Physics De-Risking Register — Kronos Fusion Energy

Abstract

This register documents an independent physics de-risking campaign that re-ran 164 design gates across Kronos Fusion Energy's two machines — the Hyperion negative-triangularity spherical-tokamak breeder and the Aegis / MetroVolt D–³He tandem-mirror burner — at their frozen design points on established first-principles codes. Nonlinear CGYRO simulations confirm the core confinement bet: negative triangularity suppresses turbulent transport by ~40% at the ion scale (converged at real electron mass) and ~80% at the electron scale, robust to collisionality and across the linear spectrum. Both machines close on the frozen physics; the distance to demonstrated closure is a set of named experiments, not open physics questions. The breeder makes no net-electricity claim: its product is tritium, helium-3, and 14-MeV neutrons.

The core confinement result

Nonlinear CGYRO confirms the central bet. Negative triangularity suppresses the turbulence that bleeds heat from the core — ~40% at the ion scale (converged and saturated at real electron mass, μ=3672) and ~80% at the electron scale — robust to collisionality and holding across the full linear spectrum. Four independent, mutually consistent gyrokinetic calculations.

Key facts

Breeder gain, Q

3.076

Breeder fusion power

85.04 MW

Plasma current

9.66 MA (no central solenoid)

Triangularity δ

−0.30 (negative)

Alpha confinement

99.98% (ASCOT5)

Tritium breeding

ratio 1.42 → ~2.0 kg-T/yr per unit

Helium-3 (fleet)

3.6–4.5 kg/yr

Burner engineering gain, Q_E

1.318

Burner fusion power

4298.5 MW

Neutron fraction

5.44%

Burner net electric

+104 / +850 / +2832 MWe @ 55 / 440 / 1400 m

Plug requirement

n_plug/n_c = 16 (WHAM-testable)

Codes used

~69 established (OpenMC, CGYRO, ASCOT5, NIMROD, Quantum ESPRESSO …)

Timeline

build Q2 2027 · first tritium 2030–31 · test burner 2032 · fleet ~2036

The mission, and why a breeder comes first

Kronos exists to build the neutron economy — a future powered by fusion and the isotopes fusion makes. Fourteen-MeV neutrons qualify the structural materials every future reactor will need; they make medical isotopes such as molybdenum-99 and targeted-alpha emitters; they make helium-3 for quantum sensing and neutron detection; and, in the aneutronic burner, they make clean firm power.

None of that arrives without four things: tritium to fuel any deuterium–tritium system, helium-3 to fuel the aneutronic burner, low-activation materials proven to survive a fusion neutron spectrum, and high-field magnets proven at reactor scale. The breeder makes all four. That is why Hyperion comes first — it is the on-ramp, not the destination.

POPCON operating diagram showing the Q=3.076 design point — The breeder operating diagram (POPCON): the design point sits on the Q = 3.076 contour at ~85 MW fusion power.

The three products

Hyperion — the breeder

A compact spherical tokamak at strong negative triangularity (δ = −0.30) that breeds its own tritium, co-produces helium-3, and floods a test volume with 14-MeV neutrons. Q = 3.076 at 85 MW; 9.66 MA held without a central solenoid; 99.98% alpha confinement. Its product is tritium, helium-3, and neutrons — not net electricity. Honest gate tritium self-sufficiency.

Aegis / MetroVolt — the burner

One D–³He tandem-mirror burner in two housings — a hardened defense build (Aegis) and a data-center build (MetroVolt). Only 5.4% of its power leaves as neutrons; the rest a direct converter catches. Q_E = 1.318, net +850 MWe at 440 m, with every one of 40,000 Monte-Carlo samples above break-even. Honest gate the WHAM-scale plug experiment.

Linear growth rate versus wavenumber, negative vs positive triangularity — Negative triangularity (gold) sits below positive (steel) across the resolved spectrum — the mechanism behind the 40–45% ion-scale turbulence suppression.

