At GDC 2026, Microsoft unveiled two significant updates poised to reshape the gaming landscape: the introduction of ML-Powered DirectX and Advanced Shader Delivery. These innovations are set to enhance gaming visuals and performance, marking a pivotal moment in the evolution of next-generation gaming experiences.
Microsoft DirectX Goes Full ML While Advanced Shader Delivery Solves Windows Gaming Stutters & Load Times
Machine Learning (ML) is emerging as a cornerstone for advancing gaming graphics and performance. Microsoft’s announcement at GDC 2026 highlighted its strategic integration of ML within the DirectX API, paving the way for enhanced capabilities in future gaming applications.
Microsoft DirectX Goes ML
The first iteration of ML-enhanced DirectX is embodied in Cooperative Vectors within Shader Model 6.9, already included in the latest Agility SDK 1.619 release. This feature facilitates various neural rendering techniques, including Neural Texture Compression and Neural Radiance Caching. Notably, NVIDIA has begun supporting these technologies through its Blackwell architecture, while AMD is preparing to incorporate them into its forthcoming architectures, as highlighted during the FSR Diamond announcement.
As the demand for ML continues to rise, the complexity of matrix-matrix operations presents challenges for existing execution models. To address this, Microsoft is introducing DirectX Linear Algebra, a framework designed to accommodate both vector and matrix-based ML workflows under a unified programming model. This innovation grants developers greater control over mathematical computations, data flow, and execution in shader-level ML scenarios.
Another noteworthy addition to the DirectX API is the DirectX Compute Graph Compiler, a new ML compiler API that executes full model graphs with native GPU performance. The key advantages of this feature include:
- Unified tooling with PIX, allowing for the visualization of graphics and ML workloads in a single capture.
- The ability to integrate full models into engines without necessitating shader rewrites.
- Automatic graph optimization, memory planning, and operator fusion.
- Consistent performance across Independent Hardware Vendors (IHVs).
AMD: “DirectX Linear Algebra and DirectX Compute Graph Compiler provide developers with innovative ways to incorporate machine learning directly into their graphics pipelines while maintaining the performance standards expected from modern GPUs. We are eager to collaborate with Microsoft to advance ML-driven graphics on Windows.” – Robert Shearer, CVP Silicon Design Engineering, AMD
Intel: “DirectX Linear Algebra offers developers a robust foundation for integrating matrix-based machine learning into real-time graphics workflows. We are excited to support Linear Algebra from day one.” – Lisa Pearce, Corporate Vice President, Software Group, Intel
NVIDIA: “With DirectX Linear Algebra and DirectX Compute Graph Compiler, developers can seamlessly integrate both shader-level and model-level machine learning into their graphics pipelines. We are pleased to support these capabilities and collaborate with Microsoft to enhance ML-driven rendering and inference workflows on NVIDIA GeForce RTX GPUs.” – Patrick Neill, Distinguished Engineer, NVIDIA
Qualcomm: “DirectX Compute Graph Compiler represents a significant advancement in making full model ML feel native within real-time engines. We are excited to work with Microsoft on a compiler-based approach that optimizes GPU workloads directly integrated into DirectX.” – Balaji Calidas, Senior Director of Engineering, Qualcomm
These enhancements to Microsoft’s DirectX API, including the Compute Graph Compiler, are set to enter private preview this summer, with Linear Algebra slated for public preview in April.
Microsoft’s Advanced Shader Delivery Saves Gamers From Shader Compilation & Long Load Times
In addition to DirectX advancements, Microsoft introduced its Advanced Shader Delivery for Windows, previously available on Xbox ROG Ally and Ally X. This initiative aims to unify the ecosystem among game developers, IHVs, and game stores to address shader compilation issues on PC. Game developers can now enable players to download fully compiled shaders in advance tailored to their specific hardware. With AgilitySDK 1.619, two new APIs have been introduced:
App Identity API:
This API allows applications to declare their identity to D3D12 and the underlying graphics drivers in a standardized manner. Developers can set a default D3D12APPLICATIONDESC and GUID to self-identify prior to creating a D3D12 device. Attaching application identity to the Shader Object Database (SODB) will be mandatory for submitting an SODB file to the Xbox Partner Center.
Stats API:
This API provides game developers with insights into the performance of a precompiled shader database (PSDB). It allows developers to assess how well a given PSDB will function for specific hardware configurations, offering valuable information on shader cache hit rates.
Developers will need to integrate the SODB collection into their game engines and submit it alongside the game package to the Xbox Partner Center to enable Advanced Shader Delivery support. The following statements from hardware vendors reflect the significance of this development:
“Advanced Shader Delivery (ASD) is revolutionizing the gaming experience by reducing load times and eliminating in-game stutter on Xbox ROG Ally devices. The collaboration between Microsoft and AMD engineering teams has yielded remarkable results in a short timeframe.” – Rodney Andre, Corp. VP Software Development
“Intel is dedicated to addressing shader compilation challenges on PC to enhance the overall gaming experience. Microsoft’s Advanced Shader Delivery is a crucial step in minimizing shader load times and compilation stutters, and we are pleased to provide driver support for this feature on our Lunar Lake and Panther Lake platforms.” – Lisa Pearce, Corporate Vice President, Software Group, Intel
“To eradicate the shader-related stutters and load times that have troubled gamers for years, NVIDIA is collaborating closely with Microsoft to launch Advanced Shader Delivery for GeForce RTX users later this year.” – Henry Lin, Director of Product Management, Gaming & AI at NVIDIA
“Advanced Shader Delivery is a vital feature for Qualcomm Snapdragon® compute platforms. By minimizing redundant shader compilation, it enhances the overall gaming experience. We are partnering with the Microsoft DirectX team to debut this feature soon on Qualcomm Adreno™ X2 GPUs.” – Nagendra Kumar, Senior Director of Engineering
NVIDIA has confirmed that it will extend Advanced Shader Delivery support to GeForce RTX users later this year, while AMD and Intel are also working to broaden support for a wider audience in the near future.
Microsoft Preps DXR 2.0 With Shader Model 6.10 Support
In a forward-looking move, Microsoft is also developing the next iteration of DXR (DirectX Raytracing), known as DXR 2.0. According to Microsoft, a device will be classified as DXR 2.0 compliant if it supports Opacity Micromaps and Shader Model 6.10. Partial support will be available in Tier 1.2 for hardware that does not include OMM, while retaining SER.
“This implies support for Shader Model 6.10, which includes a small portion of HLSL features. Shader Model 6.10 also necessitates support for TriangleObjectPositions, which a device could expose independently of the aforementioned features.”
“Features are currently under development, with an initial release anticipated around late summer 2026. The specifications are being shared early, as they align across hardware vendors, though they remain subject to refinement before final release.”
As it stands, DXR 2.0 is projected for an initial rollout in late summer 2026, coinciding with the introduction of the new Shader Model 6.10 support.