Matt Landreman’s analytic-equilibria project offers explicit mathematical formulas and supplemental artifacts for inspection. It is distinct from the separate Lean-formalized Grad-conjecture paper. For readers following AI-assisted science, the useful question is how the published formulas, checks and discovery record support the final claim, without treating them as a fusion-device demonstration.
Two analytic solution families
The paper, submitted September 22 and revised September 28, describes non-axisymmetric toroidal MHD equilibria with exact nested flux surfaces. One family has uniform rotational transform of two; the other has a sheared profile. It also relates the equations to steady incompressible Euler flow. These are the author’s analytic research claims. They should not be conflated with an experimental reactor result, a general numerical solution method or the separate Gómez-Serrano, Liehr and Taylor construction.
What the repository makes inspectable
The supplemental repository includes plotting scripts, notebooks checking equations, DESC comparisons and the original GPT-6 Astra Pro prompt and chat. It lists MIT licensing. These artifacts expose more of the path behind the result, but their presence is not an independent audit. We have not run the notebooks, reproduced the numerical comparisons or checked the derivation. A discovery conversation and a mathematical verification record answer different questions; neither should silently stand in for the other.
A focused verification worksheet
Our suggested review picks one family, lists its parameter conditions and traces each formula to the claimed equation. Preserve the exact expression used in a symbolic check and the configuration used in a numerical check. Record where a plotted example sits within the permitted domain. If the check relies on a simplification, make that assumption visible. A visually convincing surface is not a substitute for checking the stated identities. This worksheet is a proposed reader aid, not a report that we completed those steps.
Compare two plasma stories without merging them
The Lean Grad-conjecture explainer concerns a different paper and proof record. When comparing the projects, examine their statements and hypotheses before comparing significance. Ask which symmetry definition, domain and pressure conditions each uses and whether a quoted conclusion applies to both. Keep authorship and artifacts attached to the appropriate claim. Combining two related repository links into one sweeping announcement would hide the differences readers need to understand, even if both concern the same historical conjecture.
Engineering conclusions need extra evidence
For a practical follow-up, identify what an engineer would still need beyond an equilibrium formula: relevant physical conditions, an implementable setup and evidence under intended operating conditions. We do not predict stability, net energy or commercial deployment from these supplemental files. A worthwhile next article would report a bounded independent reproduction, including failures or corrections. Today’s coverage explains the source record and our proposed review method; it makes no claim that an AI-assisted discovery removes the need for mathematical or experimental scrutiny.
Frequently asked questions
Is this the same paper as the Lean Grad-Conjecture repository?
No. Landreman’s analytic-equilibria paper and supplemental artifacts are a separate work.
Did FyreLinkz run the notebooks or DESC comparison?
No. We reviewed their documented purpose without independently reproducing the checks.
Does an analytic equilibrium demonstrate a working reactor?
No. A mathematical equilibrium is not an operating-device or net-energy result.
Sources & further reading
Primary sources checked Oct 11, 2026. Vendor statements are attributed; editorial advice is our own.
- 1Analytic toroidal 3D MHD equilibria and steady Euler flows ↗Matt Landreman / arXiv
- 2Analytic 3D equilibria supplemental artifacts ↗Matt Landreman
Help us keep this useful. Check contact availability →
