Phison SSD Failures After Windows 11 August Update (KB5063878)

Phison has publicly acknowledged the storage failures linked to Microsoft’s August 2025 Windows 11 cumulative updates (identified as KB5063878 and related KB5062660) and says it is investigating the reports with industry partners while controllers that “may have been affected” are under review.

Background / Overview​

Windows Update released an August 2025 cumulative package for Windows 11 (commonly tracked as KB5063878, OS Build 26100.4946) that was intended to deliver security and quality fixes as part of Patch Tuesday. Within days of the rollout, multiple independent community tests and specialist outlets reported a repeatable failure profile: during sustained sequential writes — community reproductions commonly cite workloads on the order of ~50 GB or more — certain storage devices stop responding, vanish from Windows (File Explorer, Device Manager, Disk Management), and in some cases return with corrupted or inaccessible data. (tomshardware.com, itpro.com)
Phison — a major SSD controller designer whose silicon is used across many consumer and OEM drives — issued a concise statement acknowledging the “industry-wide effects” of KB5063878 and KB5062660, saying it had engaged stakeholders and was reviewing potentially affected controller families while promising partner advisories or remediation as applicable. That acknowledgement elevated the issue from a scatter of forum posts to an industry incident that vendors and administrators must treat seriously. (wccftech.com, overclock3d.net)

---

What users and testers are reporting​

Symptom fingerprint​

• Large, sustained sequential write operations (game installs, mass file copies, archive extraction, cloning, or video exports) proceed normally and then abruptly fail after tens of gigabytes — many accounts converge on the ~50 GB mark as a repeatable trigger. (tomshardware.com, borncity.com)
• The target SSD (or occasionally an HDD) disappears from the OS: it is no longer visible in File Explorer, Device Manager, or Disk Management. SMART and vendor telemetry often become unreadable during the failure. (windowscentral.com, notebookcheck.net)
• In many cases, a reboot temporarily restores device visibility. In a smaller set of reports, the drive remains inaccessible after reboot and partitions or files written in the failure window are corrupted or lost. (neowin.net, techradar.com)

Typical trigger profile​

Community lab reproductions consistently point to sustained sequential writes — continuous large writes that exercise controller metadata paths and caching — as the immediate trigger. The phenomenon appears workload-dependent: splitting the same data transfer into smaller chunks often avoids the failure in some test rigs. Testers commonly observed higher failure likelihood on drives that were already moderately filled (for example, >50–60% used capacity). (windowsforum.com, borncity.com)

Which hardware shows up in early lists​

Community collations and independent outlets have repeatedly named drives built around certain Phison controller families (examples cited include PS5012‑E12 and later E21T/E31T lineage) as being over‑represented in early reproductions. Multiple branded consumer SKUs (Corsair MP600 and other vendors’ drives) appeared repeatedly in community lists, while several independent reproductions also reported non‑Phison examples — meaning the dataset is heterogeneous and not a clean vendor-only recall. These model lists are community-sourced and evolving; they are useful for triage but are not authoritative until vendors publish validated inventories. (tomshardware.com, notebookcheck.net)

---

Technical anatomy: why a Windows update can expose SSD instability​

Modern NVMe SSDs are embedded systems where controller firmware, the host NVMe driver, optional on‑board DRAM, NAND channels, and host OS behavior operate in tight coordination. Two technical concepts help explain why an OS-side change can precipitate controller-level failures:

• Host Memory Buffer (HMB) and DRAM-less SSDs: Many cost-optimized SSDs omit on‑board DRAM and rely on the NVMe HMB feature to borrow a portion of system RAM for mapping tables and caching. Subtle changes in how the OS allocates or uses HMB — or timing changes in command ordering and DMA — can expose firmware race conditions on DRAM‑less designs. Earlier Windows 11 24H2 rollouts already surfaced HMB-related fragility for some drives, establishing a plausible precedent for host-driven triggers. (borncity.com, windowsforum.com)
• Controller firmware edge cases under sustained load: Sustained sequential writes stress mapping-table updates, garbage collection, and internal metadata writes. If a host-side timing or buffer behavior changes — for example altered NVMe write ordering, staging of buffered I/O, or DMA timing introduced by a cumulative update — latent firmware assumptions can be violated. The controller may hang, stop responding to NVMe admin commands, or corrupt its internal state, rendering the device invisible to the OS until a firmware reset or power cycle. The observed unreadable SMART and vendor telemetry after failures supports a controller hang hypothesis in many reproducible cases. (tomshardware.com, guru3d.com)

These mechanisms are not mutually exclusive — the most likely real-world root cause is an interaction between OS behavior and specific controller firmware code paths. That is why multiple vendors, including Phison, and Microsoft must collaborate on telemetry-driven diagnostics.

