Seagate sees big drive capacity jump coming

Original URL: https://www.theregister.com/2011/01/08/seagate_fifth_gen_pmr/

Doesn't fear NAND encroachment

Posted in Channel, 8th January 2011 10:00 GMT

Seagate expects a significant increase in disk capacities this year and reckons it's got the jump on the competition.

Stifel Nicolaus' Aaron Rakers was present at recent Seagate analyst briefing and has posted a note to clients explaining Seagate's view of the world.

First of all Seagate is unfazed by the supposed coming tablet boom, with dozens of flash-using tablets bursting through the doors blasted open by the iPad, and multi-level cell flash poised to make inroads into the netbook and notebook markets.

Second, Seagate expects 2011 to be a year of meaningful capacity increase transitions across its product line with a step up in areal density with sixth generation perpendicular magnetic recording (PMR). It's saying there could be/will be a desktop hard disk drive (3.5-inch) transition this quarter and a notebook one (2.5-inch) around the middle of the year.

Seagate thinks it is three months ahead of competitors in the desktop HDD space and up to nine months ahead in the enterprise product space. We assume this is ahead in shipment and not announcement terms.

Big words

Them there's big words, Mr Seagate CEO Lukzo, especially given Western Digital's desktop and notebook drive shipment increases in the past couple of years. If Seagate's claims are accurate then maybe the company will claw back the unit ship leadership that Western Digital wrestled away from it last year. It is surely lusting to do so.

Where are we with Seagate's current products? The 7,200rpm Barracuda desktops are at 2TB with the 4-platter XT and a 347Gbit/in² areal density. A 2TB, 5,900rpm Barracuda Green has three platters and 422Gbit/in². The enterprise Cheetah 15K.7 (15,000rpm) has 600GB and 225Gbit/in² and the enterprise 2.5-inch Savvio is at 600GB with the 10K.4 (10,000rpm) and 252Gbit/in², and 146GB with the 15K.2 (15,000rpm) and 237.1Gbit/in².

The notebook Momentus 5400.7 offers 640GB on two platters with 507Gbit/in² and is reckoned to be a fifth generation PMR drive. Where are these puppies going?

That depends on how much of a step change there is in areal density. The 4th to 5th generation PMR increase was around 30 per cent so let's use that as the gen 5 to gen 6 increase, and see what it gives us. Barrucuda XT and Barracuda Green desktops jump from 2TB to 2.6TB while the 3TB, 5-platter Barracuda XT goes up to 3.9TB - surely that would be pushed to 4TB though.

The enterprise Cheetahs go to 780GB and the 10,000rpm Savvio will reach 780GB too, while the 15,000rpm Savvio will increase its capacity to 190GB - which might get pumped up to 200GB for round number comfort. In the notebook area the Momentus would get 832GB.

With Seagate predicting Hitachi GST, Samsung Toshiba and WD will be between a quarter and nine months behind in the desktop and enterprise HDD markets respectively, it has that much time to ship as many drives as it can without any competition.

NAND Encroachment? Have no fear

Why will drive shipments be unaffected by customers deserting spinning rust for flash chips? Seagate still expects its hybrid Momentus XT, the disk drive with a 4GB NAND cache, to go gangbusters with a million drives shipped a quarter by the end of 2011. These are drives that are chosen instead of SSDs. Secondly, the installed PC base will rise and need bulk capacity disk drives to store all the digital media and general file stuff users accumulate. But these two factors are not the main thrust of Seagate's argument; that is a NAND manufacturing capacity one.

Even though users might wish to and could afford to adopt SSD storage on their computers the lovely stuff can't be made in sufficient quantities. Seagate quotes an industry estimate that there will be a total notebook HDD capacity of 95EB (exabytes) shipped in 2011 with 69EB in 2010. It says - you can almost see the glee on its face - that just 11EB of flash was made in 2010, with around 93 per cent of shipments used for smartphones and other embedded applications. This left a pathetic 0.77EB (770PB) for notebooks, which would have allowed around one per cent of notebook HDD storage to be replaced by SSDs last year.

There will be a minor boost in NAND fab capacity this year, maybe enabling 2EB to be used for notebooks. Any demand for notebook SSDs over this amount will cause shortages which will send prices up, ultimately crushing the demand. What Seagate seems to be saying is: "NAND flash may well be preferable to HDDs for notebooks but it's supply-constrained so we are alright for the time being."

It also claims there are high failure rates in the enterprise SSD market, up to an amazing 35 per cent, so there won't be as much gulping of fast Fibre Channel and SAS enterprise drives by flash as expected. Any flash impact on Seagate's revenues will be much less than some analysts think, and Seagate will gain a revenue boost from having higher capacity drives on the market before anybody else.

Apparently, Seagate will used its own heads in these new disk drive products and not TDK ones. Another interesting point is that the Seagate/Samsung collaboration will result in a third generation Seagate SSD in the second half of 2011. Although Samsung will have helped developed the product, Seagate has sole selling rights to the product for 18 months - after which Samsung can sell any competitive offering it wants to. Whoever negotiated that partnership contract did a good job for Seagate, and Samsung must expect Seagate to ship land-filling amounts of the stuff to make the contract acceptable in Korea.

Seagate may be under-estimating the NAND fab capacity increases coming. EE Times is running a story saying that Hynix, Micron, Samsung, and Toshiba are all increasing fab capacities, with an industry analyst suggesting total bit production could rise 71 per cent in 2011 over 2010. This does not square with Seagate's numbers. ®