Apple’s latest move could change how you use AI every day

For years, the artificial-intelligence explosion has been largely invisible to customers, taking place within vast data centers crammed with pricey chips.

Apple wants to bring more of it to your desk.

Apple (AAPL) on Aug. 25 introduced its M6 and M5 Ultra CPUs and new Mac mini and Mac Studio models, bringing a lot more computing power for AI that can operate right on a person’s computer.

The M6 is Apple’s first 2-nanometer Mac CPU, while the M5 Ultra is the company’s first quad-die M-series processor.

But there’s another number shoppers could notice first.

The new M6 Mac mini starts at $899, $100 more than its immediate predecessor, according to Reuters. However, Tom’s Hardware notes that this increase comes after past price spikes for Apple computers, and it will cost an additional $500 to upgrade the entry-level system from 256GB to 1TB of storage.

This is the tension at the center of Apple’s current hardware launch.

Consumer PCs are increasingly capable of executing complex AI work without constantly going into the cloud. They’re also being asked to pay extra for them.

“Today, we’re debuting the next giant leap in performance and AI compute for Apple silicon,” Sri Santhanam, Apple’s vice president of Silicon Engineering Group, said in announcing the chips.

Whether buyers believe that exchange is worth it could help determine if AI is another reason to update a computer or another cause to cling to the one already on the desk.

Apple is moving more AI from the cloud onto the Mac

The key thing about M6 isn’t just that it’s faster than M5.

Apple has a distinct approach to AI.

The M6 packs a 12-core CPU, 12-core GPU with Neural Accelerators, and a dual 16-core Neural Engine. Apple says the Neural Engine delivers up to 2x peak computing of prior generations, and peak GPU compute for AI is approximately 30% higher than M5. Those numbers are Apple’s internal benchmarks and have not yet been independently verified.

But there’s a more relevant trend underlying those numbers. Mac mini and Mac Studio computers have already proven popular for local AI inference and software development, aided by Apple’s unified-memory architecture and combination of powerful CPUs and GPUs, Ars Technica noted.

This is where the story expands beyond just chip hobbyists.

When you use many of today’s generative-AI services, a remote data center processes your command on computers owned by another company.

Local AI is able to shift that dynamic.

You can theoretically evaluate documents without leaving your computer. Local testing of models can be performed by developers. Creators can do AI-assisted work on their equipment. Businesses that process sensitive information may have greater control over their processing.

And certain jobs don’t even need an internet connection to connect to a remote AI server.

That doesn’t imply the cloud will go away. Large frontier AI models require vast amounts of computing power, and many popular AI services will continue to depend significantly on data centers.

Rather, Apple seems to be wagering that the line between cloud AI and personal AI will shift.

The Mac increasingly straddles the personal side of that line.

Apple just made the Mac a much more serious AI machine.

Bloomberg / Getty Images

Apple’s $899 Mac mini reveals the consumer catch

There is, however, a cost to adding more capability to the box.

The M6 Mac mini starts at $899 with 16GB of memory and 256GB of storage. This is $300 above the $599 original launch price of the M4 Mac mini, although Apple had already repriced that older machine to $799 in June, MacRumors noted.

And storage can greatly affect the final bill.

Taking the base Mac mini from 256GB to 1TB adds $500, Tom’s Hardware reports, pushing the machine to $1,399 before a user adds a monitor, keyboard, or mouse.

More Apple:

That makes Apple’s AI pitch much more concrete for everyday shoppers.

You don’t even need to know about neural accelerators to know $899. A family doesn’t need to understand memory bandwidth to understand another $500 for storage. And a small business owner doesn’t care how little the transistors are if the computer doesn’t save enough time or money to warrant replacing the ones they have.

This is where Apple’s technological achievement meets the realities of the consumer wallet.

What Mac buyers are actually getting

The M6 Mac mini starts at $899, while the M5 Pro version starts at $1,699, according to MacRumors. The professional Mac Studio starts at $2,499 with M5 Max and jumps to $5,499 with M5 Ultra.

Apple says the M6 in the new Mac mini delivers up to 4x faster AI performance, with storage and graphics performance up to twice as fast. Again, those comparisons are based on Apple’s testing.

The decision for buyers therefore becomes less about specifications than about economics: Does having more AI power inside the computer justify paying more for the computer itself?

Apple’s M5 Ultra pushes local AI into a different league

At the other end of Apple’s lineup is a gadget that few consumers would ever buy. Still, its capabilities go a long way toward revealing where Apple thinks personal computing is going.

You may configure the M5 Ultra with up to a 36-core CPU and an 80-core GPU. More importantly for AI, it provides up to 512GB of unified memory and 1.2TB per second of memory bandwidth, MacRumors said.

That’s enough RAM, Apple argues, to run massive language models with hundreds of billions of parameters completely on-device.

But the 512GB version will not be accessible right now. It is slated for late October and is believed to cost well beyond $10,000, but Apple has not yet published the final price for that configuration, MacRumors noted.

Obviously, this isn’t a Main Street computer.

But it may be of interest to companies, developers, and researchers who currently rent large amounts of processing capacity from cloud providers.

A powerful workstation that can run very big models locally presents a separate calculation: whether to invest more in hardware upfront to eliminate part of the recurrent cloud-computing expenditures later.

That’s a lot more fascinating concept than just making Photoshop start faster.

Why Apple’s AI push matters for AAPL stock

Apple’s advantage in this fight is not necessarily that it has the world’s most powerful AI model.

It is that Apple controls almost everything surrounding the model. It designs the computer. It designs the operating system. It designs the silicon.

And it provides frameworks, including Core AI, Core ML, and Metal, which developers may use to leverage that hardware for AI workloads.

That integration gives Apple an opportunity to make AI a factor in customers’ willingness to pay a premium for its hardware.

Reuters called the new Mac mini designed for consumers who want to host AI agents, and it retails for $899, $100 more than the previous version.

That’s the tension to watch, for investors.

Apple isn’t just attempting to sell customers on the higher specs of its Macs. It has to persuade them that a more costly Mac is worth it for the AI.

If local AI proves helpful enough, the plan may help Apple maintain its premium pricing, spark Mac upgrades, and enhance the ecosystem advantage that’s traditionally made migrating away from Apple tough.

If consumers can’t discover enough common uses for that processing power, then a faster Neural Engine is just another spec that most buyers won’t really notice.

And the rising price is becoming much harder to ignore.

The AI boom is reaching the household budget

There’s a bigger reason this launch matters.

The physical cost of AI is growing.

Data centers want GPUs. Those GPUs require memory. Servers need storage. This requires massive amounts of hardware, and the competition for components is rapidly impacting things bought by ordinary consumers.

Apple’s new Mac price is a particularly good illustration of collision. Tom’s Hardware said Apple’s recent computer price hikes are a direct result of a climate where demand for AI infrastructure has strained the technology supply chain.

So the AI revolution is not only something for Nvidia shareholders or Silicon Valley corporations to sink billions into data centers.

More often it shows up when someone is replacing the household PC.

Apple’s answer is to give buyers something in return: a whole lot more AI processing that happens on the system they actually own.

That raises one of the more significant problems for the future generation of personal computers. Consumers are accustomed to buying subscriptions for software and cloud services.

Apple is betting they may also pay more upfront for a computer powerful enough to bring some of that intelligence home.