---

What Phison said — and what it implies​

Phison’s public wording was measured: the company confirmed it had been “recently made aware of the industry-wide effects” attributable to KB5063878 and KB5062660, that it had “promptly engaged industry stakeholders,” and that “controllers that may have been affected are under review” while Phison works with partners on remediation and advisories. The statement promises updates to partners rather than a direct consumer-channel commitment, which is typical when a controller supplier must coordinate firmware fixes with each branded SSD vendor and validate firmware per SKU. (wccftech.com, overclock3d.net)
Two practical implications of Phison’s posture:

• Firmware fixes (if required) will almost certainly be delivered via SSD vendors’ support channels (vendor dashboards, vendor update tools) rather than directly from Phison to end users. Vendors must test controller firmware against their device’s BOM, factory configuration, and their own update mechanisms before public release.
• If root cause analysis shows the host OS behavior is the primary trigger, Microsoft may still need to issue an OS-level mitigation (Known Issue Rollback, targeted block, or a hotfix), either alone or in parallel with vendor firmware updates. The resolution path may therefore include both firmware updates and Microsoft mitigations depending on forensic telemetry. (windowsforum.com, windowscentral.com)

---

Who should fix what — firmware or Windows?​

This is the central practical question for remediation. There are three realistic outcomes:

• Controller firmware fix required — If analysis shows controller firmware fails to handle a legitimate NVMe command cadence or HMB allocation pattern introduced by the Windows update, SSD vendors will need to publish firmware updates for affected SKUs. These updates are distributed via vendor tools and must be validated by vendors for each model and manufacturing flash ID. Phison’s partner-centric statement strongly suggests this scenario is under active consideration. (wccftech.com, notebookcheck.net)
• Microsoft host mitigation required — If telemetry shows the update changed timing, buffering, or NVMe command ordering that violates established expectations, Microsoft can deliver a targeted update or Known Issue Rollback to restore previous host behavior while vendors prepare firmware patches. Microsoft has used KIR previously for enterprise delivery problems and host-regression mitigations.
• Hybrid approach — The most likely real-world outcome is a combination: Microsoft issues a temporary mitigation or guidance while SSD vendors release validated firmware updates for specific controller/firmware SKUs. This dual path minimizes short-term exposure while ensuring long-term robustness. (tomshardware.com, guru3d.com)

Until vendors and Microsoft publish a coordinated root-cause bulletin, users and administrators must act on conservative triage: backup first, avoid heavy writes on patched systems when possible, and monitor vendor support channels for firmware advisories.

---

Practical mitigation: what users and IT teams should do now​

These recommendations are conservative and prioritized to protect data:

• Back up important data now to an independent device or cloud. Backups are the only reliable defense against mid‑write corruption or permanent device inaccessibility.
• If you have not yet installed KB5063878/K5062660 and your workflow includes heavy writes or you use suspect hardware, consider staging the update in a test ring or delaying deployment until vendors/Microsoft provide assurance for your hardware. Enterprises should withhold the patch for production rings until validated.
• Avoid sustained, large sequential writes (> ~50 GB) on Windows 11 systems that have the August update installed, especially on drives known or suspected to use Phison controllers or DRAM‑less designs. Where possible, split large transfers into smaller chunks. (techradar.com, borncity.com)
• Check your SSD vendor’s update utility (Corsair iCUE, SanDisk Dashboard, Kioxia/Crucial utilities, etc.) for firmware advisories and install vendor‑validated firmware only after taking a verified backup. Firmware updates can fix controller bugs but carry their own small risks; follow vendor instructions closely. (notebookcheck.net, guru3d.com)
• For fleet/enterprise admins:
• Inventory drives, controller families, and firmware revisions across endpoints.
• Use WSUS/MECM/SCCM to stage or block KB5063878 where risk is unacceptable; monitor Microsoft release health and vendor advisories.
• Add sustained sequential write tests (~50 GB or more) to validation suites and run them on representative hardware before broad rollout.
• If a drive disappears mid-write:
• Stop writing to the drive immediately; power down if safe.
• Capture Event Viewer logs and vendor telemetry, and create a forensic, read-only image if data is critical (dd, ddrescue, or vendor imaging tools).
• Contact the SSD vendor’s support channel and escalate for RMA/data-recovery options if necessary.

