# Apple Announces M6 Mac mini and M5 Ultra Mac Studio Aimed at Local AI Workloads and Mac Clustering

> Source: <https://www.ghacks.net/2026/08/26/apple-announces-m6-mac-mini-and-m5-ultra-mac-studio-aimed-at-local-ai-workloads-and-mac-clustering/>
> Published: 2026-08-26 07:55:18+00:00

# Apple Announces M6 Mac mini and M5 Ultra Mac Studio Aimed at Local AI Workloads and Mac Clustering

[Apple](https://www.ghacks.net/category/companies/apple-companies/) has introduced [two new chips and updated its desktop lineup](https://www.apple.com/newsroom/2026/08/apple-introduces-m6-and-m5-ultra-for-a-big-leap-in-performance-and-ai-compute/) with the M6 Mac mini and M5 Ultra Mac Studio. Both are aimed at local AI workloads and Mac clustering.

The M6 is Apple's first chip built on a 2nm process, while the M5 Ultra is the company's first quad-die design and now its most powerful chip. There are no major new features, but both models come with higher specifications and increased prices.

Most versions will be available starting September 22, with preorders already open. The 512GB memory option is set to ship in late October. Every performance claim comes from Apple, using its typical "up to" wording against varying baselines.

## Apple's M6 and M5 Ultra Chip Upgrades

The M6 uses TSMC's 2nm process and is the first Apple chip to include all three types of CPU cores in one design. It features 12 CPU cores: two super cores, four performance cores, and six efficiency cores. While super cores first appeared in the M5 Pro and M5 Max earlier this year, the M6 is the first to combine all three core types in a single chip.

Apple claims the "world's fastest" single-threaded performance, 1.2x the multithreaded throughput of the M5, and 40% faster CPU performance than the M4. The GPU moves to 12 cores with a Neural Accelerator in each, which Apple says delivers nearly 30% more peak GPU compute for AI than the M5.

The chip includes a dual 16-core Neural Engine, storage rated at twice as fast at 15GB/s, and up to 170GB/s of memory bandwThe M5 Ultra stands out as the more notable chip. It uses Apple's UltraFusion interconnect to combine two dual-die M5 Max chips into a single quad-die package.

This design provides inter-die bandwidth above 4.4TB/s and over six times the connection density of the previous version. All four dies function as a single processor.e four dies behave as one processor.

This setup allows for up to 36 CPU cores, with 12 super cores and 24 performance cores, as well as up to 80 GPU cores and a 32-core Neural Engine. The M5 Ultra supports up to 512GB of unified memory at 1.2TB/s. Apple claims it delivers 4.5 times the peak GPU compute for AI compared to the M3 Ultra and 40 percent faster graphics.

The less expensive Mac mini uses the latest 2nm process, while the flagship M5 Ultra remains on the 3nm process used by the M5. This is typical for Ultra chips, which are built from Max dies and usually lag a generation.

As a result, the M5 Ultra name does not fully reflect the performance jump it brings, especially for AI workloads. Apple is not planning an M6 Pro or M6 Max, and will move directly to the M7 in mid-2027.

## Apple's Local AI Push, Mac Clustering, and Higher Prices

The machines are aimed less at creators and more at local AI. Apple shipped macOS 26.2 last December with support for low-latency communication between Thunderbolt 5 hosts, specifically for distributed inference using MLX.

Since then, hobbyists and developers have been daisy-chaining Mac minis and Studios to run models far larger than a single consumer machine can hold, as an alternative to buying racks. Mac Studio now officially supports clustering over Thunderbolt 5 with RDMA, allowing memory to be pooled across multiple systems.

According to Apple, four clustered Studios can deliver up to three times the inference throughput of a single machine. There is also a new Core AI framework in addition to MLX. Apple positions the Studio as its top desktop for on-device AI, while the mini is described as an always-on agentic machine.s-on agentic machine.

The economics are simple. Developers using cloud-based coding agents are seeing costs rise, while open-weight models such as Qwen and DeepSeek can do much of the same work on local machines.

However, a typical consumer Mac still cannot run models large enough to fully replace advanced APIs, which is why some users have started chaining Macs together. The Mac mini makes up about 3 percent of Mac unit sales in the US, according to Consumer Intelligence Research.

The M6 mini starts at $899 and the M5 Pro mini at $1,699, each $100 above the current M4 and M4 Pro models. Mac Studio starts at $2,499 with the M5 Max, the same as the current M4 Max model, and $5,499 with the M5 Ultra.

While these comparisons seem to favor Apple, the baseline price has already shifted. The M4 mini launched at $599 in October 2024, but Apple discontinued that version in May, raising the entry price to $799. In late June, Mac prices increased by about 15 to 20 percent, with Tim Cook citing higher component costs as the reason.

The lowest-priced Apple desktop is now 50 percent more expensive than it was two months ago, and it still comes with 16GB of memory and a 256GB SSD.ps with 16GB of memory and a 256GB SSD.

Upgrades are steep. Another 8GB costs $200, hitting the M6's 32GB ceiling costs $4These price increases are part of a broader industry trend, not unique to Apple.

## Memory Supply Pressures Behind Apple's Higher Mac Prices

DRAM prices rose by about 90 percent quarter over quarter in the first quarter of 2026, and by more than 50 percent again in the second quarter. IDC calls this a structural reallocation of wafer capacity, since wafers used for HBM in AI cannot be used for standard laptop RAM.

[Microsoft](https://www.ghacks.net/category/companies/microsoft-companies/), Asus, MSI, Nintendo, Sony, and Valve have all raised hardware prices this year for similar reasons.us, MSI, Nintendo, Sony, and Valve have all raised hardware prices this year.

Demand for the previous Mac mini and Studio models already exceeded supply, with some versions taking weeks or months to ship. Earlier in 2026, Apple removed its largest memory options and has had trouble keeping some MacBook models in stock.

The higher memory tiers are now available again on the new mini and Studio, up to 512GB, but that option will not ship until late October. Apple recently highlighted a new Mac mini assembly line in Houston, set to begin production later this year.

The company is promoting higher upfront hardware costs as an alternative to ongoing AI service fees, but this only matters if the hardware is actually available.

## Mac Availability and Apple's Leadership Transition

Tim Cook will step down as CEO on August 31, with John Ternus taking over. This makes the new Macs the last products released under Cook's leadership. The upcoming iPhone event in September will be Ternus's first as CEO, where he is expected to introduce what may be Apple's first foldable device and possibly a touchscreen MacBook.

Most M6 mini and M5 Ultra Mac Studio models will be available on September 22, with preorders already open. The 512GB memory version will not ship until late October.

It does not review Macs or AI hardware, so all performance numbers come from Apple and are based on different baselines, including the M1, M4, and M3 Ultra.

The main question is whether Apple can keep these new machines in stock, given the ongoing memory supply issues and strong demand seen with the previous generation.
