Microsoft's August 12, 2025 cumulative update for Windows 11 version 24H2 (KB5063878, OS Build 26100.4946) is causing some SSDs to drop off the system entirely during sustained large file transfers. Testers and community forums have reproduced a failure pattern where drives become unresponsive, vanish from File Explorer and Disk Management, and in severe cases leave partitions showing as RAW — with data potentially corrupted or unrecoverable. Phison, a major SSD controller supplier, acknowledged the investigation while simultaneously launching legal action over a falsified document that circulated fake lists of affected models.

The Update and the Symptoms

The update arrived on August 12, 2025, bundling security and quality fixes. Within days, independent benchmarks and user reports converged on a dramatic failure mode: after writing a large contiguous file — typically around 50 GB — certain NVMe and SATA SSDs suddenly disappear from the operating system. The drives vanish from Device Manager, SMART telemetry becomes unreadable, and any file operations in progress are truncated or corrupted. In a minority of cases, the partition flips to RAW, and a reboot doesn't bring the drive back; it may require a firmware reflash or a warranty return.

The trigger workload is consistent across reproductions: sustained sequential writes on drives that are already moderately full — community tests often cite 50–60% used capacity. Real‑world activities that invoke the bug include downloading massive game updates, extracting large archives, cloning disks, or copying video projects. The failure is not universal; millions of PCs installed the patch without incident, but for those who hit it, the consequences are severe.

Which Drives Are Affected?

Early community test matrices show an over‑representation of SSDs built around Phison controllers, though the issue isn't confined to a single vendor. Drives from Corsair, Kioxia, SanDisk, ADATA, and others appear in investigative lists. However, those lists are heuristic, not definitive. Firmware revisions, host BIOS/UEFI interactions, and specific PCIe lane configurations all influence reproducibility. Phison's statement underscores that the problem is “industry‑wide” and not limited to its own silicon.

Crucially, a document claiming to be an internal Phison summary of affected controller models circulated among customers. Phison has publicly denounced it as a complete fabrication and stated it is pursuing legal remedies. “We wish to state unequivocally that the document in question … is neither an official nor unofficial communication from Phison,” the company told Tom's Hardware. The fake document, titled “Phison SSD Controller Issues Summary,” included a table of controller models that Phison says is entirely false.

Vendor and Microsoft Signals

Phison’s official statement, issued on August 19, acknowledges the “industry‑wide effects” of KB5063878 and KB5062660 and says it is “working with partners” to identify impacted controllers. Microsoft’s KB article for KB5063878 initially listed no known storage issues — a common lag while telemetry is still being analyzed — but separate release‑health advisories confirm that Microsoft is engaging with OEMs and controller vendors. BleepingComputer and other outlets report that a fix is in the works, though no timeline has been announced.

Outlets including Tom’s Hardware, Windows Central, and TechRadar have independently validated the symptom signature and are advising users to avoid large sequential writes on patched machines. The strong alignment between community reproductions and vendor acknowledgements marks this as a genuine, high‑impact regression.

What Went Wrong? Plausible Technical Mechanisms

No cross‑vendor forensic report has been published, but the collective evidence points to a host‑to‑controller interaction rather than a simple file‑system bug. Several leading theories are under discussion:

  • Controller firmware lockup: Sustained sequential writes stress SLC caching, wear‑leveling, and metadata updates. A subtle change in Windows’ command timing or queuing could expose a latent race condition in controller firmware, causing it to hang and stop responding to the host.
  • Host Memory Buffer (HMB) issues: Many DRAM‑less SSDs rely on a small slice of system memory for mapping tables. If the August update altered the allocation or timing of HMB under heavy I/O, DRAM‑less controllers might lose critical working data. Windows 11 24H2 already had documented HMB‑related bugs earlier, making this a plausible vector.
  • Storage‑stack memory leak: Community testers have hypothesized that a buffer leak in Windows’ storage stack could prevent the drive cache from being flushed correctly, leading to partial writes and sudden controller time‑outs. This remains unconfirmed but fits the observed pattern of large, long‑duration writes.
  • Firmware diversity: Because the phenomenon appears across multiple controller brands (not only Phison), the root cause might be a Windows change that interacts with a common design pattern — for example, the way SLC caches are emptied or metadata writes are atomic — which different implementations handle with varying robustness.