---

Strengths of the current evidence — and its limits​

Strengths:

• Independent reproducibility: multiple hobbyist testers and specialist outlets reproduced a highly consistent symptom profile under similar workloads, lending credibility to the reports. (tomshardware.com, guru3d.com)
• Vendor acknowledgement: Phison publicly confirmed it is investigating, which validates that the community reports map to real, observable failures in some environments.
• Technical plausibility: the failure fingerprint (drive disappears mid-write; unreadable SMART; partial recoverability) is consistent with a controller hang or firmware-level lockup that could be exposed by host-side timing changes.

Limitations and open questions:

• No definitive public root-cause yet: at the time of reporting neither Microsoft nor SSD vendors have published a full forensic root-cause analysis tying the August cumulative update to a precise kernel or firmware fault. That means attribution to a single controller family or to Windows alone is unproven publicly.
• Community model lists are provisional: early lists of “affected models” are community-sourced and vary between test benches because firmware revisions, motherboard BIOS, and platform differences change exposure. Treat them as investigative leads, not definitive recalls.
• Reproducibility varies: not every drive with a named controller fails in every rig, which suggests that firmware version, drive fill state, host drivers, and BIOS/UEFI differences all influence whether a given device will reproduce the failure.

Because of those limits, any sweeping public claims that “all Phison drives are bricked by KB5063878” are incorrect; the real situation is nuanced and requires telemetry-validated analysis.

---

Why this matters — the systemic lesson for Windows and storage ecosystems​

This incident is a textbook example of the fragility that can arise in co‑engineered systems where OS updates alter low-level runtime behavior that embedded firmware did not previously encounter. Modern NVMe SSDs — particularly DRAM‑less designs that offload mapping structures to host memory — depend on implicit timing and allocation assumptions. A seemingly benign host update can change that choreography and exercise latent firmware race conditions.
The result is operational: data loss, disrupted workflows, and trust erosion for users and IT teams. The technical remedy is well known (coordinated telemetry, firmware fixes, and, where needed, host mitigations), but the organizational and communication challenge is greater: cross-vendor coordination under time pressure and clear, timely guidance to end users and administrators. Phison’s acknowledgement is the opening of that coordination, not its conclusion. (tomshardware.com, guru3d.com)

---

What to watch for next​

• Vendor firmware advisories or model-specific firmware updates from SSD brands (Corsair, Kioxia, SanDisk, ADATA, etc.). Firmware is the likeliest permanent fix if controller logic proves to be the causal factor.
• Microsoft release-health updates, an explicit Known Issue entry for KB5063878/KB5062660, or a KIR/block for enterprise channels if OS-side mitigation is required.
• Consolidated telemetry or whitepapers from one or more large test labs that correlate failure rates to specific controller FW revisions and host configurations — such reports will help administrators craft precise compatibility holds.

---

Final assessment and recommendations​

The available evidence supports three clear conclusions:

• A reproducible storage regression exists that manifests under sustained, large sequential writes on some Windows 11 systems after the August cumulative update(es). The failure fingerprint is strong and has been independently reproduced. (tomshardware.com, itpro.com)
• Phison has publicly acknowledged the reports and is investigating with partners; that strongly suggests affected controller firmware families are an active focus for remediation. However, Phison’s statement does not prove causation and stops short of definitive attribution.
• Practical, urgent steps are straightforward and should be prioritized: back up data, avoid heavy writes on patched systems where possible, stage updates in test rings for fleets, and apply vendor-validated firmware only after backup. These steps reduce immediate risk while remediation proceeds. (neowin.net, guru3d.com)

Administrators and cautious consumers should treat the situation as an urgent compatibility warning: the probability of encountering problems in the general installed base is low-to-moderate, but the impact of a failure (data loss, unmounted partitions) is high. That combination demands conservative actions now and careful validation before broad update deployment.

---

Phison’s acknowledgement marks the start of coordinated analysis and remediation rather than an immediate fix. The most likely remediation path will involve vendor firmware updates validated per SKU — and possibly Microsoft-side mitigations if host behavior is shown to be a co-contributor. Until those actions are published, the defensible posture for users and IT teams is clear: protect data, pause heavy writes on at-risk systems, and monitor vendor and Microsoft advisories for validated firmware or host mitigations. (overclock3d.net, windowsforum.com)

---

Source: OC3D Phison responds to Windows 11 SSD failure controversy - OC3D