.NET 10 and 11 Preview Benchmarked on PrepStack's Production SaaS
PrepStack, a multi-tenant analytics SaaS, benchmarked .NET 10 and .NET 11 previews against .NET 9 on a live production workload, revealing significant performance and efficiency gains. The Answer Up…
PrepStack, a multi-tenant analytics SaaS, benchmarked .NET 10 and .NET 11 previews against .NET 9 on a live production workload, revealing significant performance and efficiency gains.
The Answer Up Front
For most teams running ASP.NET Core, an upgrade to .NET 10 (the current LTS) is a clear win, offering immediate, measurable performance and resource efficiency improvements with minimal migration effort. PrepStack's real-world data demonstrates tangible benefits in throughput, API latency, and memory footprint. While .NET 11 previews show further promise, they are not yet stable for production. Teams should prioritize the .NET 10 upgrade now and begin piloting .NET 11 in CI environments, always validating performance against their specific application's hot paths.
Methodology
This v0 review draws on the founder's published claims at dev.to, cross-posted from PrepStack's blog; independent benchmarks are pending. Update cadence: re-tested when claims diverge from observed behavior.
PrepStack, a multi-tenant analytics SaaS with approximately 110,000 monthly active users and a peak of 3,200 requests per second, performed a direct comparison of .NET 9, .NET 10 (GA), and .NET 11 (preview). The application comprises around 95,000 lines of C# code. The benchmarks were conducted using the same harness and the same production workload to ensure relevance, avoiding synthetic microbenchmarks. The review covers runtime and JIT optimizations, garbage collection and memory management, ASP.NET Core features (OpenAPI, minimal-API validation), C# language enhancements (C# 14), Native AOT startup performance and image size, and the integration of Microsoft.Extensions.AI. The migration effort from .NET 9 to .NET 10 was also quantified. This review does not cover independent performance verification, long-term workflow impacts, or edge case behaviors beyond what PrepStack observed in their specific production environment.
What It Does
Performance & Resource Efficiency
PrepStack's benchmarks highlight substantial operational improvements from upgrading to .NET 10. They report an 11% increase in throughput per instance compared to .NET 9, alongside a reduction in API p95 latency from 132 ms to 120 ms. Memory usage also saw a notable decrease, with the working set per instance dropping from 415 MB to 380 MB. These gains are attributed to runtime and JIT optimizations, including AVX10.2 support, devirtualization, and loop optimizations, as well as GC/memory enhancements like DATAS right-sizing the heap. Early .NET 11 previews show further, albeit smaller, improvements in these areas.
Developer Experience & Language Features
.NET 10 ships with C# 14, introducing features like the field keyword and extension members. PrepStack claims these allowed them to delete approximately 700 lines of code in their 95,000 LOC codebase, indicating a tangible impact on code conciseness and maintainability. ASP.NET Core also sees built-in OpenAPI and minimal-API validation, streamlining API development and documentation. The Microsoft.Extensions.AI library is becoming a first-class citizen, signaling increased support for AI-driven features within the ecosystem.
Native AOT Improvements
Native AOT, crucial for cold start times and container image sizes, also saw significant advancements. PrepStack reports that .NET 10 reduced AOT cold start times from 84 ms to 61 ms and decreased the AOT image size from 41 MB to 33 MB. .NET 11 previews continue this trend, with further reductions to approximately 55 ms and 30 MB respectively, making .NET an increasingly attractive option for serverless and containerized deployments where startup performance and footprint are critical.
What's Interesting / What's Not
The most interesting aspect of PrepStack's analysis is its foundation in real-world production data. Unlike many benchmarks that rely on synthetic tests, this evaluation on a 110k MAU, 95k LOC multi-tenant SaaS provides highly actionable insights for founders and engineering leaders. The quantified gains—11% throughput, 120ms p95 API latency, 380MB memory footprint—are not theoretical; they represent direct operational efficiency improvements. The low migration effort of 1.5 engineer-days for a substantial codebase is also a critical data point, lowering the barrier to adoption for these performance gains. The C# 14 features, specifically the field keyword and extension members, demonstrate how language evolution can translate into direct code reduction, improving developer productivity and code clarity.
What's less detailed is the specific breakdown of how the runtime and JIT optimizations (AVX10.2, devirtualization, loop optimizations) contribute to the observed gains. While named, a deeper dive into their impact on different workload types would enhance understanding. Similarly, the
The investor read
This benchmark from PrepStack signals a healthy and actively improving .NET ecosystem, particularly in operational efficiency. The reported gains in throughput, latency, and memory usage translate directly into reduced infrastructure costs for high-scale applications, a key driver of tooling spend. The low migration cost (1.5 engineer-days for 95k LOC) makes these efficiency gains highly accessible, suggesting that .NET remains a strong contender for startups prioritizing performance and cost control. The continuous improvements in Native AOT also position .NET well for the growing serverless and containerized deployment landscape. For investors, this indicates that Microsoft's continued investment in .NET provides a robust, performant, and cost-effective platform, making it an attractive choice for new ventures and a sticky technology for existing ones.
Every claim ties to a primary source. See our methodology.