I manage a small fleet of about 30 users with a mix of Lenovo P16 Gen 2 laptops with NVIDIA graphics and mostly T16 Gen 3 models using Intel graphics. Users periodically report that networking stops working and only a reboot seems to fix it. Pings and DNS still work, but TCP-based connections such as Teams, OneDrive, web browsing, and PowerShell connectivity tests fail or time out. In some cases the display or graphics driver appears to hang, freezing the entire Windows interface. Windows occasionally recovers, but the system remains in a broken state where network applications and even UAC prompts do not respond. The problem is much more common when laptops are connected to Lenovo Thunderbolt 4 Universal Docks and external monitors, with wake-from-sleep—especially Modern Standby—being a frequent trigger. NVIDIA models tend to hard-freeze more severely, although Intel-only systems also experience the issue. Updating BIOS, firmware, graphics, dock, and other drivers has helped somewhat but has not eliminated it. Has anyone found a reliable root cause or mitigation for similar Lenovo laptops, docks, and Intel AX211 Wi-Fi adapters?
5 Answers
Do not overlook the graphics stack itself. Driver conflicts from Windows Update, Lenovo update tools, or dock display components can leave the desktop frozen while the kernel continues answering ICMP. Try a controlled graphics-driver rollback or clean installation, prevent Windows from replacing the tested package, and compare NVIDIA and Intel systems with the same dock and monitor setup. On some Lenovo models, updating less-obvious platform components such as the Intel NPU driver has also resolved freezes triggered by Teams camera use.
Dock and display paths can trigger graphics-driver failures, especially with DisplayPort adapters, older monitors, KVMs, or mismatched dock firmware. Try reproducing it with the laptop undocked, with Wi-Fi and wired networking separately disabled, and with native monitor cables instead of HDMI-to-DisplayPort adapters. Updating the dock's USB-C or Thunderbolt firmware and the graphics firmware may help; some systems have needed a motherboard or GPU replacement when the freezes continued.
I have seen similar Lenovo behavior after sleep where the NIC does not wake correctly and applications using persistent connections break. Temporarily setting the machines to performance mode and disabling sleep can help isolate the cause. A clean Windows installation with only required Lenovo drivers is another useful test, since Windows Update and vendor utilities can sometimes install a conflicting graphics or network driver over the approved package.
A practical workaround is disabling power management for the network adapters. In the adapter's advanced properties, test disabling checksum offloading and any sleep or power-saving options. Some administrators have also disabled NIC sleep through their device-management policies. Reconnecting the dock cable or disabling and re-enabling the adapter can temporarily restore connectivity, which points toward a bad resume or dock state.
This sounds like a combination of dock, graphics, and power-state problems rather than a simple loss of network service. Since ping still works while outbound TCP fails, capture traffic with Test-NetConnection and a packet analyzer while the machine is in the broken state. Also confirm whether it happens after Modern Standby or another sleep state, and make sure the BIOS, dock firmware, graphics drivers, and Wi-Fi drivers are all current.
The failures appear to happen after resume from Modern Standby (S0ix). Firmware and drivers are currently up to date, but I have not yet captured traffic from a machine while it is stuck.
The laptops use Intel AX211 Wi-Fi rather than wired Ethernet, but I will test the offload and power-management settings there as well.