Is A Nasty Mysterious Windows Update Bug Really Bricking These SSDs?
by Bruno Ferreira · HotHardwareYou may have seen reports on a particularly troubling bit of news doing the rounds, about the recent KB5063878 Windows 11 security update causing multiple types of SSD data corruption, up to and including the operating system not being detected upon rebooting. We're happy to report that there's probably not much to worry about... for now.
The data loss tale came in from X user @Necoru_cat, who witnessed their PC crashing when installing a Steam update for Cyberpunk 2077. According to Necoru, the operating system didn't come back, and the data on the drive was corrupted. Subsequent tests led to the data on the drive being unrecoverable by standard means, as the partition table eventually got corrupted, leaving the PC in a non-bootable state.
Necoru states that a sequential file transfer of 50 GB of more is enough to trigger the issue and believes this issue is related to the KB5063878 update, though we didn't spot anything to establish that exact connection. Although there's some speculation online about actual bricked SSDs, none of Necoru's report indicate that was the case -- merely that the data got corrupted, and the drives are otherwise fine. Some folks quickly remembered last year's debacle with Windows and specific Western Digital SSDs, but other than those two similarities, there's not much to link the stories.
Eventually, they did what a great tech does: attempted to faithfully reproduce the issue via exhaustive testing, across 22 SSDs of various types, ranging from SATA cheapies to fancier NVMe models. They eventually found that 12 drives experienced issues, "level 1" for needing a restart, and "level 2" for major data loss.
While Necoru's findings are definitely concerning at first glance, it's worth noting that they used the same test bench for every drive, leading us to believe that the problem may lie in that PC or its OS install, as it's the one uniting thread across all the tests. Although Necoru initially theorized the problem could be specific to DRAM-less SSDs, their tests include models that do have a cache. Additionally, the drives don't all use the same controller, or even controllers from the same vendor, making it hard to point the finger at some potential firmware bug. It didn't escape our notice that Necoru's using a Ryzen 9800X3D chip on an ASRock motherboard and high-clocked RAM. That combination had some stability issues, but given Necoru's apparent tech pedigree, they're probably aware of that and have all BIOS updates in place.
As far as we could gather, there's only one person reporting a similar issue, but there's also nothing that connects both stories together. I experienced a similar failure at one point, but it was simply a dying SSD that I got replaced. The commenter Hotz at NotebookCheck brought up a good theory, that the SSDs could be overheating and wouldn't be detected on the next boot, especially given that the temperature sensors on many models are allegedly inconsistent and take too long to throttle the drive. That would make sense in a scenario with a sudden intense load, except that once again, the SSDs tested are far too dissimilar for that to be a cause.
At any rate, if a 50 GB sequential write would be all that takes to corrupt a system, then a session of intensive benchmarking, say, with CrystalDiskMark, would be enough to reveal the problem. Although it seems like this is a system-specific issue, we'll be sure to keep our eyes peeled, on the off-chance that there's actually a widespread problem. We'd also like to salute Necoru for their intense testing, regardless of the result.