Nvidia has announced Tuesday a new open-source software designed to enhance the efficiency and scalability of AI reasoning models in AI factories called Nvidia Dynamo. This platform has been created with the intent to help service providers grow and increase revenue by optimizing AI inference requests across extensive GPU networks.
Nvidia Dynamo boosts GPU efficiency by managing inference across thousands of GPUs and optimizing different stages of large language model processing separately.
READ: Nvidia rises to second most valuable company after Q4 earnings surge (February 27, 2025)
“Industries around the world are training AI models to think and learn in different ways, making them more sophisticated over time,” said Jensen Huang, founder and CEO of Nvidia. “To enable a future of custom reasoning AI, Nvidia Dynamo helps serve these models at scale, driving cost savings and efficiencies across AI factories.”
Nvidia has collaborated with Perplexity on this. Perplexity CEO Aravind Srinivas on Wednesday said that his artificial intelligence software is set to “do a lot” in its collaboration with Nvidia on Dynamo. He thanked Huang, who praised Perplexity and Srinivas, calling them his “favorite partners.”
https://x.com/AravSrinivas/status/1902181155132309889
Dynamo, according to Nvidia, will be made available via its NIM microservices. It would also be supported in a future release by the Nvidia AI Enterprise software platform. Additionally, Nvidia has partnered with cloud service providers and other vendors such as Oracle, AWS, Microsoft, IBM, and Google Cloud to make its NIM microservices available through their enterprise AI platforms such as OCI, Vertex AI, and Azure AI Foundry among others. Dynamo supports PyTorch, SGLang, and vLLM and enterprises will be able to serve AI models across disaggregated inference scenarios as well.
The announcement was made at NVIDIA’s GTC 2025, held from March 17 to 21 in San Jose, California, showcasing significant advancements in AI and computing technologies. The conference attracted approximately 25,000 attendees with Huang delivering the keynote at the SAP Center to accommodate the large audience.
Some other highlights from the conference include Blackwell Ultra GPUs and Nvidia’s plan to invest billions of dollars in U.S. chip manufacturing over the next four years.