Apple Refreshes Mac Mini and Mac Studio as Memory Costs Bite

Apple is upgrading its Mac mini and Mac Studio desktops as demand grows for machines that can run AI applications locally, with memory and silicon shortages pushing prices higher, Bloomberg's Mark Gurman said on 'Bloomberg Tech' on August 25, 2026.

Apple Inc. (AAPL) is refreshing the two desktops that sit furthest from the company's consumer marketing — the Mac mini and the Mac Studio — and the reason has less to do with spreadsheets and video editing than with where AI workloads are starting to run. Demand is growing for desktop computers capable of executing AI applications locally, on the machine itself, rather than sending every request to a rented server. That is the market these two boxes now address.

Bloomberg's Mark Gurman laid out the refresh and its pricing consequences on Bloomberg Technology, speaking with Ed Ludlow. The detail that matters commercially is not the silicon generation but the cost of feeding it: memory and silicon shortages are contributing to higher prices on the new machines.

Why a desktop matters again in an era of rented compute

For most of the past decade the direction of travel was one way. Computing moved off the desk and into data centers, and the personal computer became a window onto someone else's hardware. Local AI inverts part of that logic. Running a model on your own machine means no per-token bill, no queue, no data leaving the building, and no dependency on a provider's rate limits or terms of service. For developers, researchers, video and audio professionals, and any business with privacy or compliance constraints on where data can travel, that is a real product category rather than a talking point.

The Mac mini and Mac Studio are the natural vehicles. Neither is a laptop, so neither is constrained by battery life or thermal budget in the way a MacBook is. Both can be specified with the large pools of unified memory that determine how big a model will fit on the machine at all. In practice, memory capacity — not raw processing speed — is the binding constraint on local AI work, which is precisely why the component cycle is now an Apple pricing story.

The component squeeze lands on the price tag

Memory and silicon shortages are pushing prices up, according to Gurman's breakdown. That is worth unpacking, because the mechanism is not subtle. High-bandwidth and high-density memory is being absorbed in volume by AI server buildouts, and advanced foundry capacity is booked. Every consumer and prosumer device that needs the same parts competes with that demand. Apple designs its own silicon and buys memory in enormous quantities, which normally gives it a cost advantage over smaller PC makers; it does not exempt it from a market where the marginal buyer is a hyperscaler with a data center to fill.

Apple has two levers. It can absorb the cost and let gross margin take the hit, or it can pass it through and let unit demand take the hit. The reported outcome — higher prices — indicates a company that believes this particular buyer will pay. That is a defensible read. Someone specifying a desktop to run models locally is comparing the purchase not against a cheaper laptop but against a recurring cloud bill, and against the cost of a discrete GPU rig with a far worse memory ceiling. Price elasticity is lower at that end of the market than it is for a consumer buying a first computer.

The risk is that it does not stay contained. If memory costs remain elevated, the pressure eventually reaches the products where price does move volume. Investors watching Apple's hardware margin should treat desktop pricing as an early tell rather than a niche curiosity.

What the tape says on the day

Markets did not treat the Mac news as a swing factor. AAPL last traded at 309.22, down 0.36% on the day from a prior close of 310.34, having ranged between 308.39 and 313.59, as of 18:45 GMT on August 25, 2026. That was a mild fade against a firmer market: the S&P 500 tracker SPY sat at $765.60, up 0.28%, the Dow proxy DIA at $535.17, up 0.28%, and the Nasdaq 100 fund QQQ at $710.34, up 0.57%.

The gap between a rising Nasdaq 100 and a slightly lower Apple is the shape the AI trade has taken for a while now. Money has flowed to the companies selling shovels — silicon, memory, networking, capacity — rather than to the companies buying them. Apple sits awkwardly in that framing: it makes its own chips, sells devices rather than compute, and is a buyer of the very components whose scarcity is now lifting its costs. A desktop line that turns local inference into a hardware upgrade cycle is one way to move onto the selling side of that trade.

