Reports are emerging that after updating to Windows 11 26H2, memory usage has dropped and everyday operations feel snappier. On October 5, Windows Latest reported that a three-year-old laptop with 16GB of RAM showed about 2GB less memory in use after the update. That is an encouraging change for PCs where adding memory isn't easy, but a single example doesn't establish that 26H2 uniformly saves 2GB. Separating the responsiveness improvements Microsoft has been rolling out from changes in post-boot memory use makes it clearer what this update should actually be credited with.
Responsiveness improvements began well before 26H2
Looking at Microsoft's official documentation in chronological order, improvements to responsiveness were already described in the May 26 update for 24H2 and 25H2, and carried into the June 9 update. They were not introduced for the first time in 26H2.
The May 26 optional preview update, KB5089573, was described as speeding up app launches and core Windows interactions such as the Start menu and search. Because this change is part of items delivered gradually, PCs that install the same update won't necessarily get the improvement at the same time.
| Official document date | Target / role | Relation to performance improvements |
|---|---|---|
| May 26, 2026, KB5089573 | Optional preview for 24H2 and 25H2 | Lists faster app launches, Start menu, search, and Action Center |
| June 9, 2026, KB5094126 | Monthly update for 24H2 and 25H2 | Rolls in the non-security improvements included in the previous month's preview update |
| 26H2 annual update | Delivered via enablement package to eligible 24H2/25H2 PCs | Enables, on a shared codebase, many improvements already delivered through monthly updates |
This table matches the May update history with the June 9 update details and the list of 26H2 changes. It compares the officially stated targets and delivery methods; it does not show when a feature became active on any individual PC or the actual speed gain.
26H2 uses the same codebase as 24H2 and 25H2, enabling some features that have already been delivered while renewing the support period. The point at which the version name changes with the annual update does not necessarily coincide with when internal improvements actually arrive. As a result, some people may notice a change only after moving to 26H2, while others were already benefiting from the same improvements earlier.
Microsoft also says that in 26H2 it improved launch time and responsiveness of File Explorer's Home page. However, even when update notes say something is "faster," not every PC will get faster by the same amount. How much everyday waiting time shrinks depends on the device configuration and how it is used.
A brief CPU speed boost reduces waiting time during interactions
The "Low Latency Profile," which temporarily raises CPU clock speed right after an app launches, has also drawn attention as a mechanism for improving responsiveness. In its own August 12 testing, Windows Latest set up identical apps on the same PC with a Core i5-13420H and compared before and after the August update. Checking CPU frequency with HWiNFO, it found that after the update the performance cores rose to near their maximum frequency at app launch, and launch times also got shorter.
According to the outlet, the higher frequency lasts about one to three seconds. Processing power is raised only right after an action begins, the necessary work finishes quickly, and the CPU returns to idle. The aim is to shorten the slight wait between a click and the screen responding.
However, the change being confirmed in August refers to that outlet's test machine; it doesn't mean it became active on all PCs on the same day. Microsoft's May explanation already included faster app launches among the improvements.
Microsoft's own explanations also include changes in this direction. In its May 1 quality progress report, it said it had been delivering performance and power-consumption tuning since mid-March for commonly used OS operations and apps. It also modified the scheduler that assigns work to CPU cores and said it would improve responsiveness for everyday tasks by better controlling C-states, the CPU's power-saving states.
One caveat: CPU utilization and clock frequency are different metrics. Even when CPU utilization is low, perceived responsiveness changes depending on what frequency short bursts of work run at and when the CPU wakes from a power-saving state.
Conversely, menus and apps opening quickly doesn't mean video encoding or game rendering performance improves by the same proportion. The effect of the Low Latency Profile is best judged by how much it shortens the wait right after an action starts.
The memory reduction may come from several improvements stacking up
Microsoft has also explained its work to reduce memory usage. In its July 31 progress report, it cited the introduction of more efficient memory allocation to reduce the memory consumed by apps and OS components. In addition to improvements to WinUI 3, used for on-screen display, it said it is also making Chromium and WebView2, used inside the OS, more efficient.
