NASA, IBM launch new AI model for studying the moon

Live updates

1. 7d ago

   NASA and IBM's AI model for lunar study expands with hardware deployments

   NASA and IBM's open-source AI foundation model, launched earlier, uses 17 years of observation data to automate lunar feature identification. The model has reduced polar-ice prediction errors by up to 22%. Recent developments include Astrolab's FLIP rover, which will test a supercomputer and AI on the Moon's south pole, supporting NASA's Artemis program. HPE's Spaceborne Computer-4 will fly aboard FLIP to test AI and advanced computing on the Moon.

   Why it matters

   These hardware deployments aim to support future lunar exploration and the lunar economy. The AI model and onboard computers will enable more autonomy in future space missions and improve the understanding of the lunar surface.

   What is confirmed

   • NASA and IBM launched an open-source AI foundation model using 17 years of observation data to automate lunar feature identification.
   • The AI model reduced polar-ice prediction errors by up to 22%.
   • Astrolab's FLIP rover will test a supercomputer and AI on the Moon's south pole.
   • HPE's Spaceborne Computer-4 will fly aboard Astrolab's FLIP lunar rover to test AI and advanced computing on the Moon.

   What to watch next

   • Launch of Astrolab's FLIP rover
   • Results of HPE's Spaceborne Computer-4 testing on the Moon
   • Future updates on NASA's Artemis program
   Sources used for this update (5)

   1. www.nasa.gov — Lunar Surface Technology
   2. www.technologyreview.com — Astrolab’s upcoming rover will run the first LLMs on the moon
   3. www.sciencetimes.com — Astrolab's FLIP Moon Rover to Test High-Performance ...
   4. aidailypost.com — Astrolab's FLIP Rover Runs First LLMs on Moon
   5. aerospaceglobalnews.com — Astrolab FLIP rover to carry AI computer to the Moon
   confidence 100%
2. 10d ago

   HPE and FLIP Rover Bring High Performance Computing to Moon

   HPE Spaceborne Computer will run AI and high performance computing benchmarks on the lunar surface to support future exploration and the lunar economy. Simultaneously, a rover named FLIP is preparing to launch the first supercomputer to the moon. These hardware deployments complement the open-source AI foundation model previously launched by NASA and IBM, which uses 17 years of observation data to automate the identification of craters, polar ice, and geological features while reducing polar-ice prediction errors by up to 22%.

   Why it matters

   Moving computation from Earth-based servers to the lunar surface reduces latency for autonomous systems. This shift enables real-time data processing for scientific research and infrastructure planning.

   What is confirmed

   • NASA and IBM released an open-source AI foundation model for lunar exploration.
   • The NASA-IBM model was trained on 17 years of lunar observations to identify craters, polar ice, and geological features.
   • The AI model reduces polar-ice prediction errors by up to 22%.
   • HPE Spaceborne Computer plans to execute AI and HPC benchmarks on the Moon.

   Still unconfirmed

   • A rover named FLIP will carry a supercomputer to the lunar surface.

   What to watch next

   • Launch date for the FLIP rover
   • Results of the HPE Spaceborne Computer AI benchmarks
   Sources used for this update (5)

   1. www.space.com — A supercomputer is about to launch to the moon for the 1st time, aboard a rover named FLIP | Space
   2. newsroom.ibm.com — IBM Newsroom - Latest News
   3. finance.yahoo.com — HPE Spaceborne Computer to Power High Performance Computing ...
   4. iri-group.jp — IBMとNASA、月面探査向け基盤AIモデルをオープン公開
   5. economictimes.indiatimes.com — nasa artemis program: Latest News & Videos, Photos about nasa ...
   confidence 90%
3. 13d ago

   NASA and IBM release open-source AI for lunar research

   NASA and IBM launched an open-source AI foundation model to automate lunar terrain modeling and scientific research. Trained on 17 years of lunar observations, the system identifies geological features, craters, and polar ice. The model improves the accuracy of polar-ice predictions by reducing errors by up to 22%. This technology replaces the previous requirement for scientists and engineers to analyze satellite imagery manually.

   Why it matters

   Automating the mapping of the lunar surface allows for faster identification of resources and hazards. Precise polar-ice data is essential for planning future crewed missions to the moon.

