Factory contamination affects "at least" 7 billion gigabytes of flash memory:
Solid-state storage devices have so far been spared from the scarcity and high prices that the chip shortage has wrought upon graphics cards, cars, Raspberry Pi boards, and innumerable other products. But that may change soon, due in part to a "contamination" at two Japanese factories used by Western Digital and Kioxia to make flash memory.
According to a short Western Digital press release, the contamination issue has affected "at least" 6.5 exabytes of flash memory, which works out to a bit under 7 million terabytes or 7 billion gigabytes—that's a lot of NAND that will suddenly be unavailable for SSDs, phones, memory cards, and USB drives. An analyst speaking to Bloomberg suggested that the final total of the lost capacity could be as much as 16 exabytes.
(Score: 3, Informative) by Anonymous Coward on Friday February 11 2022, @05:19PM (4 children)
According to the highly trustworthy page: https://en.wikipedia.org/wiki/Flash_memory#Industry [wikipedia.org] the industry produced almost 232 billion gigabytes in 2018, so a loss of 7-16 billion would be 3-8% loss of that capacity. Projecting actual output for 2022 its presumably less, maybe 5% loss ??
So not insignificant but not OMG panic flash drives either.
(Score: 1, Touché) by Anonymous Coward on Friday February 11 2022, @05:46PM
> but not OMG panic flash drives either.
If my smartphone tells me to panic I panic!
I wonder if the amount of affected hardware makes a recycling effort a thing. I'm thinking there's several tens of tonnes of chips involved-- all fused epoxy, metal (copper? gold?), and doped silicon-- which if recovered would have significant value.
...or they could just label the affected NANDs as "SX" versions and flog them off cheaper with reduced capacity & etc.
(Score: 2) by EvilSS on Friday February 11 2022, @07:42PM (1 child)
Also insert conspiracy theory here
(Score: 2) by JoeMerchant on Friday February 11 2022, @08:53PM
There's also the fact that the "outage" isn't going to hit the broad spectrum of devices on the market, but a narrower slice, so that slice is going to hurt more.
Still, 7 billion GB, that's 70 million 100GB SSDs, or 3.5 million 2TB SSDs, doesn't seem particularly catastrophic in the face of the total market capacity.
Україна досі не є частиною Росії Слава Україні🌻 https://news.stanford.edu/2023/02/17/will-russia-ukraine-war-end
(Score: 3, Touché) by Revek on Friday February 11 2022, @09:41PM
This won't stop the price of these goods from going up out of proportion to the deficit.
This page was generated by a Swarm of Roaming Elephants
(Score: 1, Interesting) by Anonymous Coward on Friday February 11 2022, @05:39PM (5 children)
I haven't seen any news source reveal the nature of the contamination, and that information seems like it would be of great interest. Does anyone have a rumor or guess about the nature of the contamination?
(Score: 5, Touché) by Anonymous Coward on Friday February 11 2022, @05:49PM
>> Does anyone have a rumor or guess about the nature of the contamination?
An analyst on Fox said it was because of Immigrants.
(Score: 3, Interesting) by drussell on Friday February 11 2022, @05:58PM (1 child)
Yeah, it would sure be nice to get more specific information on this one...
With just that vague investor-centric press release to go on for specifics, it really could be anything.
It essentially be anything from "we got a metric fukton of wafers that were contaminated with organic or metallic surface impurities that were not removed during manufacture and we somehow didn't notice that the usable yield was lower than expected, so we've lost x portion of our expected output" to "we screwed up the deposition and processing of a metric fukton of wafers, leaving some sort of metallic contamination that slowly builds dendrites or whiskers over time after fabrication, which we only noticed now, so we have to condemn all this previous production of x." Who knows.
The investor-centric press release spends all of it's verbiage trying to reassure investors going forward, with virtually no explanation of what actually happened (so those same investors could actually be able make rational decisions?) Nice one. 🤦
(Score: 2) by mhajicek on Friday February 11 2022, @09:31PM
The boss's kid put a peanut butter and jelly sandwich in the vapor deposition feedstock slot.
The spacelike surfaces of time foliations can have a cusp at the surface of discontinuity. - P. Hajicek
(Score: 2) by krishnoid on Friday February 11 2022, @06:53PM
It was in two Japanese factories, so I'm guessing ... tentacle bunnysuit supply-chain issues?
(Score: 2, Interesting) by fustakrakich on Friday February 11 2022, @09:08PM
Low prices
La politica e i criminali sono la stessa cosa..
(Score: 2) by drussell on Friday February 11 2022, @05:43PM (2 children)
How do you end up with "contamination" of such a large amount of product before you notice?
I presume these things work through the testing phase but will fail quickly in operation or somesuch?
Prepare for even cheesier cheap Chinese flash drive "storage" devices with all the retention qualities of /dev/null 🤦
(Score: 0) by Anonymous Coward on Friday February 11 2022, @05:57PM (1 child)
There are many manufacturing errors that aren't easily visible. A new employee missing a critical step without obviously failing can do a lot of damage. By the time it's detected and traced to the source, you have big problems.
(Score: -1, Troll) by Anonymous Coward on Friday February 11 2022, @06:37PM
But they don't have Mexican workers in Japan.
(Score: 5, Informative) by dltaylor on Friday February 11 2022, @06:45PM (3 children)
Way back in the day, UV EPROMs had the chip mounted on a ceramic carrier for the DIP package. Some genius at Intel found a to "save the company some money" by buying cheaper clay to fabricate the ceramic carriers. The EPROMs passed QA, but were failing at ridiculous rates in actual use: dropping bits at an alarming rate. Much time and money later the cause was found: the clay was cheaper because it was low-level radioactive. The beta (maybe gamma, it was a long time ago) emissions were penetrating the chips and, over time, reducing the stored charge in the cells.
(Score: 2) by krishnoid on Friday February 11 2022, @06:52PM
+1 womp womp [wompwompwomp.com]
(Score: 5, Informative) by EvilSS on Friday February 11 2022, @08:04PM (1 child)
(Score: 2) by dltaylor on Sunday February 13 2022, @04:02AM
May have been two different issues. I remember having to pull/replace a lot of EPROMs to restore reliability to the systems that the company I worked for produced.
(Score: 0) by Anonymous Coward on Friday February 11 2022, @07:31PM (1 child)
How many parts were shipped before the contamination was discovered?
How long will the lines be idle while the contamination is corrected?
How long will it take to catch-up on the production orders, given that some things already have a 52+ week lead time?
How many products using that flash will be affected? This includes possibly halting production on products waiting on flash parts.
(Score: -1, Flamebait) by Anonymous Coward on Friday February 11 2022, @07:39PM
>> How many parts were shipped before the contamination was discovered?
Quite a few if you do the math: most common capacity for SSD is 1 terabyte so that's 7 million SSDs. Fortunately I use Samsung so I don't have to worry about Western Digital's sloppy business practices.
(Score: 2, Touché) by fustakrakich on Friday February 11 2022, @07:34PM
Most excellent!
La politica e i criminali sono la stessa cosa..
(Score: 0) by Anonymous Coward on Saturday February 12 2022, @01:40AM
pentagon is requisitioning these things
the could be building an army of bigdogs with frikkin lasers