Apple and others use system-on-module (SoM) architecture, where discrete slabs of silicon are integrated together in a single package, which lets them use a less expensive process for their 64GB of RAM transistors versus the smaller/higher performing/much more expensive node that they’re etching the logic transistors with. But at the end of the day, every OEM is competing for time and space on the same handful of machines in the world that are capable of etching the silicon. Those machines are huge and hugely expensive and can’t be spun up on a whim, so even Apple ultimately has to pay more because the chip makers know that if they don’t they can find someone else who will as there’s no alternative.
Apple and others use system-on-module (SoM) architecture, where discrete slabs of silicon are integrated together in a single package, which lets them use a less expensive process for their 64GB of RAM transistors versus the smaller/higher performing/much more expensive node that they’re etching the logic transistors with. But at the end of the day, every OEM is competing for time and space on the same handful of machines in the world that are capable of etching the silicon. Those machines are huge and hugely expensive and can’t be spun up on a whim, so even Apple ultimately has to pay more because the chip makers know that if they don’t they can find someone else who will as there’s no alternative.