Microsoft's August 12, 2025, cumulative update for Windows 11 version 24H2, identified as KB5063878 (OS Build 26100.4946), is causing a subset of NVMe SSDs to become unresponsive and, in some cases, corrupt data during sustained large file writes. Independent testers across multiple tech communities have reproduced the failure, which typically strikes after roughly 50 GB of continuous data transfer, though Microsoft has yet to officially acknowledge the storage regression.
SSDs Are Disappearing Mid-Write—Here’s What’s Happening
Within days of the update’s release, reports began surfacing on Japanese tech forums and X (formerly Twitter), notably from user @Necoru_cat, who documented a consistent failure pattern. When performing a continuous write operation—such as copying a large video file, extracting a multi-gigabyte archive, or updating a modern game—the target SSD abruptly vanishes from File Explorer, Device Manager, and Disk Management. Vendor diagnostic tools simultaneously lose the ability to read SMART attributes or controller telemetry.
The trigger appears to be workload-dependent. Community reproductions frequently cite a threshold of roughly 50 GB of uninterrupted sequential writes, with the drive already around 60% full. Several testers confirmed that rebooting sometimes brings the drive back, but the same workload will cause it to vanish again. In more severe instances, the partition or files being written during the disappearance become corrupted and inaccessible, even after a restart.
This isn’t an isolated incident: Notebookcheck, Tom’s Hardware, Igor’s Lab, Guru3D, and Wccftech have all aggregated and verified community reports, lending weight to the correlation with KB5063878. The official Microsoft KB page for the update, however, states—at the time of writing—that the company is “not currently aware of any issues.”
Which Drives Are in the Firing Line?
The common thread in early reports points to SSDs using Phison controllers, though other controllers have also been named. The list is community-sourced and fluid, but drives repeatedly mentioned include:
- Corsair Force MP600 (Phison E16 controller)
- Phison PS5012-E12 reference design
- SanDisk Extreme Pro M.2 NVMe (Triton MP28 controller)
- Kioxia Exceria Plus G4 1 TB (Phison E31T controller)
- Fikwot FN955 (MAP1602 + WDS X3 9070 controllers)
Some drives, such as the WD Blue SN5000 and Corsair MP510, reportedly recover after a reboot but still exhibit the vanishing behavior under load. It’s important to stress: this is not an exhaustive or vendor-confirmed list. Firmware revision, host motherboard chipset, and UEFI/BIOS settings can all influence whether a specific drive is susceptible. Treat these models as starting points for your own risk assessment, not as a definitive ban list.
Why Would a Security Patch Break SSDs?
Modern NVMe SSDs are intricate systems that rely on a tight dance between the controller, NAND flash, and the host operating system. Many DRAM-less SSDs, in particular, lean on the Host Memory Buffer (HMB) feature to cache address mapping tables using the PC’s RAM. A subtle change in how Windows 11 handles memory allocation, I/O timeouts, or NVMe command sequencing can expose latent bugs in the SSD’s firmware.
When a drive is hammered with sustained sequential writes, the controller’s internal metadata paths and flash channel scheduling are already under stress. If a Windows kernel or storage driver alteration—introduced in KB5063878—now issues that workload in a slightly different order or with tighter timing, a previously unreachable firmware edge case can trigger a controller lockup. The drive disappears because it stops responding to NVMe commands entirely. Power-cycling (reboot) can reset the controller, but in-flight data is often lost or left in an inconsistent state, leading to file system corruption.
This isn’t a new phenomenon: OS updates have historically surfaced hardware firmware bugs, but the severity here—potential data loss—elevates it beyond ordinary growing pains.
Protecting Your Data Right Now
If you’ve already installed KB5063878, the single most important step is to back up any critical data to an external drive or cloud service immediately. Assume nothing is safe until you’ve verified your backups. Next, avoid performing sustained large sequential writes on any NVMe SSD that might be at risk. If you need to copy a 100 GB file, split it into smaller chunks if possible, or temporarily move the operation to a SATA SSD or hard drive that isn’t exhibiting the vanishing behavior.
Check your SSD vendor’s support page for any firmware updates released after August 12, 2025. Tools like CrystalDiskInfo, smartctl, or vendor-specific dashboards (Corsair iCUE, SanDisk Dashboard, etc.) can show your current firmware revision and allow you to capture SMART logs. Apply firmware updates only after a full backup and in a controlled environment—a botched firmware flash can compound the problem.
For those yet to install the update, consider pausing Windows Update for a short period. Go to Settings > Windows Update > Pause updates and choose a timeframe. This gives Microsoft and SSD vendors time to investigate and, if necessary, block the update on affected hardware via Known Issue Rollback or release a revised patch.
If an SSD Has Already Vanished
Stop using the drive for anything beyond diagnostic data collection. Repeated reboots to “see if it comes back” risk further corruption, especially if metadata journals are being replayed incorrectly. Instead:
- Preserve logs: Open Event Viewer (System log) and export entries around the time of the failure. Run CrystalDiskInfo or smartctl to capture SMART attributes before they reset.
- If data is critical, create a sector-level image using a tool like ddrescue or a hardware imager. This freezes the drive’s state at the moment of failure and maximizes later recovery options.
- Contact your SSD vendor’s support. Provide the logs and describe the circumstances (KB5063878 installed, heavy write operation). Vendors may have internal tools to revive the drive without a full RMA, but they’ll need evidence.
Do not format or partition the drive until you’ve exhausted recovery avenues or have a confirmed backup.
For IT Administrators: Stage and Test
If you manage a fleet of Windows 11 machines, KB5063878 should not be deployed broadly until you’ve stress-tested it on a representative sample of your hardware models. Use WSUS, SCCM, or Intune update rings to pilot the update on a subset of devices that mirror your actual NVMe SSD population.
Map out the SSD models, controller types, and firmware versions across your environment. Phison-based drives are the current red flag, but any NVMe model could theoretically be affected. Coordinate with your hardware vendors to get firmware that’s been validated against Windows 11 24H2 build 26100.4946, and deploy that firmware in a controlled manner before rolling out the OS update.
Also note that enterprise channels experienced a separate deployment issue with KB5063878: some WSUS/SCCM installs failed with error 0x80240069, which Microsoft addressed via Known Issue Rollback. Ensure your deployment tools are configured to receive the latest KIR policies to avoid a different class of update failure.
What’s Next: Waiting for a Fix
Microsoft typically takes action once its telemetry and partner feedback reach a critical mass. We can expect an updated KB page acknowledging the issue, possibly with a Known Issue Rollback that disables the offending storage stack change. SSD vendors, particularly Phison and its partners, are likely already analyzing firmware patches that harden the controller against the new host behavior.
Until a confirmed remediation lands—whether an OS-side block or a firmware update—the community’s empirical findings serve as the best early warning system. Bookmark the Microsoft Release Health dashboard and your SSD vendor’s support forum. Independent outlets that reproduce the test (like Igor’s Lab) often publish update testing results faster than official channels.
The KB5063878 episode is a stark reminder that routine cumulative updates can have hardware-level consequences. Backups are non-negotiable, and update staging isn’t paranoia—it’s risk management. As SSDs become more dependent on host cooperation, the line between OS and firmware quality will only get thinner. This won’t be the last time a Tuesday patch sends a shockwave through your storage stack.