The Moon's south pole was supposed to be the easy part. Decades of orbital data, an intentional crash into Cabeus crater in 2009, and signals from India's Chandrayaan-1 all pointed toward permanently shadowed regions packed with water ice -- enough, in the optimistic reads, to fuel a whole cislunar economy. A new peer-reviewed study published in Science Advances on March 18, 2026 says the picture is considerably more complicated. Using NASA's ShadowCam instrument aboard South Korea's Korea Pathfinder Lunar Orbiter (KPLO, also called Danuri), researchers at the University of Hawai'i at Manoa combed through the highest-resolution images ever captured of the Moon's permanently shadowed regions. Lead author Shuai Li put it plainly: "We expected to see more." AI-generated image ShadowCam imagery concept of the Moon's south pole permanently shadowed regions. Credit: AI illustration / NASA/KARI/Arizona State University What ShadowCam Actually Found ShadowCam was purpose-built for darkness. Flying on KPLO since late 2022, the instrument captures reflected light inside craters that never see direct sunlight, using photons scattered from nearby sunlit terrain. Its resolution runs to about 2 meters per pixel -- fine enough to distinguish fresh crater ejecta from smooth compacted regolith. Li's team analyzed two specific optical signatures that betray the presence of water ice: high visible reflectance (ice reflects roughly twice as much light as dry lunar soil) and forward scattering, a property in which light bounces forward off ice crystals rather than backward off loose dust. Both should be detectable at concentrations above roughly 20 to 30 percent by weight for reflectance, or about 5 to 10 percent for forward scattering. The result: no widespread high-abundance ice deposits anywhere in the permanently shadowed regions surveyed. Most brightness variations inside the PSRs turned out to trace fresh regolith exposure from recent impacts, boulder fields, and mass wasting events -- not ice. The south pole showed fewer potential ice signatures than the north pole, which was itself modest. What the Study Did (and Did Not) Rule Out The ShadowCam analysis found no evidence of ice above roughly 20-30 wt% near the surface . A small number of spots, 10 to 50 meters across, inside craters like de Gerlache, Cabeus, Faustini, and Hermite A showed forward-scattering anomalies consistent with more than 10 wt% ice, possibly kicked up by recent impacts. The study does not rule out widespread low-concentration ice, or ice buried deeper than the top meter -- it simply says the easy, accessible, high-concentration surface deposits many had assumed were there are not obviously visible from orbit. 20-30% Detection threshold (reflectance), not found 5-10% Detection threshold (scattering), not found ~2 m ShadowCam image resolution per pixel <10 Anomalous ice-consistent spots found across all surveyed PSRs 1 m Subsurface depth sensitivity of neutron spectrometers Mar 18 Publication date, Science Advances, 2026 Why the Ice Question Has Always Mattered The argument for building any permanent human presence on the Moon has always rested, in part, on water ice. Without a local source of hydrogen and oxygen, every kilogram of propellant, every liter of drinking water, every breath of air has to be hauled from Earth at staggering cost. A lunar depot that could electrolyze local water into rocket fuel would change the economics of deep space travel entirely, enabling Mars missions, cislunar fuel markets, and long-duration surface stays. NASA's selection of nine candidate landing regions for Artemis IV within 6 degrees of the south pole was not arbitrary. Those sites sit close to permanently shadowed craters while still receiving enough sunlight for solar power generation. The plan was always to land near the ice, not in the darkness, and send astronauts (and eventually robots) into the PSRs on short excursions to sample and eventually mine frozen volatiles. AI-generated image Any Artemis landing at the south pole will require astronauts to drill or sample the subsurface to confirm usable ice -- surface concentrations appear insufficient for direct extraction. Credit: AI illustration Those plans assumed that surface ice would at least be present in accessible quantities. Prior missions had offered encouragement. The LCROSS impactor in 2009 confirmed water vapor in the ejecta plume from Cabeus. India's Chandrayaan-1 M3 spectrometer detected hydroxyl and water signatures. NASA's LAMP ultraviolet spectrograph on LRO found surface frost in some cold traps. The emerging consensus was that PSRs held significant ice deposits, possibly layered from billions of years of comet and asteroid water delivery. ShadowCam does not overturn that consensus entirely. It refines it, in an uncomfortable direction. Li's team notes that prior detections are consistent with low-concentration or buried ice, which ShadowCam simply cannot see at orbital resolution. The thick near-surface deposits that would make ISRU straightforward are not there, or at least not visible at any scale ShadowCam can resolve. The Industry Response: Dig Deeper, Map Better The findings have not dampened commercial interest in south pole prospecting. If anything, they have clarified what the market actually needs: high-resolution subsurface data that orbital cameras cannot provide. Blue Origin announced Project Oasis in September 2025, and presented technical details on Oasis-1 at the Lunar and Planetary Science Conference in March 2026. The mission involves two identical SmallSats deployed from a Blue Moon Mk1 uncrewed lander into a 10-by-50 kilometer polar orbit -- far lower than any previous lunar orbiter. The periapsis, their closest approach, sits directly over the south pole's permanently shadowed regions. AI-generated image Blue Origin's Oasis-1 concept: two low-flying SmallSats mapping water ice at resolutions current missions cannot achieve. Credit: AI illustration Each Oasis-1 satellite carries three instruments: a Hybrid Gamma-Ray and Neutron Spectrometer (GRNS) for subsurface hydrogen detection down to about one meter deep, a deployable-boom magnetometer for crustal magnetic anomalies, and a multispectral pushbroom imager for surface composition at better than 5-meter resolution. The mission runs in two phases. The 90-day primary phase maps the entire south pole at resolutions roughly nine times finer than current global datasets. The final 10 days involve a controlled deorbit: the satellites descend toward the surface for ultra-high-resolution water mapping of priority PSR sites before impact. Oasis-1 Instrument Suite • GRNS (Gamma-Ray and Neutron Spectrometer): Detects subsurface hydrogen (water ice proxy) to ~1 m depth. Resolution ~15 km/pixel globally, improving to hundreds of meters during deorbit phase. • Magnetometer: Deployable boom, maps crustal magnetic anomalies. May indicate buried metal or platinum group element deposits. Resolution 15-30 km/pixel. • Multispectral Imager: Pushbroom design, captures surface composition including helium-3 proxies and regolith maturity. Resolution better than 5 m/pixel. NASA signed a NASA-JAXA Implementing Arrangement in March 2026 to integrate its Neutron Spectrometer System (NSS), developed at Ames Research Center, into JAXA's LUPEX rover for the Lunar Polar Exploration Mission. That rover will drill up to 1.5 meters into the regolith to directly sample what orbiting spacecraft cannot see. Launch on Japan's H3-24 rocket is planned no earlier than 2028. Artemis at the South Pole: Plans Unchanged, Assumptions Shifted NASA has not changed its south pole targeting based on the ShadowCam study. The nine Artemis IV candidate landing regions remain within 6 degrees of the pole, prioritizing terrain that sits near but not inside PSRs -- close enough to reach frozen volatiles on excursions, far enough to get sunlight for power and communications. What shifts is the operational calculus. If