Microsoft’s August 2025 cumulative update for Windows 11 24H2 (KB5063878, build 26100.4946) has introduced a critical regression that is disrupting live video production workflows around the globe. The patch, which rolled out on August 12, causes severe stuttering, frame drops, and choppy audio in NDI (Network Device Interface) streams when the default RUDP transport is used—particularly in multi‑PC setups relying on display capture. The company confirmed the issue on its Release Health dashboard on August 21 and is pointing users toward a temporary mitigation that requires manual reconfiguration of the NDI receive mode.

For broadcasters, esports production teams, and corporate AV departments, the fallout is immediate and costly. NDI workflows that were rock‑solid before the update are now plagued by intermittent freezing, audio drift, and dropped frames—symptoms that Microsoft says persist “even under low bandwidth conditions.” The workaround, endorsed by both Microsoft and the NDI vendor, is to switch the receive transport from Reliable UDP (RUDP) to legacy UDP or single TCP until a permanent fix ships.

The confirmed problem

The official Release Health entry for KB5063878 is blunt: “Severe stuttering, lag, and choppy audio/video might occur when using NDI for streaming or transferring audio/video feeds between PCs.” Microsoft explicitly calls out streaming applications such as OBS Studio and NDI Tools, and notes that the issue is most pronounced when Display Capture is enabled on the source machine.

Affected platforms include Windows 11 version 24H2 with the August update installed. While Microsoft’s advisory focuses on Windows 11, reports from mixed environments suggest that some Windows 10 servicing branches where companion cumulative updates were shipped may exhibit similar behavior. The common thread is the interaction between the updated networking stack and NDI’s default RUDP transport.

Crucially, the problem does not affect consumer streaming services like YouTube, Netflix, or Disney+. It is confined to local network NDI flows—the kind used to shuttle uncompressed or lightly compressed video between capture, encoding, and monitoring machines. That makes it a niche but high‑impact bug for live production.

Why RUDP breaks after the patch

NDI supports three transport modes with distinct latency and reliability trade‑offs:

  • RUDP (Reliable UDP) – Modern default. Adds sequencing and selective retransmission on top of UDP to keep latency low while handling packet loss.
  • UDP (Legacy) – Pure datagram mode with no delivery guarantees. Lowest overhead, but any packet loss translates into visible artifacts.
  • Single TCP – Ordered, reliable stream. Higher latency potential but robust against loss.

The regression is tightly coupled with RUDP traffic. Microsoft hasn’t released a root‑cause analysis, but symptoms align with a subtle mismatch between the new transport stack and RUDP’s timing heuristics. When the OS changes how it schedules or buffers packets, RUDP’s retransmission logic can oscillate: a few dropped packets trigger a storm of resends, which in turn saturates buffers and causes micro‑pauses—exactly the stutter streamers report. Switching to legacy UDP or single TCP sidesteps this because those modes don’t rely on the same adaptive reliability mechanisms.

Step‑by‑step mitigation

Microsoft and NDI’s vendor recommend manually changing the NDI Receive Mode away from RUDP. The procedure varies slightly between applications, but the principle is identical. Here’s how to apply it on affected machines:

  1. Verify the update. Check that KB5063878 (build 26100.4946) is installed. If it is, assume NDI flows defaulting to RUDP are at risk.
  2. For NDI Tools / NDI Access Manager:
    - Launch NDI Tools or NDI Access Manager.
    - Navigate to Advanced settings or Preferences.
    - Look for “Receive Mode” and change it from RUDP to either “UDP (Legacy)” or “Single TCP.”
  3. For OBS Studio (NDI plugin):
    - Open OBS and edit the properties of your NDI source.
    - In the source’s configuration panel, locate “Receive Mode,” “Transport,” or “NDI Mode.” Select “Single TCP” or “UDP (Legacy).”
  4. Restart and validate. Close and reopen your streaming applications, then run a test capture with Display Capture in the scene. Confirm stutter disappears.
  5. Test low‑latency scenes. The regression is often more visible with Display Capture, so test both normal and low‑latency configurations.

Important caveats:

  • Switching to TCP can increase end‑to‑end latency. Measure your pipeline before going live.
  • Changes may be needed on the receiver only, but if the issue persists, try adjusting the sender’s settings as well.
  • Third‑party capture appliances or older NDI builds may have different menu labels; consult the vendor documentation if the exact control isn’t visible.
  • Some NDI features (like low‑latency modes) are tuned for RUDP and may behave suboptimally under TCP. Treat this as a temporary bridge.

