Apple M7 Ultra is an Apple-announced server-class SoC, planned for release in the first half of 2027. According to Mark Gurman's Power On newsletter (Bloomberg, July 12, 2026), the chip will support up to 1.5 TB of RAM — a huge jump over current M-series capabilities (M3 Ultra: 192 GB, M4 Max: 128 GB) — and will serve as the foundation of a new Apple server product.
According to the same report, Apple is skipping the M6 Pro, M6 Max and M6 Ultra variants — the company's engineering resources are being redirected straight to M7. This shortens the lifecycle of the current generation for professional Mac Pro and Mac Studio users, but lets Apple accelerate the server chip's debut.
The architecture is descended from the discontinued Project Titan — Apple's self-driving car program, officially closed in 2024 after more than a decade of work. Ruslan Salakhutdinov (hired from Carnegie Mellon in 2017) designed a chip capable of processing sensor data from the car in real time, on-device. That work shipped in the iPhone X in 2017 as the Neural Engine in A11 Bionic and has been the foundation of Apple's on-device AI strategy ever since.
M7 Ultra is set to power Apple's own AI infrastructure (including Apple's US data centers announced in March 2026 to serve Apple Intelligence) and — potentially — be offered to enterprise customers. The strategy differs from competitors: Google TPU, Meta MTIA and OpenAI/Broadcom Jalapeño target large inference clouds, whereas Apple relies on unified CPU+GPU+Neural Engine memory — an architecture without separate VRAM that eliminates the data-transfer bottleneck between processor and accelerator.
The implication for robotics and on-device AI: a 1.5 TB RAM limit in a single SoC package would let much larger language and vision models run locally, without dependence on external servers. This is consistent with Apple's long-term strategy where user privacy (data does not leave the device) is a product advantage backed by real silicon compute power.
AI System-on-Chip (SoC) · serves as: AI acceleration, AI Inference, Compute, High-level compute.
Which group Apple M7 Ultra belongs to and how it is built
Covers integrated System-on-Chip circuits used as central processing units in embedded systems, robotics, AIoT, and edge AI devices. In this schema, parentCategory is set to 'other' because the parentCategory enum does not include a compute category.
Component type encompassing System-on-Chip designs used as primary compute units in embedded devices, edge AI, and robotics applications.
A single semiconductor component integrating CPU, GPU, AI accelerators, memory controllers, multimedia engines, and communication interfaces.