   What is confirmed

   • NASA and IBM released an open-source AI foundation model for lunar research.
   • The AI model was trained using 17 years of lunar observations.
   • The system reduces polar-ice prediction errors by up to 22%.
   • The AI maps craters, geological features, and polar ice.

   What to watch next

   • Integration of the AI model into active lunar mission planning
   • Expansion of the model to include Martian terrain data
   Sources used for this update (4)

   1. science.nasa.gov — Mars - NASA Science
   2. science.nasa.gov — Hubble Space Telescope - NASA Science
   3. www.youtube.com — NASA - YouTube
   4. www.nasa.gov — Learning Resources - NASA
   confidence 100%
4. 16d ago

   NASA and IBM launch AI model for lunar research

   NASA and IBM have released an open-source AI foundation model trained on 17 years of lunar observations to assist scientists and engineers with lunar research and terrain modeling. The model maps craters, geological features, and polar ice, reducing polar-ice prediction errors by up to 22%. This automated approach replaces manual satellite imagery analysis.

   Why it matters

   The collaboration between NASA and IBM aims to advance lunar research and exploration by leveraging AI technology. The AI model processes data from NASA's Lunar Reconnaissance Orbiter and other missions, providing a more efficient and accurate way to analyze lunar terrain.

   What is confirmed

   • NASA and IBM have released an open-source artificial intelligence foundation model trained on 17 years of lunar observations.
   • The AI model reduces polar-ice prediction errors by up to 22%.
   • The model processes data from NASA's Lunar Reconnaissance Orbiter and other missions.

   What to watch next

   • Further updates on the AI model's performance and applications in lunar research
   • Potential collaborations with other space agencies or private companies to expand AI technology in space exploration
   Sources used for this update (11)

   1. science.nasa.gov — NASA Science
   2. www.nasa.gov — NASA
   3. www.nasa.gov — NASA Live - NASA
   4. en.wikipedia.org — NASA - Wikipedia
   5. plus.nasa.gov — Home | NASA+
   6. images.nasa.gov — NASA Image and Video Library
   7. www.nasa.gov — NASA
   8. science.nasa.gov — Solar System Exploration - Science@NASA
   9. www.space.com — The moon could turn astronauts into 'walking, high voltage capacitors' that accidentally zap spacecraft
   10. timesofindia.indiatimes.com — Where India takes to the streets: Historic protest sites from Mumbai to Kolkata and Delhi | Delhi News
   11. www.nasa.gov — About NASA - NASA
   confidence 100%
5. 19d ago

   NASA and IBM Release Open-Source Lunar AI Model

   NASA and IBM have released an open-source artificial intelligence foundation model trained on 17 years of lunar observations. The tool processes decades of data from NASA's Lunar Reconnaissance Orbiter and other missions to map craters, geological features, and polar ice. According to developers, the system cuts polar-ice prediction error by up to 22% while requiring only half the labeled data needed by traditional methods. This automated approach replaces manual satellite imagery analysis to assist scientists and engineers with lunar research and terrain modeling.

   Why it matters

   The joint release builds on prior geospatial initiatives aimed at preparing for increased lunar traffic and human returns to the moon. As researchers target resources like water ice deposits, automated tools reduce the time required to review vast archives of orbital imagery. Concurrently, NASA is pursuing separate programs to investigate underground features like lava tubes in the Marius Hills pit using seismic and radar tools.

   What is confirmed

   • NASA and IBM have released an open-source AI model designed to advance lunar research using 17 years of data.
   • The lunar foundation model cuts polar-ice prediction error by up to 22% and requires only half the labeled data.
   • The model incorporates observations from NASA's Lunar Reconnaissance Orbiter alongside data from other lunar missions.

   Still unconfirmed

   • NASA has a new program aiming to find secret underground caves on the moon.

