Windows 11 cumulative update KB5063878 is being investigated by storage controller maker Phison after users reported that certain NVMe SSDs vanish mid-transfer during sustained writes exceeding 50GB, sometimes leading to permanent data loss. Released on August 12, 2025, as OS Build 26100.4946, the patch was a routine security and quality rollup for Windows 11 24H2. Microsoft’s release notes initially stated the company was “not currently aware of any issues with this update.” Within days, however, independent testers and multiple hardware outlets documented a consistent and alarming storage regression.

Background: A Routine Patch Turns Destructive

KB5063878 delivers the combined servicing stack and cumulative quality fixes for version 24H2. Typical for such updates, it landed automatically for most users via Windows Update. The trouble first came to light when X user Nekorusukii (@Necoru_cat) attempted to update Cyberpunk 2077 on a system running 24H2 with the patch installed. Mid-transfer, the target SSD disappeared from File Explorer and Device Manager, derailing the operation. Nekorusukii soon replicated the behavior and published findings, triggering a wave of community reproductions and media coverage.

Tom’s Hardware, Guru3D, Wccftech, and Ubergizmo all picked up the story and confirmed the symptom profile through their own testing or by collating reports. The pattern is strikingly consistent: during a large sequential write – typically 50 GB or more – the SSD simply drops off the bus, as if physically removed. Reboots can sometimes bring the drive back, but files being written at the moment of failure are corrupted or incomplete, and in a small number of cases the drive remains unrecoverable.

What Users Are Reporting

The symptom profile breaks down into a few precise elements. First, the trigger: sustained large sequential writes. Real-world examples include game downloads, mass file copies, cloning operations, and video exports. Second, the failure mode: the SSD disappears from Windows entirely. It vanishes from File Explorer, Disk Management, and Device Manager. Vendor diagnostic utilities and SMART telemetry become unreadable, indicating a low-level controller stall rather than a simple driver glitch. Third, the aftermath: a reboot may restore visibility, but any data in-flight during the incident is often lost. In the worst cases, the drive stays inaccessible even after power cycling, requiring vendor recovery tools or RMA.

Nekorusukii’s rigorous testing provides a crucial data point. Out of 21 SSDs tested on his own system, 12 became inaccessible during large writes. All but one – a Western Digital SA510 2TB – could be recovered after a reboot. That single unrecoverable drive highlights the very real risk of permanent data loss. Another user, running a SanDisk Extreme Pro M.2 NVMe 3D SSD, encountered the bug repeatedly in August while installing a 50 GB game update. Deleting the earlier KB5062660 update eliminated the problem, strongly implicating the Windows update pathway.

Which Drives Are Affected?

Community-compiled lists and independent tests show that the issue is not universal but concentrates around certain controller families and firmware revisions. Phison-based drives are heavily represented in reproduction attempts, leading Phison itself to acknowledge the incident. However, other controllers have surfaced in isolated reports, so a blanket claim that only Phison is affected would be premature.

Drives that have been named in the wake of these incidents include:
- Corsair Force MP600 (Phison PS5012-E12 variants)
- KIOXIA Exceria Plus G4
- Fikwot FN955
- SanDisk Extreme PRO M.2 NVMe 3D SSD

These models are best treated as investigative leads rather than a definitive blacklist. Firmware revision, drive capacity, use of Host Memory Buffer (HMB), motherboard BIOS, and storage driver versions all influence whether a given drive will trigger the bug. Not every unit of a listed model will fail, and some drives from other controller families have also exhibited similar behavior.

Plausible Technical Mechanisms

Although a formal root cause analysis remains outstanding, several hypotheses have gained traction among storage experts and community testers.

Host Memory Buffer (HMB) Allocation Behavior – Many DRAM-less SSDs rely on HMB, a PCIe feature that lets the drive borrow a small slice of system memory for its mapping tables. Windows manages HMB allocation, and a subtle change in how the OS grants or revokes that memory could push a controller into an unhandled state. Prior Windows 11 updates have exposed HMB timing sensitivities, making this a plausible contributor.

Kernel or Storage Stack Regression – The cumulative update may have altered how Windows buffers or sequences NVMe commands during sustained sequential I/O. Even a small shift in command timing or buffer pressure can expose latent firmware bugs that were never triggered under previous hosting behavior. Such regressions manifest as controller stalls, not ordinary driver crashes.

Controller Firmware Lock‑Ups Under Sustained Stress – Extended large writes stress NAND management, mapping-table updates, wear‑leveling, and garbage collection within the SSD’s controller. If the firmware contains a race condition or exception handler flaw that only appears under prolonged load, the controller can hang and cease responding to the host, causing the drive to appear removed.

Taken together, these mechanisms paint a picture of a host‑side behavior change (introduced by KB5063878) that inadvertently exposes long‑dormant firmware bugs in certain SSDs. The sudden disappearance, unreadable SMART, and occasional unrecoverable state all align with a controller‑level hang.

Phison and Microsoft Respond

Phison issued a public statement to Tom’s Hardware on August 19, 2025: “Phison has recently been made aware of the industry-wide effects of the ‘KB5063878’ and ‘KB5062660’ updates on Windows 11 that potentially impacted several storage devices, including some supported by Phison. We understand the disruption this may have caused and promptly engaged industry stakeholders. … At this time, the controllers that may have been affected are under review and we are working with partners. We will continue to provide updates and advisories to partners who may have been impacted to provide support and ensure any applicable remediation.”

