New leaks coming from Chinese social media have claimed that Apple is planning to start development of its own dedicated AI chips in the near future – but it’s not the hotly-anticipated M4 chip that I’m talking about here.
Although Apple has been making waves in the AI space recently with its upgraded Neural Engine (a dedicated neural processing unit for handling AI-related tasks on Apple devices) as seen in its powerful new M3 chip, this leak makes specific reference to server AI processors – in other words, chips to power datacenters that run cloud-based AI tools. Popular online chatbot ChatGPT, for example, runs the bulk of its operations in the cloud rather than directly on your device, which is why it requires an internet connection to use.
Apple looks to be hedging its bets when it comes to AI – investing both in cloud AI technology and on-device machine learning capabilities, with the M4 chip promising to bring the entire Mac range up to speed in today’s world of ‘AI PCs’. But what does this actually mean for consumers?
AI for the Apple guy
Well, the current rumor (which originates from well-known Apple leaker ‘Phone Chip Expert’ on the Chinese site Weibo) states that Apple is working with chipmaker TSMC to develop the AI chip on a new 3nm process, but that production isn’t likely to start until the latter half of 2025 at the earliest. Basically, we shouldn’t expect to see this making a huge impact straight away.
The upcoming range of new M4 Mac products likely won’t be affected by this decision, with Apple still aiming to remain competitive with Intel and Qualcomm’s AI PC efforts.
However, Apple users of all sorts could stand to benefit from the company’s new interest in cloud-based AI – with its own powered-up servers for offloading AI workloads on iPhones, iPads, and Macs combined with more powerful on-device AI capabilities, Apple could be poised to become a market dominator offering best-in-class AI services to everyday users.
You might be surprised just how much AI there already is in your iPhone 15 or MacBook Air. Apple’s Neural Engine tech has been lurking in its phones since the A11 Bionic chip seen in the iPhone 8, powering staple iOS features such as Face ID and Animoji. As AI-powered software becomes more common – it’s already wormed its way deep into Adobe Photoshop, for example – the need for competitive hardware in both consumer devices and data centers is on the rise.
As always with leaks such as this, it’s wise to take it with a pinch of salt – while Phone Chip Expert is a relatively reputable leaker, that doesn’t instantly guarantee that this information is legit.
Still, I reckon it’s at least somewhat accurate; while a development like this will no doubt cause further struggles for TSMC’s already-burdened manufacturing and supply lines, the fact is that local on-chip AI isn’t yet powerful enough to properly handle high-level large language models – so investing in its own AI servers is the perfect way to deliver the best possible AI experience to users.
