Hacker News Reader

Comment by greenknight | Hacker News Reader

I would expect that Apple would push its goldenchild (iphone) onto the node first. Its a small chip which they can use as a pipecleaner, makign sure they can get the yields up and optimise the process before pushing a larger die onto the node.

They easily could have been allocating risk production on the iphone for the past couple of months, ready for the launch. Apple being like yes we will take lower yields for less cost.

I do not expect any company to announce production 3N product, until apple has had one out for atleast 6-12 months. Look how long it took the rest of the industry to move to 5N. I swear part of that reason was an exclusivity agreement with Apple, and it massively paid off for their CPUs. Having a node advantage is always massive in terms of price / performance / power matrix.

View on HN

Are you suggesting they might have produced millions of A16 chips on N3 during risk production phase and launched it before TSMC even reaches HVM? Highly unlikely. Risk production is a phase where they still make changes and fix issues. It's like a beta phase. It does not come at a lower cost, it would be more expensive to throw out a big chunk of chips. The iPhone chips are very high volume, you can't produce them before reaching... high volume manufacturing phase.

The iPhone contributes to TSMC revenue in a substantial manner so that also would totally not fit what TSMC said.

The M2 Pro/Max/Ultra are much lower volume and higher margin. It makes sense to start with them.

Except that they are much larger chips, that will be much more sensitive to yield issues. They could do that, but they will be expensive. Maybe that's ok.

Their biggest (m2 max in studio) are more like chiplets with interposers, which exponentially lowers the yield problems depending on how much you can split it up. Also larger chips can still set a threshold on gpu cores to disable due to defects and that kind of thing, where a mobile one might just throw out the chip with a much lower number of (or any?) defects.

Nvidia can make tons of tiers out of the same chip by just setting different thresholds on the number of usable cores, it isn't all just price discrimination (though sometimes I think they have been found to be fusing off much more than needed for the number of defects as a pure price discrimination play, or that might have been Intel with cache size).

M2 Pro/Max chips will be huge. Only the ultra is using an interconnect between dies, but that's two full max dies. M1 Max is 432 mm^2, that's enormous. M2 is 25% bigger than M1.

There's not a lot of lower core count SKUs for these chips either. There's a few GPU tiers and CPUs tiers but not lots of room to go down. Pro has 8/10 CPU cores and 14/16 GPUs. Max is 10 CPU and 24/32 GPU.

There is no 4-7 CPU core segments like AMD does with their chiplets. Intel has much lower tiers for their dies.

The dies are just huge huge huge.

The iPhone contributes the most to apple’s revenue and margin. It wouldn’t be a great pipe cleaner because they need a ton of chips on a committed launch schedule, and can’t afford any yield screw up there.

With the Mac, they could probably afford a 10% yield, can extend ship times, and in the worst case could even push back a launch.

While the bigger chip does push down yields, my bet is they have more wiggle room than needed to compensate.

I think a lot of people who watch these things suspect exactly what you wrote at the least. Apple funds much of TSMC’s research and development for a new node, pays for production for use in iPhones, both sides make tons of money and repeat the cycle on a new node and other companies come in and buy up capacity on that cutting edge node, seemingly comfortably coasting behind apple. I guess now Apple may use some capacity themselves for their mac cpus after the newest iDevices prove the tech.

Reply on news.ycombinator.com