// 9TO5MAC — MOBILE & WEB
Apple @ Work: Next to a growing token bill, Macs might start looking like a bargain
Apple @ Work is exclusively brought to you by Mosyle, the only Apple Unified Platform. Mosyle is the only solution that integrates in a single professional grade platform all the solutions necessary to seamlessly and automatically deploy, manage, and protect Apple devices at work. Over 45,000 organizations trust Mosyle to make millions of Apple devices work ready with no effort and at an affordable cost. Request your EXTENDED TRIAL today and understand why Mosyle is everything you need to work with Apple.
Apple hardware costs more than it did a year ago, and if you have purchased a fleet recently, you are feeling it. Now, one thing I should mention: Macs have not risen in price with inflation or performance. If you think about what a Mac mini or MacBook Air cost when they launched vs. today, I would still say they are extremely affordable. With that being said, there’s a scenario taking shape in the enterprise where the price of a Mac matters a lot less than it used to. If open-weight AI models become a real part of how companies work, Apple Silicon might end up looking basically free compared to the token bill it supplements.
About Apple @ Work: Bradley Chambers has been an Apple IT admin since 2009. Through his experience deploying and managing firewalls, switches, a mobile device management system, enterprise grade WiFi, 1000s of Macs, and 1000s of iPads, Bradley will highlight ways in which Apple IT managers deploy Apple devices, build networks to support them, train users, share stories from the trenches of IT management, and ways Apple could improve its products for IT departments.
Cloud AI in the enterprise is priced by the token, and that’s a fine model when usage is moderate. The problem is that usage isn’t staying light. Agentic tools that write code, analyze documents, and run multistep tasks consume tokens at a pace a chatbot never did, and the bill scales right along with them. Earlier this year, Fleet’s IT leadership research cited a company that burned through its entire 2026 AI budget in four months after rolling an AI coding tool out to roughly 5,000 engineers. That’s the risk with a variable cost by token usage. Finance is somewhat in the dark on what the number is until the invoice shows up.
Local hardware works the opposite way, where you pay once, and the cost is capped. Every hour that machine spends running a model after that costs you very little, aside from power usage.
Someone I’ve long respected, Horace Dediu, showed some interesting numbers in a recent Asymco piece that really shows the opportunity ahead for Apple. In one example, a 26-billion-parameter model used for analysis four hours a day, around 150 million tokens a month, runs $450 to $850 per user per month through a cloud provider. A 16″ MacBook Pro with an M5 Max and 128GB of unified memory costs about $6,500 and can run a model locally. That’s a payback period of 7 to 14 months. He gives a lot more examples, and it’s worth a read to see Apple’s opportunity here.
The reason Apple Silicon is so well suited to this is unified memory. On most PCs, the GPU has its own small memory pool, and the model has to fit within it. On Apple Silicon, the GPU shares the entire memory pool, so a laptop with 128GB can hold a model that would need a far more expensive setup elsewhere.
Viewed through this lens, a price increase on a Mac Studio is a rounding error. If the machine pays for itself in under a year, and Apple Silicon Macs can stay in service for four to five years, it’s practically free over a normal refresh cycle. Now, if agentic coding keeps improving at such a rapid pace, you might end up upgrading sooner than that.
All of this hinges on one assumption: the math only works if an open-weight model running locally is good enough for your task day-to-day. For a lot of everyday work, like summarizing docs, drafting, content analysis on internal data, and basic routine coding, the gap between open-weight models and the frontier has narrowed