Rehearsing a fusion pulse before it happens
Google DeepMind and Commonwealth Fusion Systems announced a partnership in which CFS uses the open-source TORAX plasma simulator in daily workflows to test and refine SPARC operating plans before the tokamak is turned on.
- Maturity
- Emerging, stage 2 of 4
- Support
- 1 source · institution
- Evidence detail
- How we know ↓
Fusion teams must choose operating plans that balance plasma performance with machine limits before they can learn from a new device.
Exploring many combinations of magnetic coils, fuel, heating, and heat-load constraints by hand is slow and cannot replace physical commissioning.
TORAX is a differentiable plasma simulator that can integrate AI-powered models; the partners are investigating reinforcement learning and evolutionary search for pulse optimization and real-time control strategies.
What was shown, and what wasn’t
Shown
CFS says TORAX has become a linchpin in its daily workflows and that it saves time setting up and running SPARC simulation environments.
Not shown · limits
- SPARC had not yet been turned on; simulated plans do not demonstrate net fusion energy or electricity production.
- The source describes collaboration and workflow use, not validated real-time control on SPARC.
Bright editorial interpretation
Fast virtual experiments may help engineers focus on promising, constraint-aware operating plans before live testing.
Still open
How accurately will TORAX calibrate to SPARC data?
Can proposed controllers remain safe and useful under real device disturbances and heat loads?
How we know1 source · checked 2026-09-07 · no corrections
Original sources
- Bringing AI to the next generation of fusion energy ↗ · institution
Institutions: Google DeepMind · Commonwealth Fusion Systems
- Maturity
- Emerging
- Source published
- 2025-10-16
- Captured
- 2026-09-07
- Last source review
- 2026-09-07
- Editorial method
- AI-assisted source review
Bright compared this account with the linked original and supporting sources and kept reported, budgeted, projected, and observed claims distinct. Bright did not independently audit the underlying records.
Maturity describes the tested or operational setting. Confidence describes support for the particular claim; one does not determine the other.
Revision & correction history
2026-09-07 · Fast virtual experiments may help engineers focus on promising, constraint-aware operating plans before live testing.
No corrections recorded.
Keep exploring
Explore the shared question in another setting. These connections do not imply replication.