Microsoft’s August 2025 cumulative update for Windows 11 24H2 is causing NVMe SSDs to vanish during large file transfers, leaving some drives permanently corrupted, according to multiple user reports and independent hardware tests. The update, designated KB5063878 (OS Build 26100.4946) and released on August 12, introduced a storage regression that can make drives disappear from Windows mid-write, and in isolated cases, never recover. Controller maker Phison has acknowledged the problem and is investigating, while community testers have reproduced the failure across a range of SSD brands and technologies.

The first alarm came from X user @Necoru_cat, who reported that his SSD disappeared while updating Cyberpunk 2077 on a system running KB5063878. After methodically testing 21 drives, he found that 12 became inaccessible during sustained sequential writes exceeding 50 GB when the drive was more than 60 percent full. One drive—a Western Digital SA510 2TB—could not be revived even after a reboot. Other users quickly chimed in, describing identical symptoms during game installs, large file copies, and backup operations.

The Trigger: Heavy Writes, Vulnerable Firmware

The failure profile is strikingly consistent. When a drive is subjected to a long, continuous write—commonly a game installation, archive extraction, or disk cloning—the target SSD may stall and drop off the bus entirely. Tools like CrystalDiskInfo and smartctl return errors or timeouts, and the drive vanishes from File Explorer and Device Manager. A reboot often restores visibility, but resuming the same heavy-write sequence reliably triggers the failure again, signaling a deterministic interaction tied to the OS update.

Community reproductions have zeroed in on a few key variables. The workload almost always involves sequential writes of roughly 50 to 100 GB, though this is an empirical observation rather than a hard threshold. Drives at moderate fill levels (often cited near 60 percent) appear more susceptible, likely because the controller must juggle more complex mapping tables and garbage collection. The common thread is not any one brand but the underlying controller and firmware architecture.

Phison Controllers in the Crosshairs—But Others Too

Early investigation quickly spotlighted drives based on Phison controllers. Several popular NVMe SSDs from Corsair, Kioxia, Sabrent, and others use Phison silicon, making the potential blast radius wide. Phison issued a statement to Tom’s Hardware, acknowledging “the industry-wide effects of the ‘KB5063878’ and ‘KB5062660’ updates on Windows 11 that potentially impacted several storage devices, including some supported by Phison.” The company added that it is “working with partners” to identify affected controller families.

But the issue isn’t exclusive to Phison. Community test logs also contain InnoGrit-based drives and even a handful of conventional hard drives. DRAM-less SSDs, which rely on the Host Memory Buffer (HMB) to borrow system RAM for caching, are overrepresented in failure lists. That points to a possible interaction between the OS memory management changes introduced in the August patch and the way these drives handle buffer allocations under heavy write pressure.

Technical Hypothesis: An OS-Controller Tango

The exact root cause remains under investigation, but the most plausible explanation revolves around a perfect storm of factors. Modern NVMe SSDs are intricate co-engineered systems where controller firmware, host drivers, and OS memory buffering must stay in lockstep. When an OS update alters timing, buffer allocation, or power-state transitions, latent firmware bugs can surface.

With DRAM-less drives, the HMB feature is a prime suspect. These SSDs borrow a chunk of system RAM to store mapping tables and write caches. If the August patch inadvertently changed how that memory region is managed—perhaps through a timing tweak, a leak, or an allocation policy shift—a controller could find itself in an unrecoverable state during sustained writes. The failure fingerprint (drive telemetry goes silent, device drops off the bus) fits a controller lockup, not a simple timeout.

Other hypotheses under examination include a write-cache coherence bug triggered by the specific sequence of commands issued during large sequential transfers. Both Microsoft and SSD firmware teams must cooperate to determine whether the fix belongs in the OS, in controller firmware, or in both.

Scope, Impact, and the Data Loss Danger

Although the bug isn’t triggering on every Windows 11 machine, its potential for data loss makes it high impact. Many affected drives recover after a reboot, but some return with partial data corruption on the volumes involved in the transfer. The worst-case outcome, observed with the SA510 and a few other models, is total drive inaccessibility that persists across reboots and requires vendor RMA or professional recovery services.

