There is no disputing that being a hard-drive vendor is tough today. Solid-state drives (SSD) are eating into market share at an ever-increasing pace, while hard disk drive (HDD) capacity growth has effectively stalled for at least the next two years.
We have reached a major watershed in the storage industry and face the difficulties of finding the road home. Just looking at bottom line data …how much revenue, how many drives … is insufficient to identify even the next two years of industry roadmap. Will SSD take over bulk secondary storage? Will they completely replace enterprise drives? Will mobiles such as tablets move to SSD? Do the numbers even tell the story?
Let’s look at some facts. From 2010 to 3Q 2017, total HDD quarterly shipments have steadily fallen from around 170 million units to about 100 million units. The primary cause of this is HDD replacement by flash solutions. Market segments are unevenly hit.
Enterprise drives, for years the gold mine of the industry, have lost ground quite rapidly, since SSD prices and HDD prices in this class have been close for some years. The much higher performance of the SSD is the deal clincher here. Enterprise HDD declined from 32 million to around 20 million annually. Note that all-flash Arrays also ate into market share for RAID arrays
At the other end of the spectrum, small SSDs own the tablet space. Size, weight, and power are clear winners for SSD, but the ability to deliver 32, 64 and 128 GB SSDs at lower prices than the smallest HDD is a clincher. With the decline of the PC and the rise of SSD-based tablets, this segment has been a bloodbath for hard drives. Unit volumes dropped from 289 million in 2013 to 135 million in 2017.
External hard drives and consumer electronics drives have been stable segments at about 100 million units each over the last 5 years.
The only bright area for HDD is the nearline segment which is up from 36 million units in 2013 to 40 million units in 2017.
We can expect the downward trend to continue. In fact, it seems to be accelerating and that is worth exploring. PCs face wholesale replacement by the cloud/VDI combination and by mobile devices. There is literally zero market for HDDs in this segment and the rate of decline should be expected to increase rapidly through 2020.
Likewise, even the RAID ultra-conservatives can’ hold out much longer in the face of PCIe performance. This is already evident in the poor external storage revenue last quarter over at Dell-EMC. External hard drives will give way to cloud backups, eroding that market, while flash will infiltrate the consumer space as die prices continue to drop.
Nearline storage will come under serious price pressure as die costs fall and as QLC (a 40% increase in per-die capacity) and 3D NAND both mature. Moreover, the fact that 100 terabyte (TB) SSDs are announced for mid-2018 delivery means that SSDs will have a 5x capacity per box advantage over the 14TB drives that lead the HDD pack.
In the context of hard drives, technology means either HAMR (Seagate) or MAMR (WD). Both approaches seek to soften the magnetic surface in a tiny are to make writing the bit easier and to shrink bit size. HAMR uses a laser on the recording head, while MAMR applies microwaves.
Both techniques are difficult to design, never mind build in huge quantities, leading to delays getting to market. 2018/2019 is the current bet. Now, if the capacity gain were huge, this would be a no-brainer, but the best prediction is a 3.5” drive with 25TB in 2023. Given we already have 32TB in a 2.5” SSD, this is a bit of a yawn!
HDD proponents usually point to cost/TB as the main advantage of the HDD. Currently, a 1TB hard drive is around $50 on the Internet, while SSDs are perhaps 7 to 8X that. Usually, that’s the point at which the proponent’s analysis ends.
Reality is far more complex. With as much as 100X the random access performance, even a cheap SSD is a Ferrari compared with a golf-cart. That extra performance can be used for data compression or to speed up server operation. Combine the higher capacity per drive, with compression and reduced server count and you need fewer drives, appliances, servers etc. Get the picture?
Hard drives are finding life tough and it’s getting tougher. A collapse of the market isn’t likely, but market share erosion will speed up and that will especially hurt Seagate, who have little presence in the flash market. Margins are already thin, so any price war will be limited, but such a war seems inevitable if not already happening.