Chipmaker Nvidia and AI infrastructure firm Palantir Technologies announced a compact data centre system for AI on Thursday. Called the AI OS Reference Architecture (AIOS-RA) — this suite of tools is intended to help enterprises speed up the deployment of full-scale AI applications, claimed an official statement shared on Thursday.
The AIOS-RA is is designed with Nvidia Blackwell Ultra, touted as the most powerful AI hardware in the market. The product additionally features eight Nvidia Blackwell Ultra GPUs to process AI training tasks.
The system is also capable of connecting large language models (LLMs) to organizational data and operational systems.
According to Akshay Krishnaswamy, the chief architect of Palantir, this AI operating system is optimized for Nvidia’s accelerated compute infrastructure.
Essentially, enterprise clients using Nvidia’s GPU and other tools could use the AIOS-RA to skip the manual selection of servers, installation of complicated AI software, and crack ways to keep data privacy intact. The AIOS-RA is being presented as a pre-tested solution ready for sovereign cloud deployment
“From our first deployment with the United States government, our software has had to meet the moment in the most complex and sensitive environments where customers must maintain control. We are proud to deliver a fully integrated AI operating system for Nvidia,” said Krishnaswamy.
Headquartered in Florida, Palantir was launched in 2004 by Peter Theil and Alex Karp as defense contractor in the tech and intelligence sectors. It debuted its AI platform called AIP in April 2023 and has been making inroads in the private sector now.
A footage of Nvidia CEO Jensen Huang with Palantir CEO Alex Karp is making the rounds on X. In conversation with CNBC, Huang called Palantir the world’s most technically deep enterprise platform, predicting that its collaboration with Nvidia would only supercharge Palantir’s work.
Nvidia has noted that AI’s infrastructure stack requires data-sovereign environments, free of latency — which this AIOS-RA is intended to deliver.
“By combining Palantir’s sovereign AI OS reference architecture with NVIDIA AI infrastructure, industries and nations can turn data into intelligence with speed, efficiency, and trust,” said Justin Boitano, vice president, Enterprise AI Platforms, Nvidia, commenting on the development.
Nvidia is now foraying into the agentic AI space with the planned launch of NemoClaw, an open-source AI platform.