The mechanism that allocates memory apps need, the mechanism that draws the screen, and the components that display web content each operate at different layers. Overall Windows memory use is determined by the accumulation of these components plus the apps actually running. Changing a single feature doesn't determine all memory consumption.
Beyond how much memory is used, work is also under way on how quickly memory that is no longer needed is released. In May, Microsoft described efforts to reduce the widgets' normal memory consumption and release memory sooner when they are not in use.
It has also indicated a policy of expanding the degree to which users can control the pre-launching of apps and features so they can appear instantly, and of restricting it on devices with little memory. At the time, gradual rollout to Insiders was just beginning.
Given this background, it is quite plausible that memory usage improves after updating to 26H2. However, it is not known which change contributed how much to the roughly 2GB decrease reported. That Microsoft is improving memory allocation and display components is a separate matter from whether the effect of each improvement can be calculated from one PC's post-boot figures.
Also, the mechanism that briefly raises CPU clock speed does not directly explain a reduction in RAM use. Measures that improve interaction responsiveness and measures that free up memory need to be evaluated separately, even if both take effect on the same PC at the same time.
"7GB in use" and "4GB cached" can't simply be added together
On the 16GB PC tested by Windows Latest, Task Manager after a reboot showed 7.0GB "In use." The memory capacity recognized by Windows was 15.7GB, "Cached" was 4GB, and "Available" was 8.8GB. These figures show the memory situation of the whole PC at that moment.
The 7GB "In use" is not capacity consumed by Windows alone. It also includes apps and drivers that are resident at startup. The outlet's PC also had several resident apps running, so this is not a comparison of OS-only memory use.
This point is also explained in Microsoft's memory assessment documentation. That document itself covers an assessment tool for older generations of Windows, but it is referenced here for its basic explanation of memory categories and what consumes memory at startup.
Cache is a mechanism that keeps data once loaded in memory so that it can be read faster when needed again. This includes standby memory, which can be reused when an app needs memory. In its cache management documentation, which also covers Windows 11, Microsoft explains that the standby list is treated as part of "available memory."
Therefore, it is not appropriate to simply add 7GB "In use" and 4GB "Cached" and conclude that 11GB is occupied. Part of the free memory is being used as cache so that needed data can be retrieved quickly. A large cache value alone does not mean there is little memory headroom for apps.
The observation that "In use" memory fell before and after the update is meaningful. However, if the state of resident apps or app versions changes, overall PC memory usage changes too. A single measurement right after a reboot cannot support the conclusion that Windows 11 26H2 uniformly makes things 2GB lighter.
To verify the effect, compare using your usual workload
To check for improvement on your own PC, you need to measure before and after the update under conditions that are as identical as possible. Restart each time, log in, and wait the same amount of time before comparing. Keeping resident apps and power settings the same, and noting the versions of the apps used for comparison, makes it easier to judge any change.
If you stop resident apps to reduce memory use, the effect of that change will mix into the results. To see the difference due to the OS update, it is important to keep the environment the same as usual.
It also helps to think separately about what you measure. "In use" and "Available" after boot give a rough sense of how much memory headroom there is for running apps. The time it takes to open the Start menu or apps indicates responsiveness during operation. And running the browsers and work apps you normally use at the same time lets you check whether headroom has increased in your real working environment.
It is difficult to express a PC's overall "speed" with a single number. You need to look separately at whether memory use fell, whether waiting time during operation shortened, and whether headroom grew when using multiple apps at once.
In its July report, Microsoft listed optimization for PCs with 8GB or more of memory as one of its future priorities. How much benefit these efforts bring, on PCs with what amounts of memory and for what kinds of work, will continue to need comparative testing.
Even on laptops where memory can't be upgraded, if waiting times get shorter with your usual apps running together and available memory still has headroom, OS-side improvements may extend the period over which the PC remains comfortable to use.