Until vendors release validated diagnostics, these narratives should be treated as informed speculation.

Risk Assessment

The issue is not widespread; typical office and web‑browsing users are unlikely to trigger it. However, the risk is highly focused and potentially devastating. Users most at risk are those who:
- Perform sustained large sequential writes (game installers, video exports, archive extraction),
- Have SSDs that are more than half full,
- Own drives from controller families flagged in community test beds.

Even if the update works fine today, a future large game patch or bulk copy could suddenly expose the bug. A verified backup is the single most important safety net.

Immediate Action Checklist

  1. Back up critical data now. Use an external drive or cloud service and verify the backup. The 3‑2‑1 rule (three copies, two different media, one offsite) is ideal.
  2. Avoid large, sustained writes on any drive that has received KB5063878. If you must move a big file, split it into chunks of 10–20 GB and monitor drive health.
  3. Delay the update on unpatched systems, especially production or content‑creation machines. Pause Windows Update or hold the update via policy until official guidance is available.
  4. Check SSD firmware using the vendor’s toolbox, but do not flash an update without first having a verified backup. Firmware revisions will likely be the primary long‑term fix.
  5. For enterprises: Use staged rollouts. Inventory at‑risk drives, apply the update to a pilot group with representative storage hardware, and be ready to deploy a Known Issue Rollback (KIR) when Microsoft makes it available.

If a Drive Disappears Mid‑Write

  • Stop all I/O immediately. Continued writes can overwrite residual metadata and complicate recovery.
  • Do not reformat or re‑initialize the drive unless you have a validated disk image or are prepared to lose the data permanently. Formatting destroys forensic evidence.
  • Capture logs: Grab Event Viewer entries, NVMe driver logs, and output from the vendor’s diagnostic tool. This information will be essential for RMA or recovery attempts.
  • If the data is valuable, create a forensically sound image using write‑blocking hardware or read‑only imaging software, then work with a professional recovery service.
  • Contact the SSD vendor for guidance and a possible RMA. Vendors are the distribution channel for any controller‑side firmware fixes.

What a Fix Will Likely Look Like

Two parallel remediation tracks are likely:

  • SSD firmware updates: If the root cause is a controller‑side race condition or cache‑management flaw, vendors will issue firmware patches through their utilities or partner dashboards. This is standard procedure for such issues.
  • OS‑side mitigation: Microsoft can deliver a servicing update or a Known Issue Rollback (KIR) to adjust host‑side behavior — for example, disabling a problematic command sequence or altering HMB allocation. A KIR often arrives faster than a full cumulative update.

Both paths will probably be needed: a firmware change to fix the controller’s vulnerability, and an OS update to prevent the workload from re‑triggering the condition on unpatched firmware.

Timeline

  • August 12, 2025 – KB5063878 released.
  • Mid‑August – Community test cases reproduce SSD vanishing during sustained writes; reports surface on forums and tech press.
  • August 18–20 – Phison issues official statement and warns of fake document; Tom’s Hardware, BleepingComputer, Windows Central, and others confirm vendor engagement and advise caution.
  • Ongoing – Microsoft and SSD vendors continue joint investigation; no fixed resolution date announced.

Key Takeaways

This regression is serious but narrow. The rapid community reproduction helped accelerate vendor triage, and the involvement of both Microsoft and major controller suppliers signals that a coordinated remedy is in progress. Phison’s legal response to the falsified document underscores the chaotic information environment that can surround such incidents — users should rely only on official vendor channels and reputable tech press.

For now, the safest posture is to assume that any patched system could be vulnerable during a heavy write workload. A robust backup, careful staging of the update, and avoidance of large sequential transfers are the most effective defenses until firmware and OS patches materialize. Monitor the Microsoft Release Health page and your SSD vendor’s support site for the official word.