USB U0, U1, U2, and U3 Power States Explained

SuperSpeed USB defines link power management for the physical and logical connection between a downstream port and its upstream link partner. Microsoft describes four U states, U0 through U3. U0 is the active state used when the link is carrying traffic; the other states reduce link activity during idle opportunities.

This is a link-level concept, not a direct label for how much power the entire USB device consumes. Microsoft explicitly notes that USB 3.x link power states are decoupled from device power states. A device can remain fully powered while its link uses U1 or U2, so seeing a lower-power link state does not prove that every circuit inside the peripheral has powered down.

Microsoft explains that, once software has configured the link partners for U1 or U2, hardware can enter those states autonomously without software handling every transition. U1 is the shallower standby state with a faster return, while U2 permits deeper link power saving with a longer exit path.

That design lets a SuperSpeed link exploit short idle gaps even while the device remains usable. Microsoft gives isochronous traffic as an example: a device can place the link into U1 or U2 between service intervals. The important boundary is that U1 and U2 reduce link activity; they do not require the whole device to lose functionality or enter a device suspend state.

U3 is not simply “a deeper U2.” Microsoft documents U1 and U2 transitions as hardware-autonomous after configuration, while the transition to U3 is initiated by software after sufficient inactivity. When the link enters U3, the device is required to enter its suspended state, tying U3 to a broader device-power transition.

Microsoft also documents that the USB 3.0 specification does not define a direct U1-to-U3 or U2-to-U3 transition. A parent hub or controller is responsible for returning the link to U0 before sending it to U3. That sequencing is one reason link-state behavior depends on the controller, hub, device, firmware, and operating-system stack rather than only on the peripheral at the end of the cable.

A lower-power link state has to return to U0 before ordinary traffic resumes. Microsoft describes U1 as having a fast exit and U2 as having a slower exit, and its transition documentation shows that devices expose U1 and U2 exit-latency information during enumeration. The host can use those capabilities when deciding whether the states are appropriate.

Do not convert that mechanism into a universal claim that U1 or U2 adds a fixed amount of application latency, lowers SSD throughput, or hurts mouse responsiveness. Whether the link enters a low-power state at all depends on traffic and policy, and sustained transfers keep the link active. Application-visible effects require measurement on the actual device path rather than inference from the state name.

USB 2.0 selective suspend historically ties low-power link behavior closely to suspending the device. SuperSpeed USB added U1 and U2 so the link can save power more frequently without imposing the same device-suspend restrictions. Microsoft describes U1 and U2 as complementing selective suspend rather than replacing it.

U3 is the point where the concepts meet more directly: Microsoft states that the software steps used to send a SuperSpeed link to U3 are the same steps used for USB 2.0 selective suspend, and the device must enter suspended state with U3. Troubleshooting therefore benefits from distinguishing short autonomous U1/U2 transitions from a device being selectively suspended.

A USB topology can contain several links: host controller to hub, hub to another hub, and hub to device. Microsoft documents rules for hub upstream-port U1/U2 transitions based on the states of downstream ports. One peripheral therefore does not independently dictate the power state of every link above it.

This matters when diagnosing a dock or hub. A problem that appears to involve one endpoint can instead involve the hub, controller, firmware, cable path, or a link-state implementation between two ports. Microsoft has documented implementation failures where incorrect U1/U2/U3 handling can prevent expected power savings or even cause a link error and device re-enumeration.

Link power management exists to exploit idle time while preserving normal operation. Microsoft describes U1 and U2 as enabling significant power savings with relatively fast exits, especially compared with requiring a full device suspend for every idle opportunity. Removing those states can therefore increase idle power without changing the negotiated USB data rate.

If a specific USB device repeatedly disconnects, resumes incorrectly, or has vendor-documented power-management problems, changing one power-management setting can be a controlled diagnostic. Preserve the original configuration, verify firmware and drivers, isolate hubs and cables, and confirm whether the symptom changes. A blanket “disable power saving for faster USB” rule is not supported by the link-state mechanism itself.