Probably never. Here’s why.
DRAMeXchange is predicting that in 2017 some 41% of laptops will sport SSDs. As the per gigabyte price difference between disk and flash shrinks, the price differential becomes irrelevant at ever-higher capacity points.
Most people have little idea how much storage they use or need. Compulsive music/photo/video folks will use terabytes, but otherwise most users don’t need more than 120GB – and those drives are available for the $50 minimum price that any capacity disk drive needs to be viable.
Small is beautiful
For decades the hard drive industry progressed by making drives physically smaller. Smaller size reduced costs, improved performance and simplified configuration.
That’s why we had drives as small as 0.85″. But flash has killed off form factors under 2.5″.
Performance is sexier than capacity
Flash wins in notebooks and tablets despite a higher cost-per-bit because of its performance (primarily) and space, power and thermal advantages. If one notebook has a $50 disk and the other a $50 SSD, the latter will be obviously faster to consumers and more desirable.
But the external consumer storage market is different. While professionals buy Thunderbolt SSDs for maximum performance and reliability, the small office/home office user sees cost-per-bit and doesn’t require SSD performance – or Thunderbolt.
Even if SSDs fall to 2x disk prices, buyers will continue to prefer low-cost capacity for what are essentially archive applications. The home archive market will get very big because it is cheaper than the cloud. (Why it always will be is another post.)
And 2.5″ drives will remain popular despite costing more per-bit than 3.5″ drives because their size, shock resistance and bus power makes them much more convenient. Convenience sells, even storage.
The StorageMojo take
Yes, 2.5″ drives will lose the notebook market. My guess: 2020 will be the last year for disk-based notebooks.
But the SOHO archive market will save this form factor. Shingled magnetic recording (SMR) is ideal for archiving as it doesn’t handle frequent updates well, and it dramatically improves capacity, helping with the cost issue against flash.
With 4TB 2.5″ USB 3.0 drives running $30-$40/TB online – and 1TB SSDs running $300 and up – disks still retain a huge lead in cost-per-bit. Now vendors need to step up and see what else they can do to tune disks for the consumer archive market.
Courteous comments welcome, of course.
I would expect that the latest 3D NAND technology to drive MLC and SLC NAND prices down drastically.
2020 looks like a long way away when you are waiting for a windows machine that has an HDD for the C: drive. Yet, I don’t think consumers are all that aware of SSD.
My son has an old netbook with a one core atom processor; I was doing some wearable computing experiments that involved having it in a backpack and I found that if was listening to music it would skip if I moved past a lope because of the HDD shutting down from vibrations.
I put an SSD in that machine and it is actually feels much more responsive at most things than a Core i7 machine which has an HDD.
There’s also the fabrication capacity to consider as well. It seems unlikely (I know David Rosenthal has been making this argument) that given the current flash foundry capacity to even build enough to replace disk based systems, so they can very nearly by default charge a price premium from both a supply and performance standpoint.
Ian, the economics are different for disks and flash, since flash is produced in hugely expensive fabs. Even though we are far from having enough fab capacity to replace disks – and we may never have – flash fabs have to be kept running no matter what.
Since most flash goes into consumer devices with strong seasonal demand, the flash vendors build up inventory in slow seasons and need to move it, which is why we see price drops. It appears the flash industry has entered a period of sustained over capacity, so price competition will be fierce. Great for us; not so great for vendors.
However, in the case of 2.5″ drives, their use in notebooks will end in a few years, leaving the external storage their main use case. Vendors should look at building thicker drives – 19 to 23mm perhaps – to bring home the cost and capacity advantages of disks over flash.
Great post. Especially liked your point about the cloud. Something you lease has to be more expensive in the long run than something you buy.