One popular argument to explain why SSDs have not displaced the HDDs in all PCs is that there isn’t enough NAND flash production capacity to support this business and there never can be.
This argument has been posed as long ago as 2007 by WD CEO John Coyne at an IDEMA conference (the source of this post’s graphic), SanDisk’s Eli Harari at the Flash Memory Summit in 2008, and Seagate‘s CEO Steve Luczo in a Forbes interview as recently as last April. These are captains of the industry. Their arguments make people stand up and take notice.
It’s a really flawed argument.
It goes like this:
- In 2011 400 exabytes of HDDs shipped
- A typical NAND flash wafer fabrication plant (a “fab”) outputs 3.4 exabytes in a year and costs $6 billion
- To replace HDDs with NAND would require 117 of these fabs at a total cost of $700 billion
- Nobody has that kind of money
Money has a way of coming out of the woodwork when there’s a clear return. When the photovoltaics industry underwent a polysilicon shortage in 2007-8 $50-100 billion popped up to overbuild polysilicon production, resulting in today’s glut and FTC charges of dumping.
Although $700 billion is not going to be invested in a $25 billion market organically, there is no reason to rule out a $700 billion investment from sources external to the market, provided there is a clear path to profits from that investment.
The real argument is that there has not been such an investment because there is no reason to expect consumers to change their behavior. Today these consumers select a PC based, in part, on price and HDD capacity, and this rules out their choosing an SSD. This is the real reason that HDD makers need not fear that their business will be snatched away from them by NAND flash makers.
Objective Analysis regularly compiles reports covering the SSD market, and the forecasts in these reports are based on a clear understanding of the market gained through rigorous end-user surveys. This has resulted in forecasts that have consistently been the most accurate in the business. These reports can be ordered for immediate download from the Objective Analysis website.
Math in here is wrong.
– In 2011 400 exabytes of HDDs shipped
– A typical NAND flash wafer fabrication plant (a “fab”) outputs 3.4 petabytes (0.0034 exabytes) in a year and costs $6 billion
– To replace HDDs with NAND would require 117 of these fabs at a total cost of $700 billion
Nobody has that kind of money
JP,
You are quite correct!
I have revised my embarrassing 3-orders-of-magnitude error in the post, but those curious about my error can find it in your comment.
Thanks for pointing it out.