Inside the ARC physics basis for CFS’s 400 MW fusion power plant

The ARC V3A design targets a fusion gain of Q = 50 and is projected to deliver 400 MW of net electricity, with physics basis modelling published in the Journal of Plasma Physics in June 2026

(Image courtesy of Commonwealth Fusion Systems)

Commonwealth Fusion Systems has published five peer-reviewed papers and an editorial in the Journal of Plasma Physics, setting out the physics basis for its ARC fusion power plant, a design still in development ahead of planned construction in the early 2030s. The collection addresses core plasma behaviour, disruption strategy, power exhaust and magnetohydrodynamic stability of a high-field tokamak targeting 400 MW of net electricity from a site in Chesterfield County, Virginia. It represents the first full physics-basis package published by a private fusion company targeting grid-scale net electricity. Separately, CFS has announced a commercial offtake agreement with Google to purchase 200 MW of ARC’s output, roughly half the plant’s target capacity.

ARC V3A parameters and fusion power projections

The papers describe ARC Version 3A, the current design iteration. It runs at 11.4 T on axis, carries 12 MA of plasma current and has a major radius of 4.62 m. The machine targets a fusion gain of Q = 50, with modelling projecting around 1.13 GW of deuterium-tritium fusion power and 400 MW or more of net electricity to the grid. The core plasma operates above 100 million degrees Celsius, sustained by 250 MW of heating power drawn mostly from alpha particle self-heating. Plasma pulses run for 900 seconds, with 60-second intervals between flattops.

Performance projections span a wide range depending on model fidelity and input assumptions. Integrated codes including TRANSP, ASTRA and TORAX project fusion power between roughly 879 and 1295 MW, while higher-fidelity gyrokinetic modelling with CGYRO and PORTALS returns around 677 MW under nominal assumptions. That spread reflects sensitivity to profile peaking, pedestal assumptions and model fidelity rather than a single agreed projection. The papers conclude that modelling gives confidence in ARC performance approaching 1 GW, while density and temperature peaking remain open questions that SPARC operation may help resolve.

ARC physics basis for plasma stability and disruptions

Disruptions carry the most acute hardware risk in a high-field tokamak, threatening the first wall with intense heat and imposing large electromagnetic loads on all internal components. The disruption paper models both mitigated and unmitigated loads, projecting them within a factor of two of those expected in SPARC, and frames that smaller machine as the primary tool for calibrating models and testing strategies before ARC enters operation. The design goal is disruption-free running. Its pragmatic target is to withstand one mitigated disruption per day, with restart in tens of seconds and only brief intervals between pulses. Massive gas injection is the baseline mitigation approach, with a runaway electron mitigation coil proposed to address gaps in that baseline. The papers are clear that disruption risk management depends substantially on what SPARC will demonstrate.

Managing heat exhaust requires ARC to radiate most of the power crossing the last-closed flux surface before it reaches the plasma-facing components, a requirement the exhaust paper addresses through argon or neon seeding to drive the divertor into detachment. Up-down-symmetric divertors with tightly baffled outer legs are designed to hold that detachment state under normal operating conditions, and efficient pumping is projected to keep helium ash below 2% in the core. On magnetohydrodynamic stability, analysis places the baseline scenario well within the stable region, with the 2/1 and 3/2 tearing modes found linearly stable. Vertical displacement events require no in-vessel coil, with the poloidal shaping coils taking that load. Two rows of off-midplane coils are identified as a viable solution for error field correction, with the same coils capable of suppressing edge-localised modes at practical coil currents.

SPARC and the path to the ARC fusion power plant

Throughout the collection, SPARC functions as the essential validation step rather than a parallel programme. CFS is building it in Devens, Massachusetts, targeting a fusion gain greater than one in 2027. Because ARC uses a fully liquid immersion blanket around a vacuum vessel designed for replacement every one to two years, CFS can hold the final vessel shape decision until late in the design cycle, directly overlapping with early SPARC operation. That architecture means first-campaign data can feed into the first ARC unit and each subsequent vessel swap. The papers present ARC V3A as a credible, model-backed design with specific remaining risks to be retired through SPARC, and CFS publishes the work openly, treating peer review as the proper test for physics at this frontier. Commercial deployment targets the early 2030s.

Stay ahead in the electrification revolution. Explore more breakthroughs from leading innovators in electric commercial transport – visit our news page.