Nvidia’s AI agents now steer the Allen Telescope Array in the search for alien signals

The search for extraterrestrial intelligence has a new set of hands on the controls. AI agents now help run the Allen Telescope Array in Northern California as it scans the sky for possible alien signals.

NVIDIA announced the integration on October 30, 2026. Researchers can now tell the array what they need in ordinary sentences, and software agents handle the rest.

The work comes from Breakthrough Listen and the SETI Institute. It builds on NVIDIA’s Holoscan platform, the company’s toolkit for processing sensor data as it arrives.

What the agents actually do

The Allen Telescope Array uses 42 antennas. Running them has traditionally meant a lot of careful manual work: lining up instruments, cleaning up bad data and deciding what analysis to run.

Under the new setup, AI agents take on those daily chores. They can calibrate instruments, correct errors in the data and apply AI models to live radio data. According to the research findings, they can also steer the antennas themselves.

The key detail is how researchers give orders. Everything is triggered through natural-language prompts.

The hardware under the hood

Radio astronomy produces an enormous amount of data. This system can handle up to 2.4 terabits per second of sensor data.

The muscle comes from NVIDIA’s RTX PRO 6000 Blackwell GPUs. These are graphics processors, the same category of chip that powers modern AI training, repurposed here to chew through signals from the sky.

Two software pieces make the pipeline work. The first is the DAQIRI library, which moves high-bandwidth data streams directly into GPU memory.

The second is the SETI Institute’s CyberEther framework. It relies on NVIDIA’s CUDA-X libraries to process that data efficiently.

Importantly, the inference happens on-premises, right next to the telescopes. The data does not need to be shipped off somewhere else before anyone can look at it.

From offline analysis to live listening

The research findings describe the integration as a shift from offline analysis to in-flight processing. In an offline workflow, observations get recorded and examined later. In-flight processing flips that order. The system analyzes signals while they stream in, which the findings say enhances anomaly detection.

The Crab Pulsar warm-up act

In February 2025, the team achieved real-time detection of giant pulses from the Crab Pulsar.

That work was described in a peer-reviewed pipeline published in 2025. It delivered a speed-up of approximately 600 times over prior benchmarks.

The new agentic layer sits on top of that foundation. The 2025 work showed the pipeline could be fast. The 2026 integration aims to make it easier to operate.

Not a detection, a translation

It is not being presented as a detection of extraterrestrial intelligence.

Andrew Siemion, a leading figure at the SETI Institute, emphasized the technology’s potential to expand our understanding of the universe. He described it this way:

“a breakthrough in translation.”

The breakthrough is in translating a researcher’s intent into telescope operations, without that researcher needing to hand-tune every step.