Is Windows 11 26H2 raising the bar for 8GB machines?

Table of Contents

With the release of Windows 11 26H2, Microsoft has announced several performance and memory optimizations specifically targeting systems with 8 GB of memory, “Raising the bar for 8 GB Windows 11 PCs”. According to Microsoft, these improvements deliver up to 20% better responsiveness and up to 40% faster multitasking under specific testing conditions.

With the ongoing global memory shortage, extending the lifecycle of existing devices or upgrading to systems with lower hardware specifications is becoming more relevant for organizations. These improvements could therefore have a significant impact, particularly on systems with limited memory.

But do Microsoft’s bold claims translate into measurable improvements? This research explores the performance and memory efficiency of Windows 11 26H2 on systems with 8 GB of memory to determine whether the promised optimizations deliver real benefits.

Statement from Microsoft on Windows 11 26H2

In addition to these performance improvements stated in the introduction, Microsoft has introduced several memory-management optimizations. These include reducing committed-memory growth, improving memory compression and paging behavior, and keeping more active data in physical memory rather than relying on slower storage. In theory, this should mean that Windows will be less likely to page those memory pages to disk.

It is important to note that these improvements do not necessarily translate into a reduction in overall physical memory utilization, as memory efficiency can also be improved by reducing paging activity and managing available memory more effectively.

While Microsoft has introduced optimizations at the operating system level, several key applications have also been optimized to reduce resource consumption and improve the overall end-user experience. Microsoft specifically highlights Microsoft Edge and Microsoft 365 applications, including Word, PowerPoint, Excel, Outlook, and Teams. Microsoft Teams can turn on Efficiency Mode by default on 8 GB PCs, helping it launch up to 20% faster and reducing the system resources it uses while it runs.

The operating system improvements are available starting with Windows 11 26H2 build 26300.9818. Improvements to Microsoft Edge, Microsoft 365, and Microsoft Teams are delivered separately through their regular update channels or the Microsoft Store.

Since the summer of 2026, Microsoft has emphasized its commitment to improving the overall quality of Windows. Memory optimization is one of the key focus areas in this release, but Microsoft has indicated that additional improvements are planned for future updates.

  • Windows features and apps will share common parts instead of each carrying its own. Search and Widgets will share one web engine, and so will OneDrive and Photos for memory savings across all four apps.

  • Apps and experiences built on WinUI 3, including File Explorer, Notepad, Paint, and the new Windows search are being designed to open faster and use less memory.

  • Together with PC makers and silicon partners, Microsoft is optimizing how much memory a PC needs to reserve for its hardware, leaving more available for Windows and apps.

  • Microsoft is also working with third-party app developers to help their apps use memory more efficiently.

For all the details, please read the initial commitment announcement and the latest blog post about the 26H2 release.

Setup and scenario

This research aims to evaluate whether the optimizations introduced in Windows 11 26H2 translate into measurable performance and memory efficiency improvements under a standardized EUC workload in a virtual desktop.

While Microsoft reports up to 20% better responsiveness and up to 40% faster multitasking, these results are based on specific testing scenarios and primarily on telemetry Microsoft collected. By comparing Windows 11 25H2 and 26H2 using OBUX, this research provides an independent perspective on whether these improvements are reflected in overall system performance and the end-user experience.

The following scenarios are included in this research:

  • Windows 11 25H2, as the baseline
  • Windows 11 26H2

Because Microsoft’s claims focus on systems with 8 GB of memory, this research uses an Azure Standard D2as v7 virtual machine. This configuration provides 2 vCPUs based on AMD’s 5th Generation EPYC™ 9005 processors, 8 GB of memory, and a 512 GB Premium SSD.

The benchmark used for this research is OBUX version 0.1.5.94. Because OBUX measures both overall system performance and the end-user experience, it is well suited to evaluating whether the claimed improvements translate into measurable benefits. The standard GO-EUC testing methodology is applied, with ten test runs per configuration to ensure consistent, reliable results.

OBUX uses its own set of applications and workloads. Therefore, any optimizations specifically introduced for Microsoft’s key applications, including Microsoft Edge, Microsoft 365, and Microsoft Teams, are not covered by this benchmark. As a result, this research focuses primarily on identifying performance and resource utilization improvements at the operating system level rather than validating application-specific optimizations.

Hypothesis and retrospective

Based on the previous Windows 11 26H2 preview research by Leee Jeffries, the improvements in system resource utilization were limited. The memory footprint was reduced by a maximum of just 1%, which is nowhere near Microsoft’s claims. This research used a configuration with 8 vCPUs and 16 GB of memory, running a preview build from July. Therefore, it is possible that the specific optimizations mentioned by Microsoft were not yet included in this build.

For this research, it is expected to see some improvements in both responsiveness and memory consumption. However, Microsoft’s measurements are based on specific scenarios, and it is not entirely clear how these improvements were measured. As Microsoft explicitly states that these figures are not based on a benchmark, it is not expected that the results will match the published numbers directly. Nevertheless, this research should provide insight into whether these optimizations translate into measurable improvements.