Not the only regression from August’s patch

The NDI streaming bug is part of a broader wave of reliability problems following the August cumulative updates. Microsoft also acknowledged an enterprise deployment regression where systems pulling the update through WSUS or SCCM failed with error 0x80240069. The company quickly deployed a Known Issue Rollback (KIR) and re‑released packages for affected distribution channels.

Separately, community reports surfaced about a more severe storage regression: some NVMe SSDs reportedly disappeared from the OS during sustained large sequential writes—often observed at thresholds around 50 GB. These storage incidents prompted vendor escalations, but at press time they had not been universally reproduced or formally confirmed by Microsoft or SSD manufacturers. Administrators should treat these claims as high‑priority but partially verified alerts; flag any such incidents to vendor support with SMART dumps and event logs.

For IT teams, the clustering of regressions around a single Patch Tuesday underscores the delicate balancing act between security and stability. KB5063878 addressed dozens of vulnerabilities, including flaws that could enable Kerberos escalation and domain compromise. Stripping the update wholesale exposes systems to those risks, so the prevailing guidance remains: apply configuration‑based mitigations rather than rolling back the entire cumulative update.

Operational guidance for production teams

If you run a multi‑PC NDI workflow—whether for broadcast, education, or corporate events—act now:

  • Isolate test rigs. Apply the receive‑mode change on a pair of test machines and run capture scenarios that match your real‑world load. Pay particular attention to latency when switching to TCP; it’s often acceptable but can surprise you if unmeasured.
  • Coordinate with network admins. Changes to transport modes can alter network traffic patterns. RUDP’s retransmission storms might have been masking underlying issues; switching to TCP could expose them. Brief your network team on what to watch for.
  • Avoid uninstalling the update if possible. The August slate plugged critical security holes. If you absolutely must roll back on a production machine, isolate it from enterprise networks and apply compensating controls.
  • Build redundancy. For critical live events, favor a hardware encoder or a local capture chain that doesn’t depend on NDI between machines until the permanent fix lands.

Risk assessment: short, medium, and long term

Immediate risks are highest for live production and classroom capture setups where Display Capture is central. Small studios or single‑PC streamers are less likely to hit this because they don’t route NDI between machines.

Medium‑term risks include update fatigue and erosion of trust in the Windows update pipeline. Frequent emergency mitigations increase administrative overhead and can lead organizations to delay security patches—a dangerous trade‑off.

Long‑term implications point to deeper coordination challenges. Complex interactions between host drivers, OS transport stacks, and device firmware are becoming a recurring theme. Future fixes may require joint engineering across Microsoft and hardware vendors, and update delivery mechanisms (staged rollouts, KIRs) will need to become more granular to catch regressions before they hit production.

What to watch for next

  • A Microsoft technical deep‑dive or a re‑release that directly addresses the RUDP interaction.
  • Updates from NDI’s maintainer and third‑party plugin authors (NDI Tools, OBS NDI plugin) that may change default transport behavior or expose finer‑grained controls.
  • SSD vendor firmware updates or official Microsoft guidance if storage regressions are confirmed.

Until then, the mitigation is straightforward, safe, and endorsed—but it is a stopgap. Creators, IT admins, and AV engineers should monitor the Microsoft Release Health page and vendor advisories daily. A permanent patch that restores low‑latency NDI performance without sacrificing security hardening is the only fully acceptable outcome.

Checklist for administrators and engineers

  • Identify all endpoints with KB5063878 installed.
  • For NDI‑reliant machines, switch Receive Mode from RUDP to UDP (Legacy) or Single TCP and validate.
  • For WSUS/SCCM‑managed fleets with deployment failures, apply the KIR or re‑released package per Microsoft’s guidance.
  • For heavy I/O workloads, adopt a conservative posture: avoid large sequential writes on production NVMe drives until storage regressions are clarified by vendors.
  • Keep a change log and rollback plan for all modified settings.
  • Watch for official fixes from Microsoft and the NDI ecosystem.

The August cumulative update for Windows 11 brought essential security fixes but also a sharp reminder that even minor networking stack changes can unravel professional video pipelines. The workaround is immediate and effective; the permanent solution is now up to Microsoft.