Users who installed Microsoft’s August 12, 2025 cumulative update for Windows 11 24H2 are facing a demonstrable threat to their data, as multiple independent tests have confirmed that sustained large file writes can cause some NVMe SSDs to abruptly disappear, taking in-progress files with them. The failure—linked to KB5063878—is not theoretical; it has been reproduced across enough hardware combinations that the immediate advice from the enthusiast community, system integrators, and affected vendors is unanimous: back up your drives now and stop heavy write workloads until the root cause is patched.

A Routine Update, a Destructive Bug

The August 2025 Patch Tuesday release was meant to deliver the usual security fixes and stability improvements. Instead, within days of KB5063878 reaching general availability, reports flooded forums and social media. Users described a disquieting sequence: during a large file transfer, their NVMe drive would simply vanish from File Explorer, Device Manager, and Disk Management. Vendor utilities could no longer read the drive’s SMART data or controller information. Files being written at the moment of failure were often truncated or corrupted beyond repair. Some systems recovered after a reboot; others required more drastic measures to regain access to the drive.

The triggers were remarkably consistent across reproductions. Write a single file larger than roughly 50 GB, or run a sustained sequential write workload—such as cloning a disk, installing a modern AAA game, or extracting a massive archive—and the drive might drop off the PCIe bus without warning. The behavior was most frequently observed on SSDs using Phison controllers and on certain DRAM-less designs, but community collations also showed isolated incidents on non-Phison hardware, suggesting an interaction problem rather than a single-vendor firmware defect.

How the Bug Strikes: The 50 GB Danger Zone

Modern NVMe SSDs are tightly integrated ecosystems. The Windows NVMe driver, the controller firmware, any DRAM cache or Host Memory Buffer (HMB) allocation, and the NAND management algorithms must all work in lockstep. When Microsoft alters kernel behavior—even subtly, through a monthly cumulative update—controller firmware can hit edge cases that previously lay dormant. The community reproductions strongly point to a flaw triggered when the drive’s write cache and metadata engines are stressed by a long, uninterrupted stream of data. The device’s sudden disappearance from Windows and the loss of SMART telemetry are classic signs of a controller-level hang or firmware state corruption, not a simple timeout.

DRAM-less SSDs that rely on the NVMe HMB to borrow a slice of system memory are especially vulnerable. Prior Windows 11 24H2 updates had already exposed HMB-related fragility on select models, making the August update a continuation of a known friction point. That said, drives with dedicated DRAM were also affected in some tests, though less frequently. The common denominator is the host OS’s changed behavior, which means no controller family can be automatically considered immune without validation.

Vendor Response: Phison Confirms, Silicon Motion Demurs

Phison, whose controllers appear in a wide array of retail SSDs, was quick to acknowledge the reports. On August 19, the company issued a public statement: “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.” Phison said its engineers were investigating and coordinating with partners, but no definitive list of affected firmware versions has yet been released.

Silicon Motion, another major controller supplier, took a different public posture—though not through an official channel. A forum post on TechPowerUp relayed a direct reply attributed to the company: “None of our controllers are affected by the Windows 11 bug.” Windowsnews.ai could not locate a matching press release or support advisory on Silicon Motion’s corporate website to confirm the claim. The company’s public pages carry product announcements and earnings disclosures, not a Windows-update bulletin tied to KB5063878. Until Silicon Motion publishes a formal advisory or a downstream SSD vendor issues firmware guidance, the forum statement remains a provisional data point. Users with Silicon Motion-based drives should not assume immunity and should continue to practice caution.

The Falsified Document and Phison’s Legal Counterattack

As if the technical firefight were not enough, the situation was roiled by a malicious disinformation attempt. A document titled “Phison SSD Controller Issues Summary” began circulating among Phison’s customers, purporting to list affected controller models. Phison swiftly disowned it. In a statement provided to Tom’s Hardware, the company said: “We wish to state unequivocally that the document in question … is neither an official nor unofficial communication from Phison.” The company announced it was pursuing “appropriate legal processes” against those responsible for the forgery. Tom’s Hardware published images of the fake document, which included a table of supposedly impacted controllers—a table Phison declared entirely fictitious. The episode underscores the chaotic nature of the incident and the lengths to which bad actors will go to exploit industry uncertainty.

