Discover more from The Storage Stack
Replacing HDDs with SSDs
Is the necessary NAND foundry capacity in place?
One of the most important discussions in the storage industry has been opened up by Pure Storage saying no new hard disk drives (HDDs) will be sold after 2028.
Pure's claim has two main implications. First, that all the capacity of the HDDs that would have shipped in 2029 will be replaced by shipped SSD capacity. That is a large amount of extra SSD capacity.
Second, as HDDs have a life fixed by corporate accounting rules or device failure, then the entire existing global population of HDDs will have to be replaced by SSDs. That is an exceedingly large number.
Both implications have an impact on global NAND manufacturing capacity.
As an exercise in weekend back-of-the-envelope estimation I attempted to see if Pure's prediction was even possible, just from the NAND manufacturing point of view – ignoring TCO implications for HDDs and SSDs. This needs explaining
HDD and SSD price trends
Hard disk drive sales in unit terms have been consistently falling for more than five years because SSDs offer a compelling alternative in more and more drive categories. SSDs respond to IO requests faster than disk drives because, unlike HDDs, they don't have to wait for a read-write head to be moved across a disk platter's surface to the right track and then wait until the disk's rotation brings the right data block to the head; the seek time.
SSD costs have been falling steadily although at a decreasing rate as the amount of flash capacity in a die has increased. This is due to three things: the NAND die components shrinking in size, the addition of more layers of cells in 3D NAND technology, and the use of multi-bit cells with TLC (3bits/cell) now mainstream. Data reduction technologies (compression and deduplication) mean that an SSD can hold more data than its raw physical capacity, moving its cost/TB closer to HDDs when they don't use data reduction.
The overall HDD market has declined because of this, with notebook computers moving almost completely to SSDs, desktop PCs following suit, and fast (10,000rpm 2.5-inch) mission-critical HDDs being replaced by SSDs as well. The HDD business market is coalescing around nearline (7,400rpm, 3.5-inch) drives. However, QLC (4bits/cell) SSD technology brings SSD cost/TB down another notch and continuing layer count increases will sustain this downward trend.
HDD manufacturers say their technology advances, such as HAMR, will continue to bring their disk prices down and so preserve their cost advantage over SSDs. Flash drive prices will not become lower than HDD prices because HDD cost declines will sustain disk's affordability.
Wells Fargo senior analyst Aaron Rakers predicted in 2019 that enterprise storage buyers will start to prefer SSDs when prices fall to five times or less that of hard disk drives. He noted an 18x premium in 2017 for enterprise SSDs over mass capacity nearline disk drives. This dropped to a 9x premium in 2019. In 2020 Rakers said enterprise SSDs cost in general $185/TB and nearline HDDs cost about $19/TB, meaning enterprise SSDs carried a 9.7x price premium. This had stayed constant for several quarters, as a chart shows:
Customers of HDDs generally want faster data access and would use SSDs if the costs were equal to or less than HDDs, and SSDs lasted as long as HDDs. Some analysts suggest that when the SSD price premium over HDDs is 5x or less then there will be a general move to replace HDDs with SSDs. Others say that this is not likely because there simply isn't enough NAND manufacturing capacity in the world to replace all the disk drive capacity that customers buy.
Now Pure has become the first flash storage supplier to say unambiguously that flash will replace hard disk drives, with no new HDDs being sold after 2028. It says its QLC flash systems can replace nearline storage drive arrays because their TCO over 5 years is lower than a nearline drive array. The continuing decline in flash's $TB, coming from even higher density drives, will strengthen this advantage.
A TCO calculation starts from acquisition prices, and then brings in drive lifetime power and cooling costs, and also disposal costs at end-of-life. These are estimated as they lie in the future, and disk drive manufacturers can present alternative numbers based on different assumptions.
Setting the cost and price issue aside, there is the question of whether there is enough NAND manufacturing capacity to replace the millions of hard drives sold each year.
NAND manufacturing capacity
Micron Senior Director Colm Lysaght, said in 2019: “Clearly SSD price/GB will get closer to HDD price/GB over time. … However, the raw number of EB needed for a 'wholesale switch' from nearline HDD to SSD is far too large for the NAND flash industry to contemplate. The capital investment needed to generate the EB required … is prohibitively expensive.”
In his view: “SSDs may nibble (and maybe even munch) at the nearline HDD market, but both will coexist for many years to come.”
