Commonwealth Fusion Systems installed the first of 18 D-shaped toroidal field magnets for its SPARC tokamak, a milestone the company revealed at CES on January 6, 2026.1 TechCrunch 2026-01-06 CFS installed the first of 18 D-shaped 20 tesla magnets for SPARC at CES, cooled near minus 253C carrying over 30,000 amps, with a cryostat installed March 2025 and an Nvidia and Siemens digital twin partnership. Open source Each magnet weighs roughly 24 tons, generates a field above 20 tesla, and is cooled to about minus 253 degrees Celsius to carry more than 30,000 amps.1 TechCrunch 2026-01-06 CFS installed the first of 18 D-shaped 20 tesla magnets for SPARC at CES, cooled near minus 253C carrying over 30,000 amps, with a cryostat installed March 2025 and an Nvidia and Siemens digital twin partnership. Open source The stake is narrow and large at once: SPARC is the demonstration machine meant to show net energy from fusion, and its high temperature superconducting magnets are the single component the entire compact-tokamak approach depends on.3 Commonwealth Fusion Systems 2026-01-01 CFS 20 tesla HTS magnets, demonstrated in 2021, enable a smaller and lower cost tokamak, powering SPARC as the intended first net energy device and later the ARC power plant. Open source We assess with high confidence that magnet installation, not plasma physics, is now the pacing item for the fusion investment thesis, because the field strength is already proven and the remaining risk is manufacturing and assembly at cadence.

What is actually being tested

The physics case for CFS was settled in 2021, when the company demonstrated a 20 tesla large-bore HTS magnet.3 Commonwealth Fusion Systems 2026-01-01 CFS 20 tesla HTS magnets, demonstrated in 2021, enable a smaller and lower cost tokamak, powering SPARC as the intended first net energy device and later the ARC power plant. Open source That number is the reason SPARC can be small. Field strength enters the fusion power equation steeply, so a stronger magnet lets a tokamak reach net energy at a fraction of the volume that older low-field designs required.3 Commonwealth Fusion Systems 2026-01-01 CFS 20 tesla HTS magnets, demonstrated in 2021, enable a smaller and lower cost tokamak, powering SPARC as the intended first net energy device and later the ARC power plant. Open source The bet the January installation tests is therefore not whether a 20 tesla coil works, but whether CFS can build 18 of them to spec and hang them in sequence. CEO Bob Mumgaard framed the process as incremental, saying the coils would go in through the first half of 2026.2 Interesting Engineering 2026-01-07 Each 24-ton magnet generates a 20 tesla field and CEO Bob Mumgaard said the remaining coils would be installed through the first half of 2026. Open source The cryostat that houses the assembly was installed in March 2025, so the machine has been accreting toward this point for the better part of a year.1 TechCrunch 2026-01-06 CFS installed the first of 18 D-shaped 20 tesla magnets for SPARC at CES, cooled near minus 253C carrying over 30,000 amps, with a cryostat installed March 2025 and an Nvidia and Siemens digital twin partnership. Open source

Alongside the hardware, CFS disclosed a partnership with Nvidia and Siemens to run a digital twin of the reactor, using Nvidia Omniverse libraries and Siemens design software to simulate the machine in parallel with the physical build.1 TechCrunch 2026-01-06 CFS installed the first of 18 D-shaped 20 tesla magnets for SPARC at CES, cooled near minus 253C carrying over 30,000 amps, with a cryostat installed March 2025 and an Nvidia and Siemens digital twin partnership. Open source That is a tell about where the difficulty now sits. A company confident in its magnet physics but wary of first-of-a-kind operation buys simulation to compress the learning curve on parameters it cannot cheaply test on the real device.

Second order effects and the ledger

The shift from a hero magnet to a magnet production line changes who is exposed. It moves fusion risk from a laboratory question, which favors physicists, to an industrial one, which favors whoever can hold tolerances on niobium-free HTS tape and 24-ton assemblies. We assess with moderate confidence that this reframing helps CFS relative to rivals whose designs still require larger or less mature magnet systems, because CFS has already retired the field-strength unknown that others must still clear.3 Commonwealth Fusion Systems 2026-01-01 CFS 20 tesla HTS magnets, demonstrated in 2021, enable a smaller and lower cost tokamak, powering SPARC as the intended first net energy device and later the ARC power plant. Open source

