Against the background of overall pressure on the global consumer electronics industry and supply chain shocks, Apple has been exposed to a series of "compromise" adjustments in the planning of multiple product lines, including the upcoming touch OLED "MacBook Ultra" which will still be equipped with the existing M5 Pro and M5 Max chips, and the basic iPhone 18 is expected to be equipped with only 9GB of running memory, making it difficult to fully support the new generation of local AI functions.
It has been previously reported that Apple has significantly restructured its own chip roadmap: this year it will focus on the M5 Ultra and the basic version M6, and then directly launch the basic versions M7, M7 Pro and M7 Max in 2027, and move the higher-end M7 Ultra to 2028. This adjustment also disrupted the chip rhythm originally planned for the new generation of touch-sensitive OLED MacBook Pro. This device is widely expected in the industry to be named "MacBook Ultra" to highlight its flagship positioning.
Judging from the public information, this rumored MacBook Ultra will use a hybrid OLED architecture: a combination of oxide TFT and laminated OLED design to increase brightness and improve energy efficiency. It will also be equipped with a touch screen, integrate a punch camera in a pill-shaped opening to support an interactive effect similar to "Smart Island", and continue the thin and light body design. But according to the latest news from Bloomberg reporter Mark Gurman, this high-end new product will not debut in the M6 series, but will "reuse" the existing M5 Pro and M5 Max chips. It is not expected to be replaced with the M7 Pro and M7 Max until the updated model in 2027. Meanwhile, Apple still plans to launch a refreshed entry-level MacBook Pro model later this year, powered by a new base M6 chip.
Industry analysts believe that there may be two considerations behind Apple's "misplaced combination" strategy on the touch version of MacBook Ultra. On the one hand, it is rumored that the base M7 will provide significantly higher unified memory bandwidth, about 240GB/s, while the base M5 is 153GB/s. This gives the M7 series a significant advantage in edge AI scenarios. Apple therefore hopes to complete the transition to the M7 architecture as soon as possible, even if it has to sacrifice the intermediate M6 Pro, M6 Max and M6 Ultra chips. On the other hand, TSMC's tight 2nm production capacity is also regarded as one of the important factors. It is reported that Apple plans to hand over part of the M7 to Intel's 18A-P process OEM to spread the pressure on TSMC's production capacity by accelerating the release time of the M7 series.
In terms of mobile product lines, well-known Apple analyst Ming-Chi Kuo recently reported that the basic iPhone 18, scheduled to be launched in the first half of 2027, is expected to be equipped with only 9GB of RAM. There is a clear misalignment between this specification and the new generation of AI solutions. Apple previously introduced that the new version of Siri AI and Apple Intelligence operate based on three sets of cloud models and two sets of local models. The AFM 3 Core Advanced model has approximately 20 billion parameters and can dynamically activate 1 billion to 4 billion parameters according to tasks to unlock more advanced dictation capabilities and more natural Siri voice performance.
Although AFM 3 Core Advanced will only load 1 billion to 4 billion parameters into memory at any time, the related solution is still considered to require at least 12GB of RAM to run stably, which means that the basic iPhone 18 with only 9GB of memory is likely to be unable to fully support this higher-order local AI model. In other words, in the future, under the architecture of iOS 27 that deeply integrates Apple Intelligence, the basic iPhone 18 may form a more obvious functional layering with the Pro series in terms of local AI capabilities and voice interaction experience.
As for why Apple has widened this intergenerational gap in its flagship mobile phones, cost is considered to be the most direct reason. The report quoted industry data as saying that since the first quarter of 2025, the contract price of 12GB LPDDR5X has increased by about three times. From the end of the first quarter to the second quarter of 2026, the single price hovered around US$120, and has recently risen further to about US$145, while the cost of NAND flash memory is also rising simultaneously. Under the pressure of rising prices for both memory and storage, Apple has made a choice between maintaining profit margins and controlling terminal pricing by controlling the memory specifications of basic models, which has become a "realistic choice" in the current market environment.
As the MacBook Ultra may "use old chips with new shells" and the gap between the basic iPhone 18 and high-end models in terms of AI experience is further widened, Apple's "longer and longer list of compromises" has also triggered discussions in the industry: Under the multiple squeezes of tight supply chains, soaring costs and product competition in the AI era, whether consumers will still pay for such a product mix and price will become a key question that Apple must face in the next few years.