Analysts such as Wikibon's David Floyer have said NAND manufacturing capacity will not be a limiting issue. He forecast in 2021 that NAND production efficiencies will result in SSDs becoming cheaper than HDDs on a dollar per terabyte basis by 2026.
Here we are in 2023 and Pure has laid down the flash-will-replace-disk-drive gauntlet. This indicates that flash manufacturing capacity will not be a limitation. "The capital investment needed to generate the EB required" is no longer "prohibitively expensive.” Is that true? Can NAND manufacturing cope with the disk drive capacity replacement load?
Modelling NAND manufacturing
How should we do this? We start with approximately 1,320EB of HDD capacity and 277EB of SSD capacity shipped in 2022, gleaned by checking vendor and research house numbers. If the 1,320EB of shipped HDD capacity in 2022 falls to zero EB by 2029 and data growth continues, then the SSD industry will have to grow to meet its own market needs and also supply the missing HDD contribution. Does it have the capacity to do this?
It will be affected by stored data growth in the interim period. This will involve fairly simplistic assumptions, but that's acceptable here because we are only trying to see if the manufacturing capacity basics are in place.
What is the projected data growth from 2022–2030? Fortune Business Insights says: "The global data storage market size was valued at USD 217.02 billion in 2022 and is projected to grow from USD 247.32 billion in 2023 to USD 777.98 billion by 2030, exhibiting a CAGR of 17.8% during the forecast period."
Therefore the 1,320EB HDD capacity of 2022 will grow to a stored capacity need of 4,155.2EB by 2029.
What is the current SSD NAND manufacturing capacity? With 277EB shipped in 2022, say its total theoretical manufacturing capacity was 300EB that year.
If stored SSD capacity has a 17.8 percent CAGR to 2030 that means its NAND manufacturing capacity need will grow to 944.4EB by 2029. This is far, far short of the total 5,099.6EB needed to provide for SSD intrinsic market growth needs (944.4EB) plus the 4,155.2EB of replaced HDD shipped capacity by that date.
We are assuming that data growth for HDD capacity is the same as that for SSD capacity – a big assumption. But since flash supplier Pure Storage is saying SSDs (and Pure's flash drives) will replace SSDs, it seems a reasonable enough one to make.
We are currently looking at 200 or so layers. Say we have a 3x increase in manufactured SSD capacity by 2029 from additional 3D NAND layers. That brings NAND SSD manufacturing capacity to 944.4 x 3 = 2,833.1EB – still short of the 5,099.6EB needed.
Say there's an additional 33.3 percent increase with a switch from TLC to QLC NAND. That brings us to 2,833.1 x 1.33 = 3,768 EB – a shortfall of 1,331.5EB from the 5,099.6EB needed. It means extra NAND manufacturing capacity is required, and that means new NAND foundries have to be built
How long does it take from construction start for a new NAND plant to start shipping wafers and then SSDs being produced?
Say it takes four years from deciding the financial commitment and building a new NAND foundry to the new plant shipping wafers to SSD manufacturing. That means the new NAND fabs needed to produce 1,331.5EB of NAND by 2029 need to have been decided upon by 2025 at the latest.
How many new fabs? That depends upon wafers per month output and wafer capacity. A foundry with a 200,000 wafers/month output of 180TB wafers will produce 432EB per year. Three of those would produce 1,296EB, four 1,728EB.
The financial commitment will be substantial.
Replacing installed HDD population
If no new HDDs are sold after 2029 and the existing HDD population then has a 3–5-year life span, due to corporate accounting rules, then that means that all the global installed HDD capacity will have to migrate to SSD by 2032–2034.
What is the total global installed HDD capacity?
Say it is 10 times the 2022 shipped HDD capacity – that would equal 13,200EB. That means the NAND manufacturing industry would have to supply capacity for native SSD needs plus 13,200 extra EB in HDD replacement capacity by 2034.
Say it actually manufactures 5,099.6EB by 2030 – needed for intrinsic SSD market growth and HDD replacements. After that it needs to satisfy internal SSD market growth needs, plus add 13,200EB to replace the global population of end-of-life HDDs by 2034. In other words it needs to increase NAND manufacturing capacity more than 250 percent in four years just to replace the end-of-life HDDs.
That may not be possible. Either something's wrong with my estimations – quite likely – or installed HDDs will have to spin for a lot longer.