A wafer of 3D XPoint chips. Image source: Intel.
Last year, Intel (INTC +7.14%) and memory maker Micron (MU +4.80%) announced a new generation of memory technology known as 3D XPoint (pronounced "3D cross point"). The technology itself is quite compelling, promising to serve as a "middle ground" between typical NAND flash and DRAM in terms of performance and cost.
A couple of Apple (AAPL +2.77%)-centric websites -- namely MacRumors and Apple Insider -- raise the interesting possibility of Apple potentially using Intel's 3D XPoint-based drives at some point in the future.
Here's why a 3D XPoint win inside the MacBook could be beneficial to both Intel and Apple.
Apple cares about storage performance; 3D XPoint is ridiculously fast
Apple has made it crystal clear that it cares about the storage performance of all of its devices, from iPhone to MacBook Pro. This makes a lot of sense due to the simple fact that the faster that data can be moved from a storage device to main memory so that the CPU can use it, the snappier the overall system performance is going to be.
The transition from hard disk drives to much faster NAND flash based drives was arguably one of the most important transitions in recent computing history. Although I doubt that the transition to 3D XPoint transition has the potential to be as transformative in consumer PCs (in data centers, it's a whole different ballgame), it should still be good for computing performance.
The introduction of 3D XPoint into some MacBook models (I would expect that it would be the MacBook Pro systems first, with the mainstream MacBook/MacBook Air getting the technology later once it is more cost effective), could be a win/win for both Apple and Intel.
Apple would win by offering customers a massive performance improvement that it could conceivably charge a fat premium for and Intel, which currently does not supply the storage solutions into Apple's Mac product line, would gain additional content share in Apple's Macs.
How could Apple pull this off?
The major problem with 3D XPoint in the Mac is, no doubt, going to be cost. It's not clear how much Intel plans to charge for the technology, but on a per-gigabyte basis, 3D XPoint is very likely going to be much more expensive than NAND flash (but cheaper than DRAM).
If Apple were to simply swap out NAND flash for an equivalent amount of 3D XPoint memory, this would lead to one of two things:
- Apple's per-device margins would go down if it tried to keep price points flat; or
- Apple would charge customers (substantially) more to preserve margins
The first choice would simply be bad for shareholders and the second would probably lead to a decline in MacBook Pro unit shipments. Apple could conceivably offer up variants of its systems with and without 3D XPoint, but it would be very hard to convince the average customer to buy up the stack for a benefit that he/she probably won't understand the true value of.
Such systems would become niches within niches and not be worth the bother for the Mac maker.
What Apple could do, though, is to apply the same fundamental idea that it uses with its Fusion drives: combine a small amount of 3D XPoint with a fairly large amount of NAND flash -- Apple's take on a hybrid drive.
According to AnandTech's write-up on the Fusion Drive, OS X works to keep frequently used files/programs on the fast NAND flash portion of the drive while less frequently used/important applications are moved to the hard disk drive.
Apple could do the same thing with 3D XPoint and NAND flash taking the places of NAND flash and a typical hard disk drive. This would be fairly cost effective and likely deliver a better experience than just NAND flash alone.
