A critical storage regression in Microsoft’s August 2025 cumulative update for Windows 11 24H2 (KB5063878) is causing NVMe and SATA SSDs to vanish from the operating system—and in some cases from the system firmware—during sustained write operations. Multiple community testers have reproduced the bug, with at least one drive suffering irreversible data loss after repeated trigger attempts. The issue has drawn an official response from NAND controller maker Phison, but Microsoft has yet to acknowledge the storage problem on its support page.

The Bug: How SSDs Disappear Mid-Transfer

The failure mode is alarmingly consistent. Users performing large sequential writes—installing a game, copying multi-gigabyte archives, or extracting compressed files—watch their target SSD drop from File Explorer, Device Manager, and even the platform firmware. Restarting the PC sometimes recovers the drive, but in several documented cases the device remains invisible through multiple reboots, rendering its data permanently inaccessible.

Initial public attention began when X (Twitter) user Nekorusukii posted a reproducible sequence: updating Cyberpunk 2077 on a system with KB5063878 installed caused the SSD hosting the game to disappear. Subsequent experiments expanded the dataset across 21 drives, with 12 becoming inaccessible during heavy writes. The standout casualty was a Western Digital SA510 2TB, which could not be recovered even after a power cycle and reboot, marking a permanent data loss event.

Reproduction steps have been published by multiple testers and independently verified by technology outlets. The typical trigger involves writing 50–100 GB of sequential data—such as copying a large game folder, creating a 60+ GB compressed archive, and then decompressing it directly on the target SSD. The bug strikes most frequently when the drive is around 60% full, though testers have triggered failures at different utilization levels, indicating the percentage is an observed pattern rather than a strict precondition.

Affected Hardware: Beyond Phison Controllers and DRAM-less Designs

Early reports flagged Phison-based controllers as disproportionately represented, prompting Phison’s own investigation. However, community testing quickly demonstrated the problem is not confined to a single controller vendor or to DRAM-less SSDs. The test pool included drives from multiple manufacturers with different controller architectures—some with onboard DRAM—and failures occurred across the spectrum. This suggests an interaction between Windows’ storage stack and a subset of controller firmware implementations, rather than a singular broken SKU.

Not all drives are susceptible, and the bug does not strike in every write-heavy scenario. The mixed results across the 21-drive test—with roughly half affected—highlight that this is neither universal nor fully predictable. Still, the inclusion of DRAM-equipped SSDs in the failure list rules out simplistic explanations that pin the blame solely on Host Memory Buffer (HMB) caching in budget drives.

Vendor and Microsoft Responses: Phison Acknowledges, Microsoft Remains Silent

Phison issued a measured statement to Tom’s Hardware, confirming it is “aware of the industry-wide effects” of KB5063878 and KB5062660 and that “controllers that may have been affected are under review.” The company emphasized collaboration with partners and promised future advisories. This official acknowledgment—though short of a root-cause admission—validates the community reports and signals that firmware-level fixes may be forthcoming from drive manufacturers.

Microsoft, conversely, has not listed any storage regression on the support page for KB5063878. As of publication, the page states that “Microsoft is not currently aware of any issues with this update.” The company has, however, tracked and resolved unrelated problems with the same update—such as a Group Policy-based Known Issue Rollback addressing WSUS/SCCM installation errors—so its silence on the SSD failures is conspicuous.

Mainstream technology outlets including Tom’s Hardware, Guru3D, Neowin, and TechRadar have independently reproduced or aggregated user findings, adding weight to the community’s alarm. The growing body of evidence puts pressure on both Microsoft and SSD vendors to accelerate their investigations.

Technical Deep-Dive: Why SSDs Vanish from BIOS, Too

Modern SSDs depend on a fragile choreography between operating system drivers, the NVMe command stack, and device firmware to manage mapping tables, garbage collection, and caching. A common theory centers on HMB—a feature that lets DRAM-less SSDs borrow system memory for their mapping tables. Under sustained write loads, any timing perturbation or cache-coherency flaw can trigger a firmware crash that makes the controller unresponsive.

But the presence of failures in DRAM-equipped drives points toward a broader issue. The August update may have altered I/O flushing, caching behavior, or storage driver timing in a way that some firmware versions cannot tolerate. A subtle race condition or corrupted command path could cause the controller to enter a locked state or corrupt on-drive metadata.

