In the big tech industry, even if large-scale initiatives don’t work out, they can still pay big dividends in the future. It may be the case that apple And that “”apple intelligence“This is a strategy.” AI technology waveis often regarded as . AI lag Among them, we are friends. I claimed otherwise in lots of works Here at AI-RTZ.
Now, an Apple chip roadmap has emerged that fleshes out this argument that Apple has some chip technology that could have unintended benefits.Local AI processing‘. Today I would like to unpack it.
This appears to be the intended result of Apple’s multi-year, multi-billion dollar development. “Level 5” self-driving car own vs. Tesla etc. Ann abandoned effort for Longer timeline than expected towards that goal. paper We have had a long discussion Over 3 years ago.
But the effort creates an unexpected roadmap for AI chips that could lead to bigger changes across Apple’s platforms than expected.
in A new roadmap from AI to AI agents, Inference, reinforcement learning, training and inference loops, and more. above local device.
Bloomberg covers this AI chip roadmap in the following article: “Apple’s M6, M7, and M8 chips show how AI is reshaping the company.”:
“Apple’s new Mac chip roadmap This represents the company’s latest move to reposition its business for a world of artificial intelligence. ”
“When Apple Inc. Self-driving car project canceled In 2024, the move looked like it would undo a decade of work. This effort involved thousands of employees, the creation of hundreds of patents, specialized facilities, and spending of more than $10 billion. ”
“But even if the Apple car never came to market, and probably never will, the effort had an important legacy. It helped lay the technical foundation for the company’s latest artificial intelligence hardware strategy. Without that push, Apple would likely be even further behind in AI than it is now. ”
Just in time for Apple’s new management to take the lead. Next CEO John Tarnas In September of this year, and Apple’s longtime head of Apple Silicon, Johnny Sluge.
The twists and turns in that journey are certainly worth noting.
“Early on, Automotive program historyApple decided it needed to stand out from its rivals by having all of its vehicles be fully autonomous, with Level 5 capabilities, the highest level of self-driving cars. The move necessitated one of the company’s first large-scale AI initiatives, requiring breakthroughs in machine learning and custom silicon chips. ”
“Apple needed processors that could handle massive AI workloads in real time, and while they never finished chips designed for their vehicles, their work on these components ultimately became the foundation for technology far beyond vehicles. Most notable is their work.” created the neural enginea dedicated part of the Apple chip responsible for on-device AI processing. ”
And these new chips appeared in AI machine learning-related applications almost a decade ago. In 2017, AI researchers google I came up with something extremely important.All you need is attention“AI Paper” connected to transformers supply power to AI technology wave today.
“That technology was introduced with the iPhone X in 2017 and helped power features like Face ID and Animoji before expanding across the product lineup. Since the introduction of Apple’s in-house Mac processors in 2020, every new computer has had Neural Engine. This has made the Mac one of the most powerful platforms for running AI locally, for both consumers and (increasingly) professionals.”
Apple wasn’t as far behind AI hardware as people think even today.
“The same efforts influence the company’s powerful ultra-class Mac chips and the custom processors that currently run Apple Intelligence servers. Today, AI is largely synonymous with generative models and chatbots, but Apple has spent more than a decade building the hardware foundation needed to power the technology.”
“That doesn’t mean Apple isn’t like that.” I had a hard time To create compelling software and services that can take advantage of that hardware. The company has certainly stumbled on this front, with upgrades to Apple Intelligence and the Siri digital assistant taking much longer than expected. ”
“But as AI applications expand, more processing takes place on devices, and Apple works to reimagine its software strategy with its new Siri AI assistant, those early hardware investments are starting to pay off.”
:Since then M1 chip launched in 2020Apple has been steadily improving the Neural Engine, including significant improvements in M4. These advances are now beginning to reshape Apple’s silicon roadmap itself. ”
Bloomberg will continue to provide updates on the current M-series chip roadmap.