Phison’s ubiquity in the NVMe market means that a wide range of consumer and business systems could be exposed. However, exposure does not equal certainty. The bug demands a specific combination of workload, drive fill level, and firmware version—conditions that not every user will hit. Nevertheless, the coordinated acknowledgment from Phison and the volume of independent reproductions elevate this from anecdote to a clear, actionable risk.

Microsoft’s Silence and Vendor Coordination

As of this writing, Microsoft’s official KB page for KB5063878 lists no storage-related known issues. The company did fix a separate WSUS/SCCM installation error for the same package, indicating that deployment telemetry is being monitored. Historically, Microsoft uses Known Issue Rollback (KIR) or safeguard holds to block problematic updates once a pattern is confirmed, but no such measure has been announced yet.

Meanwhile, SSD vendors are the frontline for remediation. Phison says it will provide updates and advisories to partners. End users should expect firmware patches delivered through vendor utilities—not a single universal fix. Corsair, Kioxia, WD, Seagate, and others are likely auditing their SKU lines for compatibility with the patched Windows kernel.

Protecting Your System: Immediate Steps

For consumers and IT administrators alike, the priority is data integrity. Until an official fix arrives, conservative precautions are the smart play.

1. Back Up Critical Files Now

  • Use the 3-2-1 rule: three copies, two different media types, one off-site. Cloud sync, external HDDs, or NAS devices all qualify.
  • Do not rely on the affected SSD as the sole copy of anything you cannot afford to lose.

2. Delay or Pause the August Update

  • If you haven’t installed KB5063878 or its predecessor KB5062660, consider holding off until vendors publish a mitigation.
  • In managed environments, use WSUS, SCCM, or Intune to block the update on at-risk fleets.

3. Avoid Large Sequential Writes

  • On systems where the update is already installed, refrain from transferring very large files (>50 GB) in a single operation to NVMe drives—especially DRAM-less or Phison-based models.
  • Split large archives, game installs, or video exports into smaller chunks to reduce the chance of hitting the failure cliff.

4. Inventory and Monitor Storage Hardware

  • Map out all NVMe SSDs in your environment: note model, firmware version, and controller type (use vendor tools or CrystalDiskInfo).
  • Flag DRAM-less drives and those equipped with Phison or InnoGrit controllers as high-priority for testing and firmware updates.

5. If a Drive Disappears Mid-Write

  • Stop all write activity immediately.
  • Power off the host and move the affected drive to a quarantine system for imaging.
  • Use a forensic tool like ddrescue or FTK Imager to capture a sector-by-sector image before attempting any recovery.
  • Collect system logs (Event Viewer, NVMe traces), note the exact Windows build and KB install dates, and save firmware dumps. These artifacts are essential for vendor RMA requests.
  • Do not reformat or run repair utilities that write to the device; this can aggravate low-level corruption.

The Fix Timeline and What to Watch

The road to resolution is multi-pronged. Phison and affected SSD vendors will likely need days to weeks to reproduce the bug, craft firmware fixes, validate across SKUs, and distribute updates through official channels. Microsoft may simultaneously issue an OS-side mitigation—such as a microcode update, a registry safeguard, or a targeted Known Issue Rollback—that disables the problematic behavior while vendors ship permanent firmware patches.

Keep an eye on vendor support pages for firmware changelogs mentioning “improved compatibility with Windows 11 August 2025 update” or similar language. Furthermore, watch for Microsoft to update the KB5063878 known-issues section or release a standalone advisory through the Windows Release Health dashboard.

Separating Fact from Panic

Amid the flurry of reports, a few claims require extra scrutiny. There is no evidence that all Phison drives will fail—only that they are overrepresented in test logs. The 50 GB and 60-percentfull numbers are practical heuristics, not guaranteed failure points. And early community compilations of affected SKUs are investigative leads, not vendor-confirmed lists. They serve as useful triage aids but shouldn’t be treated as gospel.

The broader takeaway is that modern storage stacks are fragile ecosystems where OS changes can tip latent hardware bugs into active failures. For IT teams, this is a renewed call for rigorous update staging, storage firmware audits, and ironclad backup discipline. For Microsoft and the storage industry, it’s a reminder that whole-platform stress testing—covering sustained I/O, high fill levels, and DRAM-less configurations—must be baked into the development cycle.

Until the coordinated fix arrives, treat every large file transfer on a recently patched Windows 11 24H2 system as a calculated risk. Your data’s safety hinges on the precautions you take today.