The most severe cases—where the SSD disappears from the system BIOS/UEFI—indicate that the controller has stopped responding to enumeration or has corrupted its NVMe namespaces beyond platform firmware recovery. This is a hallmark of controller-level metadata damage, not a simple software hiccup, and it explains why some devices suffer permanent data loss. When a drive is only invisible to Windows but still visible to firmware, a reboot or power cycle often restores access. Losing firmware visibility, however, signals a deeper corruption that professional recovery tools may struggle to address.

Risk Assessment: Who Should Be Most Concerned

Desktop and laptop users who store critical data on a single SSD without recent backups face the highest risk of permanent loss. The bug has already claimed at least one drive irreversibly in community testing. Systems that routinely perform large sequential writes—game installs, content creation scratch disks, local backup targets—are especially exposed because the trigger centers on sustained write IO.

Enterprise and IT administrators must consider latent fleet-wide failures. Even a low failure rate could have outsized operational impact if large batch jobs are executed on affected machines. Systems with older or unpatched SSD firmware remain particularly vulnerable, as vendor firmware may contain unaddressed edge-case bugs that the Windows update now exposes.

Immediate Mitigation: Steps to Protect Your Data

The following practical measures prioritize data safety while vendor investigations continue:

  • Back up critical data now. Create full copies of irreplaceable files on a separate, known-good storage device—external drive, network share, or cloud. Do not perform large write operations on at-risk SSDs until the situation clarifies.
  • Pause installation of KB5063878 if it has not already been applied. Use Windows Update’s “pause updates” feature for individual machines, or hold the patch in WSUS/SCCM for managed environments. Note that the cumulative includes a servicing stack update, complicating uninstall paths, so delaying it is the safest course.
  • Avoid heavy sequential writes to potentially affected drives. Postpone large game updates, archiving, cloning, and bulk copy tasks targeting suspect media until a fix emerges.
  • Update SSD firmware and motherboard BIOS where vendor updates are available, but only after confirming the update addresses this specific issue. Check manufacturer support pages and apply firmware updates carefully, keeping copies of current firmware for rollback.
  • If a drive becomes inaccessible, do not write more data to it. Writing after corruption increases the chance of permanent loss. Power down the system and consider professional data recovery services for valuable information.

Recovery Options: What Works and What Doesn’t

Recovery outcomes vary widely depending on the severity of the corruption. Soft attempts—restarting the machine, power-cycling the drive (removing and re-seating an M.2 module or cycling a SATA/HDD), or testing the device in a different port/system—restore access in some cases but provide no guarantee. When the drive remains invisible to firmware, these steps rarely succeed.

Some SSD vendors offer specialized utilities to reinitialize firmware or repair metadata. These tools can regain visibility but often risk data loss; they should only be used after creating a forensic image or under vendor guidance. For business-critical or irreplaceable data, engage a reputable data recovery service immediately and avoid repeated power-on attempts that could overwrite salvageable metadata.

What Remains Unknown and What Comes Next

Community tests pinpoint KB5063878 (and the July KB5062660) as a common trigger, but the exact causal chain remains unverified in a controlled, vendor-backed lab. The update may alter OS timing, storage driver behavior, Defender caching, or another kernel I/O surface that exposes firmware defects. Without Microsoft’s or a manufacturer’s formal post-mortem, the observations are strong warning signals—not a definitive root-cause diagnosis.

The scope of affected devices is also unclear. Current reporting suggests a minority of drives reproduce the failure, but the real-world percentage across global hardware permutations cannot be estimated without vendor telemetry. The conservative posture is to consider all susceptible models at risk until firmware updates or Microsoft guidance declares otherwise.

Remediation will likely require coordinated action. SSD vendors can push firmware updates that harden controllers against the OS behavior, while Microsoft can adjust driver, NVMe stack, or Defender components to avoid triggering fragile controller states. Phison’s engagement signals movement on the storage side, but no timeline has been announced.

Conclusion

The August 2025 Windows 11 24H2 cumulative update (KB5063878) has introduced a high-severity, low-frequency storage regression that can cause SSDs to vanish during heavy writes—occasionally with permanent data loss. Independent reproduction, vendor acknowledgment, and widespread media coverage all justify immediate caution: back up critical data, pause the update, avoid heavy write workloads on affected machines, and await coordinated remediation from Microsoft and SSD manufacturers.

Until firmware patches or an operating system fix materializes, the safest posture for home users and administrators alike is to assume risk exists and act preemptively. This incident underscores the brittle boundary between modern OS storage stacks and intricate SSD firmware—a boundary that demands fast, coordinated response when large-scale updates shift timing and behavior.