Getty Images/iStockphoto
Kioxia, WD NAND contamination to affect flash pricing
A contamination last month of WD NAND that also affected Kioxia has been corrected, but the fallout from disrupted flash memory production is not yet known.
A contamination at two Western Digital Corp. and Kioxia Corp. fabrication facilities, which caused the production of 3D NAND to be suspended, has been fixed.
Western Digital (WD) and Kioxia revealed in early February that a contamination at the Yokkaichi and Kitakami facilities affected 6.5 exabytes, or 6.5 million terabytes, of flash due to a contamination in materials used in the manufacturing process. In early March, WD said the facilities were forced to close for two weeks, which amounted to an additional loss in NAND production, but are now fully up and running.
The loss included 13% of NAND produced in the first quarter of 2022, but will ultimately amount to about 3% of the total 2022 output lost to the contamination, according to research from TrendForce, a market research firm based in Taipei, Taiwan. The NAND produced at the facilities was primarily aimed at client SSDs and embedded MultiMediaCard.
Experts said the disruption to 3D NAND production will affect flash pricing, but to what extent is yet to be seen. It could see an increase beyond the market trend or have little effect. Market share from the joint WD-Kioxia production, which makes nearly one-third of the NAND produced, was as high as 32.5% in the third quarter of 2021, according to TrendForce.
Don Jeanette, vice president at Trendfocus, an analyst firm in Cupertino, Calif., said that before the contamination, he had projected NAND pricing to decline in the first and second quarter of 2022, and estimated prices would increase in the second half of the year. His projection for Q2 has changed.
"It is about an 18 to 20 exabyte output production shortfall," he said, due to the loss of product and manufacturing downtime. "That shortfall coupled with the way the market moves puts prices at an increase going into the second quarter."
Two weeks and bad chemicals
Tom Coughlin, president of Coughlin Associates, said that while the exact cause of the contamination isn't known, various factors could create a production disruption, including a loss of power or not being in a full vacuum.
"Potential contamination [in these] devices can lead to either immediate faults or latent faults that could be a problem later on," he said. "They're probably reacting to some anomaly that occurred in the manufacturing plant."
Jim Handy, general director and semiconductor analyst at Objective Analysis, said the word contamination is likely too strong, as it makes it sound as though entire facilities needed to be shut down and scrubbed. But the fabs were up and running in a couple of weeks.
"It is two weeks' worth of production," he said. Although, he added, "it could have taken them a while to discover that there was a problem."
The high degree of purity of the materials used in semiconductors "is phenomenal," he said, which means the smallest impurity can cause significant problems.
"It seems like they just got a batch of bad chemicals," Handy said. "Everything that was made using those chemicals was going to have to be chucked, but it isn't like they have to clean all other tools out that were using those chemicals."
Impact on the market
How the contamination will affect the price of NAND is still unknown. WD and Kioxia are the two largest manufacturers of 3D NAND, a product that is increasingly being used for consumer devices, automobiles, factory automation and data centers.
But, Coughlin noted, rather than supply keeping pace with demand, the NAND industry runs on a boom-bust cycle. When demand starts to climb, NAND production gets underway and prices begin to rise, eventually creating an oversupply. This is where the market was when the contamination happened, he said.
"It is probably not a bad time to have a problem like [the contamination]," he said. "It will probably decrease the supply of NAND somewhat. If NAND prices were dropping, they won't drop as much as they would have without the shortage."
Coughlin said that the amount of NAND affected by the contamination is finite, and if pricing is affected, it will only be for a limited time.
Handy echoed the point. Companies that rely on NAND would likely begin to prioritize what products need flash memory the most, shifting away from less critical uses and keeping the effects of the facility contamination to a minimum.
"If there's a shortage in any NAND application, then I would expect it to be a shortage of off-brand USB flash drives," Handy said.