Windows 11 24H2 Storage Regression: Backup Now, Vendor Guidance

Silicon Motion has told at least one forum poster that “none of our controllers are affected” by the Windows 11 storage regression that surfaced after Microsoft’s August 2025 cumulative update — but the larger picture remains unsettled and users should treat that single claim cautiously while they back up and avoid heavy writes until vendors and Microsoft publish confirmed guidance.
What happened (short version)

• On August 12, 2025 Microsoft shipped the combined Servicing Stack Update + Latest Cumulative Update for Windows 11 24H2 (identified in community reporting as KB5063878). Within days, multiple independent testers and enthusiast sites reproduced a repeatable failure mode: during sustained, large sequential writes (community reproductions commonly cited a threshold around ~50 GB), some NVMe SSDs momentarily vanish from Windows (File Explorer, Device Manager, Disk Management), vendor utilities lose access to SMART/controller telemetry, and files written during the event can be truncated or corrupted. In some cases a reboot restores the drive; in a smaller set of reproductions the drive remained inaccessible. (windowscentral.com)
• Early community collations showed a disproportionate number of affected drives using Phison-based controllers and some DRAM‑less designs, prompting Phison to publicly acknowledge it was investigating possible “industry‑wide effects” related to the August update. At the same time, hands‑on testing and aggregated lists showed that the issue was not strictly limited to one controller vendor, and a number of non‑Phison devices also appeared in isolated reproductions. (tomshardware.com)
• Following those reports, the situation became an industry triage: Microsoft said it was “aware of” reports and asked affected customers to submit telemetry through the Feedback Hub while vendors (most visibly Phison) began collecting data and coordinating with partners. (bleepingcomputer.com)

What Silicon Motion reportedly said — and what we can confirm

• A TechPowerUp forum thread (linked by the user) includes posts relaying a statement attributed to Silicon Motion indicating none of the company’s controllers were affected by the Windows 11 bug referenced in community reports. That thread is one of several community collations that summarize vendor responses, test reproductions, and mitigation guidance.
• I could not find a matching, formal press release or support advisory on Silicon Motion’s official press/news pages or corporate channels confirming that public assertion. Silicon Motion’s press pages and recent newsroom items list product announcements and earnings, not a Windows‑update advisory tied to KB5063878. That absence does not prove the forum post is false, but it does mean the claim appears to be a direct reply in a community thread rather than a documented vendor bulletin on Silicon Motion’s official site. Readers should therefore treat the forum-sourced statement as an informative lead that needs vendor confirmation through official channels or direct support contact. (siliconmotion.com) (ir.siliconmotion.com)

Why this matters (technical anatomy)

• Modern NVMe SSDs are tightly coupled systems: host OS kernel + NVMe driver + NVMe controller firmware + NAND media + optional DRAM or Host Memory Buffer (HMB). Small changes in host behavior (timing, buffer allocation, command ordering) can expose firmware edge cases on some controllers under heavy sustained I/O. Community reproductions strongly indicate the failure manifests during long, continuous writes that heavily exercise controller caching and metadata paths. That operational fingerprint — device disappearing from the PCIe/NVMe topology and unreadable SMART telemetry — points at a controller-level hang or firmware state corruption triggered by a host interaction.
• DRAM‑less controllers which use the NVMe Host Memory Buffer are particularly sensitive to changes in how the host allocates or times shared memory; prior Windows 11 24H2 rollouts have already shown HMB-related fragility on select models, so HMB/DRAM‑less interactions are a plausible vector. Early reproductions over‑represented Phison families and some DRAM‑less designs, but testing also included controllers from other suppliers and several non‑Phison drives showed issues in isolated runs — so the root cause appears to be an interaction rather than a simple single‑vendor firmware bug.

Cross-check with multiple independent sources

• The phenomenon has been covered and reproduced by multiple specialist outlets and community test benches (Tom’s Hardware, Windows Central, ITPro, BleepingComputer, Guru3D, NotebookCheck among them). These outlets performed hands‑on reproductions or aggregated community data and reached similar conclusions about the workload trigger (sustained, large writes), the symptom set (drive disappears, telemetry unreadable), and the early vendor involvement (Phison publicly investigating). That independent agreement across outlets increases confidence that a real regression exists even as forensic root cause work continues. (tomshardware.com, windowscentral.com, itpro.com)

What to do right now — practical guidance for end users
1) Back up immediately

• If you’ve installed the August 12, 2025 cumulative update (KB5063878) and rely on any NVMe/SSD for important files, create verified backups now (external drive or cloud). Data integrity is the priority because the failure mode can corrupt files written during the incident. Multiple outlets and community threads emphasize that backups are the single most important mitigation. (windowscentral.com)

2) Avoid big, continuous writes until clarified

• Don’t run large game installs, disk clones, multi‑GB archive extractions, or long sequential copies to at‑risk drives until vendors or Microsoft publish validated mitigations or firmware updates. Splitting large transfers into smaller chunks reduces exposure to the sustained‑write trigger many testers reproduced.

