Apple appears ready to rewrite part of its Mac playbook by pushing its future AI ambitions ahead of its usual chip-release cadence.
According to Bloomberg’s Mark Gurman, Apple is preparing a significant shift in its in-house silicon strategy by skipping high-end M6 variants and accelerating development of an AI-centered M7 generation. Instead of releasing the usual Pro and Max versions alongside the next chip family, Apple is reportedly planning to launch only a standard M6 processor for entry-level Macs before moving its premium lineup to M7 processors.
The move would break a pattern that has existed since Apple introduced its custom silicon lineup. Every generation from M1 through M5 included multiple performance tiers, typically a base chip, followed by Pro and Max versions and, in some cases, an Ultra model.
What Apple is reportedly building
Bloomberg reports that the standard M6 processor could arrive as early as this year in entry-level Mac products, including a refreshed MacBook Pro. Internally, the chip is said to include upgrades intended to strengthen AI performance, graphics processing, and memory capabilities.
The M6 reportedly targets around 200GB/s of memory bandwidth, up from roughly 153GB/s on the current M5 architecture. It’s also expected to feature an updated Neural Engine, broader performance gains across processor cores, and a redesigned graphics setup with up to 12 GPU cores.
The bigger change, however, appears to be what comes next. Bloomberg reports that Apple plans to release the M7 lineup beginning in 2027, with M7 Pro and M7 Max variants expected later that year, and an M7 Ultra potentially following in 2028. The base M7 is reportedly designed around stronger on-device AI processing and could reach approximately 240 GB/s of memory bandwidth.
Why this matters
The reported strategy suggests Apple sees AI workloads becoming central to how future computers are used.
Large language models, image generation tools, video processing, and AI-assisted workflows require systems capable of moving large amounts of data rapidly. Higher memory bandwidth and more powerful AI engines could help Macs process more of these tasks directly on the device rather than relying on cloud services. That shift could also give Apple another way to separate itself from competitors that rely heavily on external chip suppliers.
Consumer and market fallout
For consumers, this creates immediate friction and a difficult buying dilemma.
High-end users looking to upgrade their MacBook Pros, Mac minis, or Mac Studios face a massive dead zone. With M7 Pro and Max hardware pushed out to late 2027, power users must choose between buying older M5 tech or waiting more than a year for the next true performance leap.
There is a brief silver lining for absolute top-tier buyers: Apple still intends to launch an M5 Ultra chip (Sotra D) as early as this year for a delayed Mac Studio refresh, pushing specifications to a massive 36 CPU cores and 80 GPU cores.
However, supply chain component constraints have already forced Apple to limit new Mac Studio orders to just 96 gigabytes of memory, down significantly from the 512 gigabytes supported during the M3 era in 2025, making the upcoming M5 Ultra release complex and highly constrained.
Additionally, this timeline disruption throws other highly anticipated hardware roadmaps into complete uncertainty. Rumors had pointed to a late 2026 release for the high-end MacBook Ultra featuring a touchscreen and an OLED display. With the pro-tier silicon pipeline significantly delayed, the launch window for Apple’s first touchscreen laptop is now entirely unclear.
Coupled with Apple’s June 25 price increases across its entire Mac and iPad lineup, consumers are being asked to pay more for hardware that faces a prolonged waiting period before its next major architectural leap.
Related reading: If you’re also weighing an iPhone upgrade, check out our roundup of everything we know so far about the rumored iPhone 18 Pro, including its expected AI features, design changes, and release timeline.
Read the full article here