This marks a significant escalation from purely community‑driven troubleshooting to an official industry investigation. Microsoft’s KB article for KB5063878, however, still does not list any storage‑device issue as a known problem. The software giant has not yet published an advisory or Known Issue Rollback (KIR). Enterprise administrators can use servicing controls to pause or rollback the update in managed environments, but consumers lack such direct levers.

SSD vendors and OEMs are monitoring the situation. In past incidents of this nature, the fix has typically come either as a firmware update from the drive maker or a Microsoft mitigation such as a KIR. Given Phison’s commitment to partner advisories, firmware updates are the likely long‑term resolution.

Scope and Risk: Narrow but High‑Impact

So far, the problem appears targeted rather than widespread. Reports are technically consistent and reproducible in controlled tests, but they account for a tiny fraction of the overall Windows 11 NVMe install base. Nevertheless, the risk profile is severe: low probability in the general population, but extremely high impact for anyone affected. A drive that becomes unrecoverable can mean complete loss of files, game libraries, or work projects that were not backed up.

Anyone who has already installed KB5063878 and uses an NVMe SSD should assess their risk and take immediate precautions.

Immediate Defensive Steps for Windows Users

  • Back up critical data right now. Copy important files to a second physical disk or reliable cloud storage. Do not rely on the at‑risk SSD as the primary backup destination.
  • Avoid sustained large writes until a fix is available. That means pausing large game downloads, mass file copies, cloning operations, or lengthy video exports that target the suspect SSD.
  • Check your SSD vendor’s support utility. Tools like Corsair iCUE, SanDisk Dashboard, or Kioxia/Crucial software may have firmware advisories. Only apply firmware updates after a full backup – the update itself could trigger the bug if it performs a large write.
  • In enterprise environments, stage the update in a test ring. Validate a representative workload (50 GB+ sustained write) against hardware that matches your production inventory. Use Microsoft’s deployment controls to pause or rollback the patch where risk is unacceptable.

What to Do If a Drive Disappears Mid‑Write

If your drive suddenly becomes invisible during a file transfer:

  1. Stop all write attempts immediately. Do not initialize, reformat, or attempt to repair the disk – that can destroy partition metadata and reduce recovery chances.
  2. Reboot the system once. In many cases, power‑cycling restores the drive temporarily, but repeated reboot/write cycles raise the risk of permanent damage.
  3. Run vendor diagnostics, if possible. The drive may still appear to diagnostic tools even when not visible in Windows. Collect SMART data and controller logs for support cases.
  4. If the drive remains inaccessible, create a block‑level forensic image. Use read‑only tools such as dd from a Linux live environment or a hardware imager. This preserves the exact state for vendor recovery and RMA claims.
  5. Engage vendor support with reproduction logs, firmware revision, motherboard/BIOS version, and workload details. Vendors will often request images and logs before accepting RMA tied to firmware behavior.

Evidence Strengths and Open Questions

The current body of evidence is compelling but incomplete. On the plus side:
- Independent reproducibility: Multiple outlets and hobbyist testers have replicated the exact symptom fingerprint under similar workloads, ruling out isolated hardware faults.
- Vendor acknowledgement: Phison’s public statement confirms the industry‑wide nature of the problem and validates community reports.
- Technical plausibility: The controller‑hang scenario aligns with known interactions between host storage stacks and SSD firmware, especially on DRAM‑less drives.

However, significant gaps remain:
- No definitive root cause: Neither Microsoft nor SSD vendors have published a full forensic analysis linking KB5063878 to a specific regression. The exact host‑side change – be it HMB timing, NVMe command sequencing, or something else – is still a well‑supported hypothesis, not a confirmed fact.
- Incomplete affected‑drive list: Community compilations are useful for triage but not exhaustive. Firmware revisions, OEM customizations, and system variables materially alter susceptibility.
- Potential confounders: Motherboard firmware, PCIe link training, and thermal throttling can mask or exaggerate the issue. Variability outside carefully controlled test benches makes broad generalizations difficult.

Anyone claiming “only Phison controllers are affected” should be cautious. Phison‑equipped drives are over‑represented, but other controllers have surfaced in isolated reproductions. The safest position is to assume risk for any NVMe SSD running the patched Windows build until vendor telemetry clarifies the scope.

What to Watch for Next

Three developments will clarify the path forward:
1. Vendor firmware advisories and updates. Watch your SSD manufacturer’s support page for targeted firmware that addresses controller‑level behavior under Windows 11 24H2.
2. Microsoft release health updates. An eventual Known Issue entry or KIR would provide official mitigation and confirm the regression.
3. Wider independent verification. Large‑scale testing across motherboard, BIOS, and CPU platforms will reveal whether the bug is strictly a controller/firmware issue or a host OS timing regression that interacts with specific controller families.

Conclusion

KB5063878 presents a credible and actionable risk for a narrow but important slice of Windows 11 users. Sustained large sequential writes – a routine operation for gamers, content creators, and power users – can trigger SSD controller stalls that make drives disappear and, in rare cases, cause permanent data loss. The convergence of community testing, specialist outlet reproductions, and Phison’s official investigation elevates this above mere rumor.

For now, the pragmatic stance is defensive: back up, avoid heavy writes on recently patched systems, and stage the update in test rings. Apply vendor‑issued firmware fixes only after safeguarding your data. This incident reinforces a hard truth about modern NVMe storage: reliability depends on an intricate dance between firmware, drivers, and OS behavior. When a routine security patch disrupts that choreography, the consequences can be immediate and irreversible. Stay protected, stay informed, and coordinate with Microsoft and your SSD vendor for authoritative remediation as it becomes available.