Results

Let’s first start with the OBUX scores, which provide insight into both the overall system performance and the end-user experience. The benchmark distinguishes between the system score, which reflects how efficiently the system completes computational tasks, and the UX score, which measures the responsiveness of applications during user interactions.

The OBUX results show a very modest improvement in overall system performance with Windows 11 26H2 compared to Windows 11 25H2. The system score increased from 1137 to 1159, showing an improvement of only 1.93%.

Although the increase is measurable, the difference is relatively small. This seems to suggest that the optimizations introduced in 26H2 provide some benefits at the operating system level, but these do not translate into a substantial increase in overall system performance under the standardized EUC workload used in this research.

However, the UX Score shows no measurable difference between the two versions. While Microsoft claims an improvement in responsiveness, this is not reflected in the OBUX results, which specifically measure how quickly users can interact with applications.

It is worth noting that the OBUX UX Score is based on application responsiveness and is specifically designed to measure response times during user interactions with applications. Any improvement in system responsiveness should therefore result in a higher score.

For more details about the UX Score, please refer to the OBUX documentation.

Let’s dive into the system performance metrics to better understand these results.

The CPU footprint has been reduced in Windows 11 26H2, which helps explain the improved system score compared to 25H2. Throughout the entire benchmark, CPU utilization is consistently lower in 26H2, with an average reduction of 7.5%.

Memory utilization remains identical between Windows 11 25H2 and 26H2, with no measurable reduction observed during the OBUX benchmark.

However, Microsoft’s memory optimizations are not necessarily focused on reducing overall physical memory consumption. Instead, several improvements target memory efficiency by reducing committed-memory growth, optimizing paging behavior, and keeping more active data in physical memory rather than relying on slower storage.

To better understand whether these optimizations have an impact, look beyond overall memory usage and validate paging activity. The Windows Performance Monitor metric Memory Pages/sec measures the rate at which memory pages are read from or written to storage to resolve hard page faults. While this activity does not necessarily indicate a shortage of physical memory, it provides additional insight into how Windows manages memory during the benchmark.

It is important to note that while overall memory usage remained identical between Windows 11 25H2 and 26H2, there was a measurable difference in paging activity.

Memory Pages/sec measures the rate at which Windows reads memory pages from or writes them to disk to resolve hard page faults, making it a useful indicator of paging activity and its potential impact on system responsiveness, particularly when evaluating memory efficiency improvements. Windows 11 26H2 recorded an average of 3112 pages per second, compared to 3536 on Windows 11 25H2. This is a reduction of approximately 12%, suggesting that fewer pages needed to be transferred between physical memory and storage during the benchmark.

These optimizations could potentially improve system responsiveness and multitasking performance, particularly under memory-constrained conditions. However, as overall memory utilization remained unchanged and no measurable improvement was observed in the OBUX UX score, the results do not demonstrate a direct benefit to the end-user experience under the tested conditions.

Overall, disk activity is very comparable between Windows 11 25H2 and 26H2, with no significant differences observed. While disk reads are slightly higher in 26H2, the difference is minimal and is unlikely to have any noticeable impact on the end-user experience.

During the benchmarking, the number of running processes is also collected. This metric is particularly relevant, as a reduction in the process count could help explain a lower CPU load. However, this is not the case, as the average process count remains exactly the same. This suggests that Microsoft has introduced optimizations that reduce CPU utilization without affecting the number of running processes in the operating system.

Conclusion

Is Microsoft raising the bar for systems with 8 GB of memory? Based on this research, Windows 11 26H2 does not showcase any significant measurable improvements and does not improve the end-user experience.

Microsoft’s statements cannot be confirmed. Furthermore, the blog post does not provide sufficient information to determine what data these claims are based on.

The most notable improvement is the reduction in CPU utilization, which is on average 7.5% lower compared to Windows 11 25H2. This is accompanied by a slightly improved OBUX system score, while the average number of running processes remains unchanged. However, the OBUX UX score shows no improvement, and both overall memory consumption and disk activity remain identical between the two versions.

Interestingly, Windows 11 26H2 also shows a reduction of approximately 12% in the Memory Pages/sec metric. This suggests less paging activity during the benchmark and could indicate improvements in memory efficiency. As paging operations can introduce additional CPU overhead, this reduction might also contribute to the lower CPU utilization observed in Windows 11 26H2. However, the collected metrics do not establish a direct relationship between these findings.

What does this mean for organizations considering Windows 11 26H2? While the reduced CPU utilization and paging activity are positive developments, the results do not provide sufficient evidence to justify reducing memory allocations or switching back to systems with 8 GB of memory purely based on the upgrade. Organizations should validate performance and memory requirements using their own representative workloads before making such decisions.

Have you already upgraded to Windows 11 26H2 and noticed improvements in performance or memory consumption? Please let us know!

Photo by Harrison Broadbent on Unsplash