James Ding
May 19, 2025 20:12
NVIDIA and Microsoft collaborate to enhance AI applications, introducing breakthroughs from cloud to PC, with significant advancements in AI inferencing and development tools.
NVIDIA and Microsoft are enhancing their collaborative efforts to advance agentic AI applications, bridging the gap between cloud and PC environments. This initiative marks a significant stride in AI innovation, according to NVIDIA’s blog.
Microsoft Discovery: Revolutionizing Research
At the Microsoft Build event, the tech giant unveiled Microsoft Discovery, a versatile platform designed to transform research processes with agentic AI. This platform is expected to expedite product development and enhance the scientific discovery timeline across various industries. It incorporates NVIDIA’s ALCHEMI NIM microservice, optimizing AI inference for chemical simulations, and NVIDIA BioNeMo NIM microservices, which streamline AI model development for drug discovery.
Microsoft Discovery’s capabilities were demonstrated through a novel coolant prototype identification, accomplished in under 200 hours, a task that traditionally takes months or even years.
Scaling AI Infrastructure with NVIDIA GB200
To further boost AI capabilities, Microsoft is deploying NVIDIA GB200 NVL72 rack-scale systems in its Azure data centers. These systems, featuring up to 72 NVIDIA Blackwell GPUs per rack, significantly enhance performance and efficiency, delivering up to 35 times more inference throughput compared to previous models.
These advancements are supported by custom server designs, NVIDIA NVLink interconnects, and NVIDIA Quantum InfiniBand networking, which facilitate seamless scalability for demanding AI applications.
Enhanced AI Reasoning on Azure AI Foundry
Building on existing NIM integrations, NVIDIA and Microsoft are expanding Azure AI Foundry with the NVIDIA Llama Nemotron models and BioNeMo NIM microservices. These additions deliver advanced reasoning capabilities and are tailored for complex decision-making and domain-specific AI workloads, offering up to 20% higher accuracy and five times faster inference.
The healthcare sector is set to benefit from these advancements, with applications in digital biology, drug discovery, and medical imaging, aiming to improve patient care and accelerate scientific innovation.
Generative AI on Windows 11
NVIDIA and Microsoft are also focusing on enhancing generative AI experiences on Windows 11 PCs. At Microsoft Build, they introduced an AI inferencing stack to boost development and performance for Windows 11, leveraging NVIDIA TensorRT for RTX AI PCs. This integration promises seamless AI deployment across over 100 million RTX AI PCs, enhancing tools for developers and creative professionals.
As these technological advancements unfold, the collaboration between NVIDIA and Microsoft is expected to set new benchmarks in AI-powered applications, streamlining processes from scientific research to everyday PC applications.
Image source: Shutterstock
>