Who gains: HTS tape suppliers and the broader superconductor supply chain, whose order books now depend on a machine that needs 18 coils and a follow-on power plant, ARC, that needs more.3 Commonwealth Fusion Systems 2026-01-01 CFS 20 tesla HTS magnets, demonstrated in 2021, enable a smaller and lower cost tokamak, powering SPARC as the intended first net energy device and later the ARC power plant. Open source Nvidia and Siemens gain a reference deployment of industrial simulation in a frontier-energy setting, useful marketing well beyond fusion.1 TechCrunch 2026-01-06 CFS installed the first of 18 D-shaped 20 tesla magnets for SPARC at CES, cooled near minus 253C carrying over 30,000 amps, with a cryostat installed March 2025 and an Nvidia and Siemens digital twin partnership. Open source Who loses on a relative basis: fusion approaches that are not tokamaks and cannot point to a comparably concrete assembly schedule, because a machine visibly accreting magnets is easier to underwrite than one still arguing physics. The loss is reputational and financial, in the competition for the next round of private capital, not technical.

The counter-case

The main risk to this read is that installing the first magnet says little about installing the eighteenth. Assembly of large superconducting coils is where schedules historically slip, and a single coil hung in January does not prove the cadence Mumgaard described will hold through the first half of the year.2 Interesting Engineering 2026-01-07 Each 24-ton magnet generates a 20 tesla field and CEO Bob Mumgaard said the remaining coils would be installed through the first half of 2026. Open source First plasma remains a forward target rather than an achieved result, and net energy is a further step beyond first plasma.1 TechCrunch 2026-01-06 CFS installed the first of 18 D-shaped 20 tesla magnets for SPARC at CES, cooled near minus 253C carrying over 30,000 amps, with a cryostat installed March 2025 and an Nvidia and Siemens digital twin partnership. Open source If the installation rate stalls, the very concreteness that makes the thesis attractive becomes the metric that exposes it. The digital twin deal can also be read less charitably: heavy investment in simulation can signal that the team expects surprises it cannot yet bound.1 TechCrunch 2026-01-06 CFS installed the first of 18 D-shaped 20 tesla magnets for SPARC at CES, cooled near minus 253C carrying over 30,000 amps, with a cryostat installed March 2025 and an Nvidia and Siemens digital twin partnership. Open source For the optimistic assessment to fail, CFS would have to miss its own stated first-half installation window by a wide margin, which would reset every downstream date.

What to watch

  • All 18 magnets installed on Mumgaard's timeline. If the full toroidal field set is hung by the end of summer 2026 as described, the manufacturing thesis is tracking; slippage into 2027 would signal that coil assembly, not physics, is the binding constraint.2 Interesting Engineering 2026-01-07 Each 24-ton magnet generates a 20 tesla field and CEO Bob Mumgaard said the remaining coils would be installed through the first half of 2026. Open source
  • First plasma date holds. A first-plasma milestone in the stated window would validate the schedule; a quiet deferral would tell the market the machine is harder to commission than the magnet demo implied.1 TechCrunch 2026-01-06 CFS installed the first of 18 D-shaped 20 tesla magnets for SPARC at CES, cooled near minus 253C carrying over 30,000 amps, with a cryostat installed March 2025 and an Nvidia and Siemens digital twin partnership. Open source
  • Net energy claim, independently checked. Watch whether any net-energy result carries external validation rather than a company statement alone, the same discipline that separated credible fusion claims from marketing in prior years.3 Commonwealth Fusion Systems 2026-01-01 CFS 20 tesla HTS magnets, demonstrated in 2021, enable a smaller and lower cost tokamak, powering SPARC as the intended first net energy device and later the ARC power plant. Open source
  • Digital twin shows up in operations. If CFS reports using the Nvidia and Siemens simulation to set real commissioning parameters, the partnership is load-bearing; if it stays a press item, it was mostly signaling.1 TechCrunch 2026-01-06 CFS installed the first of 18 D-shaped 20 tesla magnets for SPARC at CES, cooled near minus 253C carrying over 30,000 amps, with a cryostat installed March 2025 and an Nvidia and Siemens digital twin partnership. Open source

The reframing to watch is the one already underway: fusion's hardest question at CFS is no longer can the magnet reach 20 tesla, but can the company build eighteen of them on time, which is the kind of problem investors know how to price.