This spring, spare a thought for Apple’s operations, silicon, and hardware development teams. Those people will probably be trying to test heavily disguised hardware containing the first M2 processors to roll off the TSMC production line.
And they’ll be doing so while attempting to surreptitiously build and test production capacity – all in the context of a pandemic that threatens to shutter global tech supply lines.
This doesn’t sound like an easy job. On the surface, we may be looking at the next phase in Apple’s silicon development plans, in which the final M1-series Mac appears to be followed by the introduction of the M2 chip (with a 3nm Mac processor not terribly far behind and likely to appear in 2024). But behind these plans sit a host of uncertainties the company’s teams need to manage.
Take raw materials, for example.
We know there’s shortages of some of these because of the war in Ukraine. We haven’t yet assessed the impact of grain price increases in developing nations on access to hard-to-get materials such as rare earth components. We have no insight whatsoever into any challenges securing supplies of neon and other materials for use in Apple/TSMC’s chip-cutting machines. We can’t know what impact all these additional challenges combined with shutdowns in key manufacturing centers in China will have on the company’s ability to secure supply of some of the more mundane components on which its products also rely.
But we do know that meeting these challenges is part of the job Apple’s operations and related teams must achieve before new hardware can ship. We also know the company is on the cusp of a massive extension into new hardware and potentially new platforms.
What we think will happen in the next few months is that Apple will introduce its highest-end M1 series Mac Pro, followed by the first M2-powered systems in fall.
Those M2 Macs are likely to be accompanied by an M2-powered iPad Pro on a cadence in which all Macs and iPads are upgraded with new processors in the subsequent 18 to 24 months.
Apple’s teams will be working closely with TSMC. They will want to optimize the production process for the entire family of M1 series processors, while also building capacity to manufacture the future family of M2 chips.
Apple is allegedly testing nine new Macs equipped with four different M2 chips.
They will also be working on the next iteration of Apple Silicon, including the first ever mass market 3nm chips for smartphones and Macs. TSMC is expected to begin manufacturing 3nm chips in the second half of the year, according to Digitimes. The chipmaker says it expects demand for these processors will be driven by the needs of high-performance computing and smartphones.
I think they will introduce these chips in new iPads first, followed by iPhones (17?) before graduating from A to M. When Apple puts them inside Macs in or around fall 202,3 it will also become the first mass market PC maker to introduce computers equipped with 3nm processors.
I have no doubt at all that Apple will celebrate the computational advantages of these 3nm Mac chips when they ship. I anticipate they will deliver eye-wateringly significant performance gains in comparison to the M1 series processors, and will consume even less energy, perhaps 30% less.
Apple’s already proud of its performance-per-watt. But when it shares its pride, it does so in full knowledge of what’s coming next. Just watch that high-end performance catch up — and exceed — that of less-efficient chip configurations over the coming years.
These chips may carry as many as 40-cores, which implies significant upside to the M1 Ultra. For reference, the M1 has 8 CPU cores, while the M1 Ultra hosts 20. What will you do with your 80+ core Ultra? More to the point, what computational task will you be unable to do?
The thing is, there’s a lot riding on these designs.
What should be happening is that over the next three years, we get to watch Apple’s Macs glide straight to the top of any PC performance benchmark test. They’re already close to that summit, but that max performance graph Apple showed with M1 Ultra is a beginning, not an end.
The M3 Ultra (and its successors) will inevitably deliver far more performance for far less power, cementing Apple’s advantage. But the devil is in the details, and the complexity of solving all the many current challenges in supply, design, process design, and logistical support may yet delay elements of the plan.
Unfortunately, there’s another set of challenges the teams must resolve. They are working on another strategically important component: an Apple-designed 5G modem. We think the company intends on placing this inside its first devices starting in 2023.
Now, let’s swiftly review the scale of the tasks Apple’s teams face:
Apple has executed well across the last few years. It has set revenue records, introduced new products, and while supply may at times be constrained, they have eventually shipped. Apple has also openly admitted to shortages that affected some product lines, principally (with a little strategic component redeployment) the iPad.
Even before the introduction of the last couple of iPhones, there were claims Apple had problems ensuring manufacturing capacity was up to speed. Execs were unable to visit the factories to ensure production was optimized.
In the context of factory shutdowns across the supply chain — and with so much at stake as the company pushes its Mac platform to a leadership position in the PC industry — it’s hard not to wonder how the teams tasked with taking these projects forward sleep at night when every week sees a brand-new crisis.
Apple’s road map really does seem beautiful, but the path itself appears to be full of cracks.
Please follow me on Twitter, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.