But I admit that my post was kind of skimpy on context. I don't mean to suggest that Apple Silicon (ASi) will lead to the end of x86. But I also think ASi has the potential to be more than just "niche", though I guess it depends on how you define niche.
Let's first acknowledge that it's a minority share of the PC market that really cares about performance, or performance/watt. Many, many PC buyers could get by just fine on Sandy Bridge or even Core 2 Duo levels of performance. And so, one could argue that the market for leading edge processors is itself "niche." But I don't think that's an interesting definition of "niche."
Maybe examples would help. The rumor is that the first Mac ASi will be a 12 core SOC (8 performance, 4 efficiency) fabbed on TSMC 5nm. Presumably that SOC will achieve IPC better than the A13, plus higher clocks. It's not hard to imagine that with such specs, Apple could drop into an $800 MacBook Air or a Mac Mini the performance of Rocket Lake. Chromebook buyers would not care about this. But anybody who cares about performance might care quite a bit. I think the interest from people who care about performance could be enough to nudge Mac marketshare up a few percentage points. Is that niche? Maybe, but it's profitable niche.
In terms of desktop and server --- it could be interesting. As you note, the Mac Pro is server-ish. It's a rack-mountable design. The Mac Mini is amenable to high density deployments. Intel thermals suck, but ASi could thermals could make these very interesting products.
One last thought... by moving to ASi for the Macs, Apple suddenly needs to care much more about the volume of sales than before. The name of the game in profitability for silicon is Volume. I'm guessing that Apple is going to price these ASi Macs to move. I don't mean $350 laptops. But I do mean that 12 core SOC in an $800 MB or Mini.
Buub wrote:Xolore wrote:blastdoor wrote:The devil is in the details. I can imagine Apple’s price points will remain the same, but there might be a big uplift in performance and performance/ watt. The question is — how big is that uplift?
Given the process advantage and apples excellent design team, it could be big.
Apple does not manufacture chips. Any belief that they have any form of process advantage is absolutely false because they don't have a process at all. Don't mistake TSMC for Apple and don't think nobody else has access the the same processes.
Substitute Apple for AMD and your statement is pretty much unchanged.
You are pedantically correct. But I don't think anyone here is beating a drum about "process advantage". Intel is still the king of process, even if the independent fabs are close on their heels.
But that's irrelevant, and who's arguing that anyway?
Apple has some world-class chip designers on staff, and their chips are designed in-house. This is their advantage, not fab process. Like AMD (now), they design the chips and someone else makes them.Xolore wrote:Regardless, Apple doesn't play in servers. They don't actually have a traditional desktop product either.
Apple has dabbled in servers in the past, and the Mac Pro is a server-based workstation. The only reason they aren't in servers now is because it's a distraction from their main market, not that they couldn't do it.
And this very thread is precisely about them entering the desktop market with their CPUs. So, while technically correct, again, you are talking around the discussion going on here, which is what the effect will be when Apple fully establishes their CPUs and SoCs on the desktop.