3) Identify your controller and firmware level

• Use manufacturer utilities (CrystalDiskInfo, vendor dashboard, HWInfo, Samsung Magician, Corsair iCUE, etc.) to identify the drive model and the controller family. Vendor firmware dashboards are the distribution channel for controller firmware updates; if a vendor posts a fix it will usually be on their utility or support page first. If your drive uses a Silicon Motion controller and you saw the TechPowerUp/forum post, check Silicon Motion official channels and your SSD vendor for confirmation and firmware packages before assuming you’re safe. (treexy.com, siliconmotion.com)

4) If you hit the bug: stop writing, collect logs, contact support

• If a drive disappears mid-write: stop further writes, reboot and check whether the device reappears, and collect logs (Event Viewer entries near the time of failure, screenshots of Device Manager, SMART readings). Submit the details to the SSD vendor’s support and to Microsoft via Feedback Hub/Support; vendors and Microsoft asked users to provide telemetry and diagnostic data to help reproduce and triage. (bleepingcomputer.com)

5) Consider uninstalling the LCU only as a last resort (and only if you know what you’re doing)

• Some community mitigations for update-related regressions use DISM /Remove-Package or other advanced uninstall methods to remove the LCU; these are delicate and can break system patching if done incorrectly. Prefer waiting for vendor guidance or consulting support unless you have firm instructions from Microsoft or your vendor.

What system administrators and organizations should do

• Stage or pause large deployments: hold off wide deployment of KB5063878 in production rings until vendors and Microsoft publish confirmed mitigations or validated firmware. Perform representative stress tests (sustained writes) on sample hardware and firmware revisions before a full rollout.
• Monitor vendor dashboards and Microsoft Release Health: Microsoft has asked for telemetry and continues to investigate; vendors (Phison prominently) are coordinating with partners. Expect firmware updates published by SSD vendors and possible Microsoft mitigations (Known Issue Rollback or hotfix) depending on root‑cause attribution. Submit organizational incidents via Microsoft support channels so telemetry can be correlated. (bleepingcomputer.com)

How credible is Silicon Motion’s “not affected” claim?

• If Silicon Motion indeed told a forum poster that none of their controllers are affected, that is a useful datapoint — especially for users whose drives actually use Silicon Motion controllers — but a single forum reply is not the same as an official vendor advisory or an audited telemetry statement. At the time of writing I could not find an official Silicon Motion press release or support bulletin corroborating the forum-sourced claim on Silicon Motion’s public newsroom or support pages; their public pages instead carry product announcements and earnings disclosures. Until Silicon Motion publishes a formal advisory or an SSD vendor posts tested, SKU‑specific firmware status, it’s prudent to treat the forum reply as provisional and continue to follow backup/avoid‑heavy‑writes guidance. (siliconmotion.com, ir.siliconmotion.com)

Why vendors may be cautious (and why the problem is hard to summarize as “one vendor at fault”)

• The event’s reproducibility depends on combination of firmware revision, drive fill level, sustained‑write profile, motherboard/UEFI NVMe settings, and OS driver/patch level. That complexity means a controller family could be implicated in some configurations and unaffected in others; it also explains why community lists varied by tester. Vendors therefore prefer to confirm affected firmware IDs and validated SKUs before posting public advisories, because firmware fixes must be tested per branded module, not just per controller chip. That validation process slows public statements but reduces risk of incorrect or incomplete advice.

What we know about Phison and Microsoft responses (for context)

• Phison publicly acknowledged it was investigating the issue and working with partners; Phison later disowned a circulated, falsified internal document that purported to list affected controllers and indicated it had taken legal steps regarding that falsified material. Microsoft said it was “aware” of reports and asked affected customers to submit Feedback Hub reports and Support tickets while it worked with storage partners to reproduce and diagnose the failure. Those vendor and platform responses are consistent with a cross‑stack incident that needs both host and controller telemetry to root cause. (tomshardware.com, bleepingcomputer.com)

How we’ll know the issue is resolved

• The right evidence will be:
• Vendor advisories listing confirmed affected firmware IDs and SKU-level guidance (or confirming no affected SKUs where applicable).
• SSD vendor firmware updates posted through named vendor utilities plus release notes tying the fix to the Windows update regression.
• A Microsoft KB/Release Health entry or Known Issue Rollback when host behavior is implicated (or a follow-up Windows patch if Microsoft changes how the storage stack behaves).
• Independent reproducibility tests demonstrating the failure no longer appears on fixed combinations.
• Until that chain of confirmation exists, the safe posture is conservative: back up, avoid heavy writes, and keep firmware/BIOS/driver channels monitored.

Final takeaways (quick bullets)

• The August 12, 2025 Windows 11 cumulative (KB5063878) is associated with a reproducible storage regression in community tests; heavy sustained writes are the common trigger and the symptom is the device vanishing mid-write with a risk of data corruption.
• Phison acknowledged and is investigating; other vendors and Microsoft are working together to triage. Multiple specialist outlets have reproduced or aggregated the failures independently. (tomshardware.com, windowscentral.com)
• A forum post (TechPowerUp thread) relays Silicon Motion saying its controllers are not affected; however, there is no equivalent formal advisory on Silicon Motion’s official channels at the time of writing, so treat the forum claim as provisional and verify via official vendor support pages or direct vendor contact before assuming immunity.
• Immediate user action: back up, stop large continuous writes, check drive firmware and vendor support pages, and submit telemetry to Microsoft and your SSD vendor if you experience the failure. (bleepingcomputer.com)

If you want, I can:

• Walk you step‑by‑step through how to identify your SSD’s controller and firmware (HWInfo/CrystalDiskInfo screenshots and what to look for).
• Draft a short message you can send to your SSD vendor’s support that includes the exact logs and event‑viewer excerpts useful for triage.
• Monitor and summarize official vendor advisories (Phison, Silicon Motion, WD, Kioxia, SanDisk) and Microsoft Release Health and return with a clear status update when formal confirmations or firmware downloads appear.

Which of those would be most helpful right now?

---

Source: TechPowerUp Silicon Motion: None of Our Controllers Affected by the Windows 11 Bug