Microsoft Joins the Fray—But Offers No Immediate Fix

Microsoft told multiple outlets it was “aware of” user reports and asked affected customers to submit telemetry through the Feedback Hub. The company has not yet published a Known Issue Rollback or an emergency out-of-band fix, though the volume of reports makes it likely the issue is actively being triaged by storage stack engineers. For enterprise customers and power users, the lack of an immediate mitigation from the OS vendor means the responsibility for data safety falls squarely on their own shoulders.

What You Must Do Right Now

The steps are straightforward but non-negotiable if you have installed KB5063878 on any Windows 11 24H2 machine with an NVMe SSD:

  • Back up immediately. Use an external drive or a reputable cloud backup service. The corruption risk is real, and a backup made before any heavy write activity is your only guaranteed safety net.
  • Halt massive sequential writes. Avoid game installs, disk cloning, multi-gigabyte extractions, or video rendering to an NVMe target until a fix is confirmed. If you must transfer a large file, split it into chunks under 10 GB with pauses between transfers. This reduces the likelihood of hitting the sustained-write threshold that reproduces the failure.
  • Identify your drive’s controller and firmware. Utilities like CrystalDiskInfo, HWInfo, Samsung Magician, Corsair iCUE, or your drive manufacturer’s own dashboard can reveal the controller family and current firmware revision. Check your SSD vendor’s support page regularly for firmware updates. Any vendor-issued fix will be distributed through these utilities first.
  • If your drive disappears mid-write: stop all writes immediately. Reboot and check whether the device reappears. Collect Event Viewer logs around the timestamp of the failure, take screenshots of Device Manager and SMART readings, and submit the data to your SSD vendor and to Microsoft via the Feedback Hub. Telemetry from real-world failures is critical for root-cause analysis.
  • Consider removing the cumulative update only as a last resort. Advanced users can use DISM or other package-removal tools, but incorrectly uninstalling a servicing stack update can leave a system in an inconsistent patching state. Wait for vendor or Microsoft guidance unless you have professional support.

For system administrators and IT organizations, the advice is even more pointed: pause the rollout of KB5063878 in production environments. Perform representative stress tests—sustained writes on sample hardware—before any wider deployment. Monitor Microsoft’s Windows Release Health dashboard and your SSD vendors’ bulletins for firmware advisories. Report any incidents through official Microsoft support channels so telemetry can be correlated.

Why Full Clarity Remains Elusive

The reluctance of vendors to issue definitive “affected/not affected” lists is frustrating but understandable. SSD behavior depends on a matrix of variables: specific firmware build, NAND configuration, motherboard/UEFI NVMe parameters, Windows driver version, and even the drive’s fill level. A controller family might be vulnerable on one OEM’s implementation and perfectly stable on another. Validating a fix entails verifying every branded SKU that uses a given controller, a process that takes time and cannot be rushed without risking incorrect advice.

Phison’s involvement of external partners suggests that the root cause may span both host and firmware, requiring coordinated changes on both sides. The falsified document episode further complicates matters by injecting noise into the triage process at a critical moment. Nevertheless, the independent reproducibility of the failure by outlets such as Tom’s Hardware, Windows Central, ITPro, BleepingComputer, and Guru3D leaves little doubt that the regression is genuine and dangerous.

The Road Ahead

Resolution will be signaled by a chain of confirmations: vendor advisories listing specific firmware IDs and SKU-level guidance, firmware updates posted through official utilities with release notes referencing the Windows update regression, a Microsoft Knowledge Base article or Known Issue Rollback acknowledging the issue, and independent tests showing the failure no longer occurs on patched combinations. Until that chain is complete, treat every large write to an NVMe drive as a potential data-loss event.

For now, the wisest path is the simplest: verify your backups are intact, postpone disk-intensive tasks, and keep your eyes fixed on your drive vendor’s support page. The industry is working on a fix, but your data cannot wait.