Sophelio launched the Fusion Equilibrium Challenge, the first fusion energy competition accepted to the NeurIPS Competition Track, on Aug. 11, 2026, in Austin, Texas. The challenge releases 133 GB of experimental tokamak data and asks the global AI community to reconstruct plasma magnetic geometry using only non-magnetic diagnostics. Organized with the DIII-D National Fusion Facility, UKAEA’s FAIR-MAST program, and the University of Texas at Austin’s Institute for Fusion Studies, with data hosted on Hugging Face, the competition targets future reactor-class devices like ITER, SPARC, ARC, and CFETR, where neutron fluxes degrade magnetic sensors. The dataset, under the CC BY 4.0 license, contains 9,121 curated plasma discharges: 7,915 from DIII-D in San Diego and 1,206 from MAST in Culham. Participants must infer the two-dimensional poloidal flux map, ψ(R,Z), and key equilibrium parameters using non-magnetic features such as poloidal-field coil currents and Thomson-scattering electron profiles. The challenge also examines whether machine-learning models capture machine-specific correlations or discover cross-device representations. Sophelio CEO Craig Michoski stated that open datasets and benchmarks have driven machine-learning advances and that fusion is ready for the same transition.