NVIDIA today launched an open source initiative to build a platform for deploying artificial intelligence (AI) applications at the network edge while at the same time moving to integrate its platforms with quantum computing platforms.
Announced at a NVIDIA GTC event, NVIDIA Aerial software includes an Aerial CUDA-Accelerated RAN, Aerial Omniverse Digital Twin (AODT) and an Aerial Framework for building applications that is based on the Compute Unified Device Architecture (CUDA) framework developed by NVIDIA. That framework converts Python code into CUDA code to run on NVIDIA Aerial RAN computer platforms.
NVIDIA CEO Jensen Huang told conference attendees the goal is to provide 5G and 6G base stations that enable AI models based on graphical processor units (GPUs) to run at the network edge. Organizations participating in this initiative include Northeastern University, Virginia Tech, Arizona State University and DeepSig along with researchers from WINSLab and LIDS at MIT.
Additionally, NVIDIA today announced a wireless stack for 6G developed in collaboration with Booz Allen, Cisco, MITRE, ODC and T-Mobile following the launch of an AI-WIN project earlier this year and launched NVIDIA IGX Thor, an industrial-grade platform designed to enable IT teams to build real-time physical AI directly to the network edge. Early adopters include Diligent Robotics, EndoQuest Robotics, Hitachi Rail, Joby Aviation, Maven and SETI Institute.
NVIDIA also revealed the NVIDIA BlueField-4 data processing unit (DPU) capable of supporting 800Gb/s of throughput and enabling high-performance inference processing, reaffirmed its commitment to open source software by making available additional open models for physical AI, robotics and biomedical use cases along with a set of open blueprints for building AI factories.
At the same time, NVIDIA signaled its intention to meld GPUs and quantum processors via NVIDIA NVQLink, a high-speed interconnect that has been adopted in supercomputing labs managed by Brookhaven National Laboratory, Fermi Laboratory, Lawrence Berkeley National Laboratory, Los Alamos National Laboratory, MIT Lincoln Laboratory, Oak Ridge National Laboratory, Pacific Northwest National Laboratory and Sandia National Laboratories.
NVQLink already provides quantum integration with 17 builders of quantum processes and five controller builders. Researchers and developers can access NVQLink through its integration with the NVIDIA CUDA-Q software platform to create and test applications that invoke CPUs and GPUs alongside quantum processors. Technology partners working on NVQLink include quantum hardware builders Alice & Bob, Anyon Computing, Atom Computing, Diraq, Infleqtion, IonQ, IQM Quantum Computers, ORCA Computing, Oxford Quantum Circuits, Pasqal, Quandela, Quantinuum, Quantum Circuits, Inc., Quantum Machines, Quantum Motion, QuEra, Rigetti, SEEQC and Silicon Quantum Computing.
NVIDIA also revealed it is working with General Atomics to build a digital twin for a fusion reactor with support from San Diego Supercomputer Center at UC San Diego School of Computing, Information and Data Sciences, the Argonne Leadership Computing Facility (ACLF) at Argonne National Laboratory and National Energy Research Scientific Computing Center (NERSC) at Lawrence Berkeley National Laboratory, and that in collaboration with Oracle is building an AI supercomputer for the U.S. Dept of Energy (DOE) that will be based on 100,000 NVIDIA Blackwell GPUs.
Finally, Lilly revealed it has deployed the largest, most powerful AI factory based on NVIDIA Blackwell GPUs that is wholly owned and operated by a pharmaceutical company, while Uber is working with NVIDIA to build a mobile network for autonomous vehicles and Palantir Technologies is embedding support for NVIDIA GPUs into its data intelligence platform.
It’s more than clear the pace of AI innovation is only going to continue to accelerate. The only thing left to determine now is how long it will take to realize the business benefits.