“As I recently reported, Apple It will be completely overhauled in the next few years. Overview of chip plans for Mac. The next cycle should start this fall with the familiar pattern of base-level M6 chips. The traditional strategy would be for this processor to be followed by the M6 Pro and M6 Max, and eventually the M6 Ultra in the company’s top-of-the-line desktop computers. ”
But there’s a twist to this.
“But for the first time, the company is skipping high-end chips for a new generation. Rather than completing the M6 family with Pro, Max, and Ultra versions, Apple is moving directly to the M7. In fact, just six months after starting the M6 process, Apple started development on the M7 (the stage of finalizing the chip design). That means the M7 should arrive in the first half of 2027, followed by the M7 Pro and M7 Max.” The M7 Ultra is expected to arrive at the end of 2027 and then in 2028. ”
“Apple has omitted Ultra chips in the past, such as with the M4 generation, but abandoning all high-end versions of a new lineup is unprecedented.”
“The reason for the break with tradition is AI. Apple had planned significant neural processing upgrades for the M7 family, but ultimately decided those improvements were important enough to justify accelerating the next generation rather than completing the M6 lineup. These changes come into full swing with the M7 Ultra. We’ve heard that this processor dramatically upgrades AI performance, bringing it closer to the class of dedicated AI accelerators such as Nvidia Corp.’s Blackwell.”
Even if it’s an unexpected opponent to catch up with, Nvidia leads the way Beyond Blackwell with Vera Rubin today and Feynman tomorrow. Apple’s chips are competitive on the local side, focusing on today’s AI data centers.
“This chip could eventually become the basis for a complete overhaul of its AI server strategy going forward. Apple will soon roll out a more powerful server based on the M5 Ultra with the internal code name J246, but engineers are already developing another new server chip built around the M7 Ultra’s capabilities for release by 2029.”
I We discussed Apple’s opportunities. Comes with a chip server side in cloud, of Eq. This is another unexpected direction for the company’s long-standing Apple silicon strategy.
“The new Ultra is designed to support up to 1.5 terabytes of memory, which is roughly double the capacity planned for the M5 Ultra, but whether Apple ultimately offers that configuration will depend on industry conditions. Widespread memory chip shortages are making this component difficult to obtain and increasing prices.”
“In addition to the M7, Apple is already working on an M8 chip with even greater AI capabilities, including the codename Soko, which will arrive by 2028. It’s also working on a new chip for high-end Macs called Cardinal. The 2028 generation of processors is moving to a 1.4-nanometer manufacturing process, which will make it even more efficient.”
“The takeaway here is that AI is no longer just a feature that Apple’s chips need to support. It’s now shaping how those products are designed and when they ship. This is a change from the days when processing speed, graphics, battery life, thin design, etc. were primary concerns.”
Again, you get windfall dividends from investments that don’t perform well.
“In retrospect, Apple’s car project probably shouldn’t be remembered as a wasted effort. Sure, the company abandoned the car, but it kept hold of a technology that could ultimately prove to be much more valuable.”
“AI hardware efforts originally intended to power self-driving cars are now shaping Macs and AI servers. In the end, one of Apple’s most expensive failures may be seen as one of its most significant technology investments.”
of whole It’s worth a read to learn more about new updates for iPad, Macbook, Apple Pencil, and more.
But the bigger point is Apple is in an advantageous position for the future Local AI era. Now again, NVIDIA, Microsoft, etc..
Therefore, it will complement the ongoing mainframe AI era. in spite of ‘“RAMagedon” and its long-term impact What I discussed.
of AI technology wave By 2030, the eighth year, it will be very different than it is today. Local AI processing using much more powerful chips is also part of that picture.
And Apple has a competitive chip roadmap and an important role to play going forward. with Nvidia, Google, And others. stay tuned.
(Note: The discussion here is for informational purposes only and is not intended as investment advice. join us here)