   What to watch next

   • Adoption rates of the open-source lunar foundation model among independent researchers and developers
   • Further developments regarding NASA seismic and radar surveys targeting potential lava tubes in the Marius Hills pit
   Sources used for this update (11)

   1. completeaitraining.com — NASA and IBM release open-source lunar foundation model ...
   2. guavy.com — NASA and IBM Release Open-Source AI Model for Lunar Research
   3. www.space.com — NASA's hunting for a secret underground cave on the moon
   4. writersofthefuture.com — Podcast | Writers & Illustrators of the Future
   5. www.space.com — Can we protect historic sites on the moon from the coming lunar rush?
   6. www.nasa.gov — Super Intelligence
   7. www.natureworldnews.com — An Apollo-Style Seismic Survey Will Test Whether a Lava Tube Hides Beneath the Moon's Marius Hills Pit - ....
   8. businessupturn.com — NASA and IBM open-source lunar AI model trained on 17 years of Moon data · Business Upturn
   9. guavy.com — NASA and IBM Release Open-Source Lunar Foundation Model for ...
   10. www.scrabbl.com — NASA-IBM Lunar Foundation Model Advances Moon Science
   11. economictimes.indiatimes.com — geospatial technology: Latest News & Videos, Photos about ...
   confidence 95%
6. 25d ago

   NASA and IBM Open-Source Lunar AI Model Trained on 17 Years of Data

   NASA and IBM Research launched an open-source artificial intelligence foundation model on September 10, 2026, to analyze 17 years of lunar observations. The tool enables researchers and developers to adapt the system for crater detection, terrain modeling, and predicting water ice deposits. By converting nearly two decades of spacecraft data into a reusable format, the model replaces manual satellite imagery reviews with automated geospatial analysis. This capability supports engineering teams as NASA plans for human returns to the lunar surface.

   Why it matters

   Planetary scientists previously analyzed only a small fraction of available lunar photos due to the labor of manual review. This AI model expands the scale of analysis for volcanic features and polar ice deposits. These insights are critical for selecting landing sites for future crewed missions.

   What is confirmed

   • NASA and IBM Research launched an open-source lunar AI foundation model on September 10, 2026.
   • The model is trained on 17 years of lunar observations.
   • The tool supports tasks including crater detection, terrain modeling, and the prediction of ice deposits.
   • The system allows researchers and developers to adapt lunar data for machine learning.

   What to watch next

   • Publication of peer-reviewed studies using the open-source model
   • NASA announcements regarding specific landing site selections based on AI ice predictions
   Sources used for this update (5)

   1. ai-brainer.com — NASA and IBM train lunar model for water ice search
   2. www.archyde.com — NASA and IBM Launch Open-Source AI Model to Map ... - Archyde
   3. stratnewsglobal.com — Lunar Foundation Model Advances Moon Science - StratNews Global
   4. zglg.work — NASA and IBM Open-Source a Lunar AI Model Trained on 17 Years ...
   5. guavy.com — NASA and IBM's Lunar Foundation Model Advances Moon Science ...
   confidence 100%
7. 27d ago

   NASA and IBM Launch Lunar AI Model

   NASA and IBM have officially launched a joint artificial intelligence foundation model designed to extract new insights from nearly two decades of lunar data. The open-source system processes data from orbiting spacecraft to investigate craters, volcanic features, and potential ice deposits in dark polar regions. Benchmark results show accuracy gains of up to 22 percent in mapping potential ice deposits. Planetary scientists previously relied on manual methods to review satellite imagery, restricting analysis to a small fraction of available photos, but this tool improves geospatial analysis capabilities for engineering teams as NASA prepares for human returns to the lunar surface.

   Why it matters

   Planetary scientists previously faced major bottlenecks because manual reviews of satellite imagery limited analysis to a tiny fraction of available photos. The new open-source system makes decades of accumulated lunar data vastly more accessible for research and engineering teams. This development arrives as space agencies ramp up technical preparations for future crewed missions to the moon.

   What is confirmed

   • NASA and IBM launched an artificial intelligence model to extract new insights from nearly two decades of lunar data.
   • Benchmark results demonstrate mapping gains of up to 22 percent for potential ice deposits.

   What to watch next

   • Adoption rates of the open-source model among planetary science research teams
   • Further benchmark results regarding crater mapping and volcanic feature analysis
   Sources used for this update (4)

   1. colombiaone.com — NASA Uses AI to Read the Moon Before Humans Return to Its Surface
   2. www.siliconrepublic.com — Juice: ESA’s Jupiter mission completes third gravity assist
   3. www.exterrajsc.com — Open-Source AI Model Trained on Decades of Lunar Data ...
   4. www.glbgpt.com — Das Lunar Foundation Model der NASA und IBM ...
   confidence 100%
8. 29d ago

   NASA and IBM Expand Reach of Open-Source Lunar AI Tools

   NASA and IBM released a joint artificial intelligence foundation model trained on nearly two million spacecraft image tiles to map the lunar surface and improve ice detection accuracy by up to 22 percent. Planetary scientists previously relied on manual methods to review satellite imagery, restricting analysis to a small fraction of available photos. The open-source system processes data from orbiting spacecraft to investigate craters, volcanic features, and potential ice deposits in dark polar regions, advancing geospatial analysis capabilities for engineering teams.

