What changed
NVIDIA's Holoscan platform is now being used to process massive streams of sensor data from the Allen Telescope Array, allowing AI agents to steer the instrument, enhance and correct incoming data, and apply AI models to the incoming stream — all from a single natural language request. This marks a significant advancement in the search for extraterrestrial intelligence, as AI agents can now perform tasks that would take humans weeks or years in just hours.
What this means in practice
The system, which includes DAQIRI, a Holoscan platform library, enables the telescope to pull terabits per second of sensor data straight into GPU memory for real-time processing. CyberEther, a signal-processing framework developed at the SETI Institute, uses NVIDIA CUDA-X libraries to perform the necessary computations. Nexus, a web-based observatory orchestrator, acts as the agent harness, allowing researchers to monitor packet loss, read live spectrograms, and rewrite signal-processing blocks through natural language prompts. The system can handle 2.4 terabits per second on NVIDIA RTX PRO 6000 Blackwell GPUs, with inference running on premises next to the instrument.
Limitations and availability
The system is currently in use at the Allen Telescope Array in Northern California and has already achieved the first real-time AI detection of a pulsar in the Crab Nebula. The team is evaluating open models including NVIDIA Nemotron for cases where data locality and turnaround matter most. The system is not yet available for general use outside of this specific scientific application.