What “closing” means

The register makes one claim, repeated at every gate: the design closes. That word is used precisely. Closes on model means every governing equation, re-run from first principles at the frozen design point, gives a self-consistent, physically admissible answer. Demonstrated closure is the remaining distance to hardware — spanned by named, funded-scale experiments, not open physics questions. A stated hard requirement is a strength, not a gap: a design that hides its one hard number is unfalsifiable; one that states it — “we need a plug 16× the central-cell density, and here is the experiment that shows it” — is engineering.

Frequently asked

Does the Hyperion breeder make net electricity?

No. It is a strategic-isotope, tritium, and 14-MeV neutron-source platform; net electricity is a category error for it. Success is measured in breeding ratio and isotope yield. The Aegis / MetroVolt burner carries the power story.

What is the breeder's fusion gain?

Q = 3.076 at 85.04 MW fusion power (config-22021 design point), reproduced from first principles with 99.98% alpha confinement (ASCOT5).

How much tritium and helium-3 does it produce?

Up to ~2.0 kg of tritium per unit per year at the advanced blanket (breeding ratio 1.42); 3.6–4.5 kg/yr at fleet scale. Helium-3 follows from tritium decay plus lunar supply.

Is the burner net-energy-positive?

On the frozen model, yes: engineering gain Q_E = 1.318 with net +850 MWe at 440 m, and every one of a 40,000-sample Monte-Carlo lands above break-even. Demonstrated closure requires the WHAM-scale plug-potential experiment.

What is the one open physics gate?

The burner plug-potential requirement (n_plug/n_c = 16, 347× the demonstrated gas-dynamic-trap value) — a requirement-class result retired by a named WHAM-scale experiment, not open physics.

How was it validated?

By re-running 164 gate analyses on ~69 established, community-maintained codes (OpenMC, CGYRO, ASCOT5, NIMROD, SOLPS-ITER, Quantum ESPRESSO, WarpX …) under a deterministic seed, cross-checked against 19 frozen anchors.

When could this be built?

Construction Q2 2027; FOAK first tritium ~2030–31; a ~100 MW test burner in 2032; commercial fleets ~2036. Fleet scale is gated on lunar helium-3 (~2038–40), not a calendar.

The Kronos 2026 series — companion papers

This register is Paper 1.0, the hub of a 44-paper series. Each companion is archived on Zenodo under the community kronos_fusion_energy, with the DOI shown (resolving at doi.org on publication), and each draws on the same frozen gates re-run in this document.

Foundations — the five product volumes

1.1 Hyperion — compact ST tritium/He-3 breeder — the compact-ST strategic-materials platform — gain, breeding, He-3, 14-MeV neutrons · DOI 22132235
1.2 Aegis / MetroVolt — D–3He tandem-mirror burner — the low-neutron D–³He burner — engineering gain, the plug requirement, mirror confinement · DOI 22132474
1.3 High-field REBCO magnets & tape — the high-field REBCO superconductor basis · DOI 22132475
1.4 Direct energy conversion — quasineutral expander — recovering charged-particle power directly · DOI 22132480
1.5 AI, ML & quantum control — the computational-methods layer — ML surrogates, control, quantum · DOI 22132484

Breeder physics

2.11 Negative-triangularity confinement — negative triangularity suppresses turbulence 40–45% (TCV/DIII-D corroborated) · DOI 22132098
2.12 Disruptions & runaway electrons — disruptions, runaway electrons and SPI mitigation as a requirement · DOI 22132102
2.13 Alpha / fast-product confinement — 99.98% alpha confinement (ASCOT5) · DOI 22132106
2.15 Ideal-MHD stability — ideal-MHD stability, Troyon margin, Mercier/ballooning · DOI 22132110
2.23 Current drive & bootstrap — holding 9.66 MA without a solenoid · DOI 22132145
2.24 Solenoid-free startup — solenoid-free startup and ramp · DOI 22132147
2.39 Admissible operating-window synthesis — the admissible operating window · DOI 22132231

