SKUs.
Memory supply has also emerged as a broader industry concern. AI infrastructure buildouts have driven higher global RAM demand, contributing to shortages already impacting PCs and smartphones. Apple’s vertically integrated architecture, which incorporates memory directly into its SoC designs, helps buffer the company from some of the supply pressure affecting competitors.
As Jessop noted, widespread supply issues would likely disrupt more of Apple’s lineup. Jeronimo added, “If Apple can’t get memory, no one else will.”
Maybe you would like other interesting articles?