   Why it matters

   Traditional planetary science relied on slow manual review of satellite photos, limiting the analysis of lunar data. The release of the NASA-IBM model modernizes this work by automating the processing of massive volumes of orbital imagery. Meanwhile, broader commercial developments in orbital computing continue to attract venture capital funding.

   What is confirmed

   • NASA and IBM released a joint artificial intelligence foundation model trained on nearly two million spacecraft image tiles.
   • The open-source system improves lunar ice detection accuracy by up to 22 percent.
   • The tool processes data from orbiting spacecraft to investigate craters, volcanic features, and potential ice deposits in dark polar regions.

   Still unconfirmed

   • Satlyt raised $8M to run AI on satellites and function as the Android of orbital computing.

   What to watch next

   • Further releases or engineering adoption metrics for the NASA-IBM Lunar Foundation model
   • Deployment updates regarding orbital computing software platforms
   Sources used for this update (3)

   1. finance.yahoo.com — Satlyt, founded by a former Google and SpaceX product manager, raises $8M to run AI on satellites
   2. techbytes.app — NASA-IBM Lunar Foundation open-Source Geospatial AI Model
   3. newsroom.ibm.com — IBM Newsroom
   confidence 100%
9. 31d ago

   NASA and IBM AI Model Boosts Lunar Ice Detection Accuracy

   NASA and IBM have released a joint artificial intelligence foundation model trained on nearly two million spacecraft image tiles to map the lunar surface. The open-source system improves lunar ice detection accuracy by up to 22 percent. Planetary scientists previously relied on manual methods to review satellite imagery, restricting analysis to a small fraction of available photos. The new tool processes data from orbiting spacecraft to investigate craters, volcanic features, and potential ice deposits in dark polar regions.

   Why it matters

   Analyzing decades of lunar observations manually has long slowed planetary research. This collaboration builds upon ongoing artificial intelligence development that began taking shape in 2023 through NASA teams. The technology aims to unlock hidden water hotspots and optimize data use from space missions.

   What is confirmed

   • NASA and IBM released a joint foundation model trained on nearly two million spacecraft image tiles.
   • The AI model improves lunar ice detection accuracy by up to 22 percent.

   What to watch next

   • Further deployment of the NASA-IBM Lunar Foundation Model by planetary researchers to investigate dark lunar regions.
   • Additional performance benchmarks and research findings regarding the open-source model.
   Sources used for this update (12)

   1. jen.jiji.com — EBRD supports Kazakhstan’s tech company Abris to accelerate digital transformation
   2. quantumzeitgeist.com — NASA and IBM’s AI spots lunar ice with 22% better accuracy
   3. www.livescience.com — 'We only expect to encounter that once every 130 years': NASA's LRO spacecraft discovers giant new lunar ....
   4. stockanalysis.com — IBM23
   5. ansys.synopsys.com — How Does Space Technology Benefit Earth?
   6. www.futura-sciences.com — AI just unlocked the Moon’s hidden water hotspots ... - Futura
   7. www.nasa.gov — AI/ML News - NASA
   8. science.nasa.gov — Artificial Intelligence for Science - NASA Science
   9. aviationweek.com — WEBINAR: Time to First Article - NASA, Airbus & Agentic AI
   10. stockanalysis.com — IBM06
   11. www.glbgpt.com — NASA-IBM Lunar Foundation Model: Uses, Results & Access
   12. newsroom.ibm.com — IBM Newsroom - Latest News
   confidence 100%
10. 42d ago

    NASA and IBM Deploy AI Model to Analyze Lunar Data

    NASA and IBM released the NASA-IBM Lunar Foundation Model to help researchers analyze diverse datasets from lunar missions. This open-source AI tool processes millions of images from five orbiting spacecraft to map craters, volcanic features, and potential ice deposits. The Universities Space Research Association provided scientific evaluation, dataset development, and planetary science expertise to build the system. NASA scientists intend to use the technology to investigate the darkest regions of the moon's surface.

