Room-Temperature Superconductor
A material that conducts current without loss at everyday temperature and pressure. Reported records exist only at about -120 C or at over a million bar.
Open in the interactive tree →Superconductors carry current with zero resistance but need cold. Hydrides such as LaH10 reach 250 K (-23 C), but only at around 150 to 170 GPa (1.5 to 1.7 million bar) in diamond anvil cells. Several room-temperature claims failed: Dias's hydrides (2020 and 2023, retracted) and LK-99 (July 2023, debunked within weeks as a copper-sulfide impurity effect).
As of October 2026
October 2026: no confirmed room-temperature superconductor at normal pressure. The reported record at normal pressure is 151 K (Hg-1223 after pressure quenching, University of Houston, PNAS, March 2026; retained Tc 139-151 K across samples, not yet independently replicated; the previous record was 133 K since 1993) and about 250 K under high pressure (LaH10, 2019). Research concentrates on ternary hydrides that stay stable at lower pressure and on nickelates (92 to 96 K under pressure, about 40 K as films at normal pressure). AI predictions help with candidate searches but do not replace an explanation of the mechanism.
What is missing
- A theory that explains cuprate and nickelate superconductivity and so allows predictions
- A material that keeps a high transition temperature stable at normal pressure (hydrides need over a million bar)
- Reproducible, openly checked measurements (zero resistance, Meissner effect) after LK-99 and the Dias scandal
- Ways to make wires and films from brittle ceramics
Becomes possible once solved
- Lossless power grids and storage
- Light, cheap magnets for MRI, maglev trains and fusion
- Quantum computers without heavy cooling
Open steps
- Hydrides stable at lower pressure High AI leverageFind hydrides that keep a high transition temperature well below the roughly 1.5 million bar needed today, ideally near normal pressure.
- Theory of cuprate and nickelate pairing Low AI leverageExplain why cuprates and nickelates superconduct at high temperatures so that new materials can be predicted from first principles.
- Higher-Tc nickelates Medium AI leverageRaise nickelate transition temperatures beyond 96 K under pressure and beyond 40 K as films at normal pressure, by tuning chemistry and strain.
- Open, reproducible verification Medium AI leverageEstablish shared protocols and open raw data for zero resistance and the Meissner effect, so claims like LK-99 are tested in days.
- From prediction to measured sample Medium AI leverageLink candidate prediction to automated synthesis and measurement so that a predicted superconductor is made and tested within days.
Where AI could help
Medium AI leverage. ML speeds up candidate screening, but the mechanism is not understood and stable, reproducible samples must still be made and measured in labs.
- Screen billions of compositions and rank candidates by predicted transition temperature and stability
- Use machine-learned potentials to find hydrides that stay stable at lower pressure
- Mine papers and raw data to flag irreproducible claims early
- Steer autonomous labs to synthesize and test the best candidates quickly
Shown so far
- In July 2026 an Aalto-led team published ML-guided screening, first-principles checks and lab synthesis that confirmed two new kagome superconductors (YRu3B2, LuRu3B2), with transition temperatures below 1 kelvin. source
- In December 2024 a Princeton-led team used a neural network on chemical composition alone to predict Mo20Re6Si4, then synthesized it and confirmed superconductivity below 5.4 kelvin. source
Prerequisites
- Superconductivity1911
- High-Temperature Superconductors1986
- Graphene & 2D Materials2004Twisted bilayer graphene became superconducting in 2018, a new route to study the effect
- Nickelate Superconductors2023
Unlocks
- Lossless Power Grids2050s?