Empowering our People
AstroSage Llama 3.1 8B is a breakthrough domain-specific AI model tailored for astronomy, featuring 8 billion parameters and trained on an extensive corpus of astronomy literature. It enhances research accuracy, efficiency, and contextual understanding. Openly accessible, it democratizes advanced AI, promoting inclusivity in scientific discovery while addressing key challenges for future development.
In June 2025, research highlighted AI’s transformative impact on gravitational wave astronomy, improving LIGO’s signal detection through deep learning and other techniques. AI enhances true positive detection rates, reduces false positives, and enables real-time analysis. As gravitational wave science evolves, AI will be essential for interpreting and understanding cosmic events.
In April 2025, “AI Cosmologist I” introduced autonomous scientific agents capable of conducting entire research workflows. This AI integrates diverse functions, significantly impacting cosmology by automating hypothesis generation, simulation, and analysis. While enhancing scientific efficiency, it raises ethical questions regarding authorship and accountability, paving the way for a new era in scientific inquiry.
In June 2025, Europe launched Jupiter, its first exascale supercomputer, achieving 793 petaFLOPS performance. Powered by NVIDIA Grace Hopper superchips, it enhances research in AI, physics, and climate science while prioritizing energy efficiency. Jupiter promotes technological sovereignty and global collaboration, marking a significant milestone in Europe’s computing landscape and scientific innovation.
In July 2023, astronomers observed supernova SN 2023zkd, marked by an AI-assisted detection. This unique event featured an extended brightening over four years, gradual dimming, and a secondary flash, suggesting a massive star interacting with a black hole. The discovery enhances understanding of binary star systems and the evolving role of AI in astrophysics.
In 2025, researchers discovered that Sagittarius A*, the Milky Way’s supermassive black hole, spins at nearly its maximum rate. This finding, derived from 12 million AI-enhanced simulations, reveals that its emissions stem from hot electrons in the accretion disk rather than powerful jets, reshaping black hole dynamics and cosmic evolution understanding.
A research team at Emory University used AI to uncover non-reciprocal forces in dusty plasmas, challenging existing theories in plasma physics. This marks a shift from AI as a mere tool to a collaborator in scientific discovery, suggesting significant implications for fields like fusion energy and materials science, and redefining core scientific assumptions.
In 2025, astronomers discovered a new moon orbiting Uranus, named S/2025 U1, using the James Webb Space Telescope. This 10-kilometer satellite, the first discovered since 2003, enhances our understanding of planetary systems and interactions within Uranus’s rings. It highlights JWST’s capabilities and suggests more hidden moons may exist.
IBM and NASA’s AI model “Surya” significantly enhances solar flare prediction, improving accuracy by 16% over past models. Utilizing extensive solar data, Surya integrates advanced deep learning techniques, contributing to critical space weather monitoring. Its open-source nature promotes global research, thereby safeguarding modern technological infrastructure against solar-induced disruptions.