    Why it matters

    Lunar mapping was previously limited to specialized research teams. This tool democratizes access to advanced analysis for the broader scientific community.

    What is confirmed

    • The NASA-IBM Lunar Foundation Model is an open-source AI tool for analyzing lunar mission datasets.
    • The model was trained using over two million images from five spacecraft orbiting the moon.

    Still unconfirmed

    • NASA scientists plan to use AI to explore the darkest parts of the moon's surface.

    What to watch next

    • Publication of new lunar maps generated by the foundation model
    • Confirmation of specific ice deposits identified via AI analysis
    Sources used for this update (2)

    1. tech.yahoo.com — NASA turns to AI to uncover new clues hidden across the moon’s surface
    2. finance.yahoo.com — USRA Contributes Planetary Science Expertise to NASA-IBM Lunar Foundation Model
    confidence 90%
11. 45d ago

    NASA and IBM release open-source Lunar Foundation Model

    NASA and IBM have launched the NASA-IBM Lunar Foundation Model, a free, open-source AI tool designed to analyze decades of lunar observation data. The system helps researchers map craters, locate potential ice deposits, and study volcanic features on the Moon. The model was trained using data from multiple NASA missions and instruments, including more than two million images from five spacecraft orbiting the Moon. This tool provides public access to advanced lunar mapping capabilities that were previously restricted to specialized research teams.

    Why it matters

    The project aims to improve the efficiency of space research by automating the analysis of massive datasets. Accurate mapping of ice and volcanic sites is essential for planning future lunar landings and sustaining human presence on the Moon.

    What is confirmed

    • The NASA-IBM Lunar Foundation Model is an open-source AI tool.
    • The model identifies craters, ice deposits, and volcanic features on the Moon.
    • The AI was trained using data from multiple NASA missions and instruments.

    Still unconfirmed

    • The AI spots Moon water ice 22% more accurately than previous methods.
    • The model was trained on more than two million images from five Moon-orbiting spacecraft.

    What to watch next

    • Peer-reviewed validation of the model's mapping accuracy
    • Integration of the tool into active lunar mission planning
    Sources used for this update (4)

    1. finance.yahoo.com — IBM Takes its AI Strategy Beyond Earth with NASA Lunar Model
    2. tech.yahoo.com — NASA-IBM Lunar Foundation Model puts free AI moon mapping in anyone’s hands
    3. www.martincid.com — NASA and IBM released a free AI that spots Moon water ice 22% more accurately
    4. events.fastcompany.com — Schedule | Fast Company Innovation Festival
    confidence 90%
12. 47d ago

    NASA and IBM Launch Lunar AI Model

    NASA and IBM have launched a new artificial intelligence foundation model designed specifically for lunar science. The system aims to assist researchers in mapping craters and identifying ice on the moon. Financial markets responded to the announcement as IBM shares gained two percent following reports that the lunar artificial intelligence outperforms standard mapping methods by twenty-three percent. The collaborative project leverages advanced technology to improve space research capabilities and enhance the analysis of lunar data.

    Why it matters

    The partnership between NASA and IBM expands on previous technological cooperation in space exploration and earth science applications. Artificial intelligence models of this scale help process massive volumes of remote sensing data collected by lunar orbiters and missions. This initiative arrives alongside other major space agency undertakings, such as the recent launch of the Nancy Grace Roman telescope designed to explore the edge of the observable universe.

    What is confirmed

    • NASA and IBM have launched a new AI foundation model for lunar science to help map ice and craters on the moon.
    • IBM shares gained two percent following the announcement that the lunar AI beats mapping methods by twenty-three percent.

    What to watch next

    • Further deployment details and performance metrics from NASA and IBM regarding the lunar AI model.
    • Additional integration of the foundation model into upcoming lunar research missions and mapping projects.
    Sources used for this update (6)

    1. Reuters — IBM, NASA launch AI model to help map ice, craters on Moon
    2. NASA Science (.gov) — NASA, IBM Launch AI Foundation Model for Lunar Science
    3. IBM Research — Introducing IBM and NASA’s new foundation model for the Moon
    4. mynews13.com — NASA and SpaceX successfully launch Nancy Grace Roman telescope
    5. finance.yahoo.com — IBM Gains 2% as Lunar AI Beats Mapping Methods by 23%
    6. Space — NASA, IBM launch new AI model for studying the moon
    confidence 95%