Burner physics

2.5 RF alpha-channeling — RF alpha-channeling — the efficiency lever · DOI 22132168
2.6 DCLC loss-cone microstability — the DCLC warm-fill knee, stated honestly · DOI 22132170
2.9 Red-team of the binding gates — red-team of the binding gates · DOI 22132187
2.29 Synchrotron radiation transport — synchrotron transport and the max-gain temperature · DOI 22132195
2.33 ECH thermal-barrier decoupling — the ECH thermal-barrier plug mechanism · DOI 22132209
2.34 MetroVolt HVDC-native power interface — the HVDC-native power-electronics interface · DOI 22132211

Materials, magnets & structures

2.1 Centrepost & magnet integrity — consumable-cartridge centrepost & magnet integrity · DOI 22132096
2.3 Centrepost fluence & shielding — centrepost fluence & shielding · DOI 22132162
2.17 Low-activation alloys — near-neutral low-activation refractory alloys · DOI 22132119
2.19 Flowing-lithium vapor-box divertor — the flowing-lithium vapor-box divertor · DOI 22132127
2.20 Quench detection & NV sensing — in-winding NV-diamond quench sensing · DOI 22132131
2.30 First-wall neutron-damage lifetime — first-wall neutron-damage lifetime · DOI 22132198
2.32 REBCO conductor materials under irradiation — REBCO conductor under irradiation · DOI 22132203
2.38 ML-potential materials-discovery pipeline — the ML-potential materials-discovery pipeline · DOI 22132229
2.16 Divertor portfolio — the divertor detachment portfolio · DOI 22132115

Fuel cycle, isotopes, safety & supply

2.2 Blanket tritium-breeding lever — the blanket tritium-breeding lever · DOI 22132129
2.14 Helium-3 fuel cycle — the He-3 fuel cycle (lunar-gated closure) · DOI 22132108
2.18 Tritium fuel cycle & breeding ladder — the tritium breeding ladder 0.29 → 0.9 → ~2.0 kg-T/yr · DOI 22132123
2.31 Nuclear-data UQ for the TBR — the credible error bar on the TBR — nuclear-data uncertainty · DOI 22132200
2.25 Isotopes & 14 MeV applications — 14-MeV isotope & applications platform · DOI 22132152
2.21 Inherent-safety & environmental case — the inherent-safety & environmental case · DOI 22132133
2.22 Licensing (byproduct material, Part 30) — byproduct-material (Part 30) licensing · DOI 22132139
2.35 Supply chain & fleet buildability — supply chain & fleet buildability · DOI 22132221

Methods, verification & the platform

2.4 Verification & validation (47-code provenance) — verification & validation — 47-code provenance · DOI 22132164
2.7 Control-barrier safety clamp — the certified control-barrier safety clamp · DOI 22132176
2.8 Quantum computing for fusion — quantum computing for fusion, honestly · DOI 22132180
2.28 Tensor-network kinetic solver — the tensor-network kinetic solver · DOI 22132192
2.26 Integrated platform overview — the integrated three-product platform overview · DOI 22132156
2.27 Advanced-fuel scorecard — the advanced-fuel scorecard · DOI 22132158
2.36 Centrepost prognostics & RUL control — centrepost prognostics & remaining-useful-life control · DOI 22132223
2.37 Maturity-gated actuation authority — maturity-gated actuation authority · DOI 22132227

Data & reproduction

The public edition is deposited on Zenodo at 10.5281/zenodo.22133057. Each analysis re-runs on established, community-maintained codes with a deterministic seed and is cross-checked against 19 frozen anchors; the companion papers above carry the per-domain data and code. Economics are held in a separate internal edition and appear in no public deposit — this record carries physics only.

Explore the platform

Complete Volume — the full combined edition of the series
Publications index — all 44 papers in one place
Interactive 3D model — the machine, component by component
Live physics validator — reproduce the frozen anchors in-browser
Kronos toolkit — the open-source physics engine
Zenodo community — every deposit in the series