Mac revenue is small, but the mix is changing

The Mac has long been the quiet segment of Apple's business — smaller than the iPhone, less discussed than services. Local AI does not change its scale overnight, but it changes what a Mac buyer is buying. A machine specified for model work carries far more memory and storage than a general-purpose desktop, and memory and storage are the highest-margin upgrades in the configurator. Even flat unit volumes with a richer mix lift revenue per machine. That is the arithmetic to watch in the next reporting cycle, and it is why the pricing decision on these two products is more informative than the segment's size suggests.

It also gives the Mac a replacement argument it has lacked. Desktops in professional settings get held for years. A workload that did not exist when the current machine was bought, and that is gated on memory the current machine does not have, is exactly the sort of trigger that shortens a refresh cycle.

Oura's IPO puts a price on the wearable that is not a watch

Gurman also discussed Oura, the smart-ring maker, which plans to raise as much as $3 billion in an initial public offering that could value the company at more than $16 billion. Both figures are large for a single-category hardware business, and they say something about how public investors are now pricing health data.

The ring format has succeeded by avoiding a fight it could not win. Rather than compete with the smartwatch on notifications, apps and screens, it competes on continuous overnight measurement — sleep, recovery, temperature trends — from a device the wearer forgets is there. That positions it as a health-monitoring subscription business with a hardware entry point, which is a more attractive multiple story than a gadget business.

A raise of that size would also make Oura one of the more consequential consumer-hardware listings in a market that has been dominated by AI infrastructure names. What it prices, and how it trades afterwards, will tell wearable makers and their private backers whether public capital is available on those terms.

What to watch next

Three things. First, whether the higher desktop pricing holds or gets discounted, which will indicate how real local-AI demand is beyond the enthusiast tier. Second, whether memory cost inflation shows up in Apple's other hardware lines. Third, the terms Oura ultimately sets against the $3 billion raise and the $16 billion-plus valuation now attached to it.

Frequently asked questions

What is Apple changing about the Mac mini and Mac Studio?

Apple is upgrading both desktop lines as demand grows for computers able to run AI applications locally on the machine rather than in the cloud. Bloomberg's Mark Gurman, speaking on 'Bloomberg Tech' on August 25, 2026, said memory and silicon shortages are contributing to higher prices on the refreshed models.

Why does running AI locally require a desktop?

Desktops are not limited by battery life or the thermal constraints of a laptop, and they can be configured with far larger pools of unified memory. Memory capacity, more than processing speed, determines whether a given AI model fits on a machine at all, which makes high-memory desktops the practical choice for local inference work.

How did Apple shares trade on the day of the news?

AAPL last traded at 309.22, down 0.36% from a prior close of 310.34, with a day range of 308.39 to 313.59 as of 18:45 GMT on August 25, 2026. That was a modest fade against a firmer tape, with the Nasdaq 100 fund QQQ up 0.57% at $710.34 the same afternoon.

Why are memory shortages pushing computer prices up?

High-density memory and advanced chip capacity are being absorbed in volume by AI data center construction. Consumer and professional devices compete for the same parts, so scarcity raises input costs. Apple designs its own silicon and buys memory at scale, but that does not insulate it from a market where hyperscalers are the marginal buyers.

How big is Oura's planned IPO?

Oura, which makes smart rings, plans to raise as much as $3 billion in an initial public offering that could value the company at more than $16 billion, according to Bloomberg's Mark Gurman. That would make it one of the larger consumer-hardware listings in a market otherwise dominated by AI infrastructure names.

Does the Mac matter much to Apple's revenue?

The Mac has long been a smaller segment than the iPhone or services. Local AI does not change that scale immediately, but it changes the mix: machines specified for model work carry more memory and storage, the highest-margin configurator upgrades. Flat units with a richer mix still lift revenue per machine.

Sources

• Apple Upgrades Macs for the AI Era — Bloomberg Technology

Photo: Dzenina Lukac · Pexels Licence — source