How to get a fusion energy job in 2026

Why does a $200M fusion raise matter for your job search?

Fusion hiring is now tied to construction schedules rather than research grants, and that changes who gets hired. On 6 October 2026, Type One Energy announced a $200 million Series B led by Breakthrough Energy Ventures and Clutterbuck Capital, with Lowercarbon Capital, Siemens Energy Ventures and SiteGround Capital joining the round. That takes the Knoxville company past $400 million raised in total. The money funds the first phase of Project Infinity: Infinity One, a prototype stellarator due for commissioning in 2029, followed by Infinity Two, a 400-megawatt plant targeting power on the grid in 2034. Both machines are slated for the Tennessee Valley Authority's Bull Run site. Read those dates as a staffing plan — every milestone needs engineers, technicians and schedulers hired years ahead of it.

The structure of the programme tells you where the jobs actually sit. Type One plans to design the plant and many of its components in-house, then rely on a bespoke network of suppliers to build them — going further down that road than most fusion developers, who outsource only part of the machine. AECOM is handling engineering for Infinity Two, and the high-temperature superconducting magnet technology is licensed from Commonwealth Fusion Systems. For anyone switching fields, that's encouraging. One fusion programme spreads paid work across machine shops, magnet winders, cryogenic plant specialists, vacuum houses and engineering consultancies. You don't have to land at the fusion company itself to be working in fusion, and the supplier route is usually the easier door to push open first.

This isn't one company's hiring spree, either. The fusion industry logged a record $4.48 billion in private funding in the 12 months to July 2026, bringing the total past $14.24 billion since 2021, and the sector now employs more than 16,000 people. Supply-chain jobs supported by fusion developers have climbed from roughly 3,000 in 2023 to 10,497 in 2026. Here's the part nobody tells career changers: engineers already outnumber scientists in that workforce by about 40% to 23%, and the industry expects manufacturing and operations headcount to grow faster than physics headcount from here. If you've assumed fusion is closed to anyone without a plasma physics doctorate, the hiring data says flatly otherwise.

Fusion hiring is now tied to construction schedules rather than research grants, and that changes who gets hired.

Which fusion energy jobs are companies actually hiring for?

Fusion companies hire overwhelmingly for hardware disciplines: vacuum systems, cryogenics, high-temperature superconducting magnets, pulsed power and power electronics, controls and instrumentation, precision welding and fabrication, radiological licensing, and project controls. Plasma physics roles exist, but they're a small share of open headcount. Scan a dozen 2026 listings and the pattern is unmistakable — these are mechanical, electrical and manufacturing jobs attached to an unusual machine. A stellarator or tokamak still needs a vessel that holds ultra-high vacuum, coils that stay superconducting under load, helium loops that keep them cold, a power train that fires on cue, instrumentation that tells you what happened, and someone who can prove to a regulator that every gram of radioactive inventory is accounted for.

The specifics inside the postings are your roadmap. One magnet cryogenics lead role advertised this autumn asked for 10 to 15 years of industry experience owning cryogenic design, thermal-hydraulic modelling and forced-flow cooling distribution for cabled or stacked-tape HTS coils. Elsewhere you'll find a cryogenics and vacuum engineer building test infrastructure for magnet prototypes, a mechanical engineer hired explicitly for ASME code work, a project scheduler at an isotope company working next to tritium systems, a multiphysics modelling engineer coupling thermal, structural, tritium-transport and electromagnetic models, and fuel-cycle process modellers expected to know ASPEN and CFD. Not one of those descriptions requires you to have touched a fusion machine before. They require evidence you've done the underlying engineering somewhere.

Pay is competitive and the spread is wide. Pulsed power postings in 2026 ran from roughly $97,000 for an operations engineer troubleshooting capacitor banks and pulse-forming networks, up to $150,000–$270,000 for a subject-matter expert, while senior research scientist and engineer bands at one inertial-fusion company were advertised at $200,000–$350,000. Two practical warnings before you start applying. Some national-lab pulsed power roles require a DOE Q clearance, which can add months to your start date, so weigh lab against startup with that in mind. And application windows are brutally short — one principal engineering role posted on 29 September closed on 3 October. Check company career pages directly and weekly instead of waiting for an aggregator to catch up.

Which nuclear, oil and gas or semiconductor skills transfer one-to-one?

Three backgrounds transfer almost one-to-one. Semiconductor equipment engineering maps straight onto fusion's vacuum work: ultra-high vacuum practice, leak detection, contamination control, pumping systems and tool commissioning are the same craft in a different chamber. Fission experience transfers through radiation protection, health physics, materials licensing, tritium and waste handling, and nuclear-qualified welding procedures. Oil and gas transfers through large capital project execution — commissioning, QA/QC, long-lead procurement, schedule and cost control on safety-critical builds — and, if you've worked LNG, through genuine cryogenic process knowledge. A nuclear to fusion career change is often the shortest of the three hops, because the regulatory vocabulary and the safety culture already live in your head and don't need re-teaching.

Be honest about what doesn't carry over, because a good recruiter spots the gap faster than you can talk around it. Deep reactor-licensing experience under the fission construction and operating regime is less portable than it feels: the 2026 US framework treats fusion machines as accelerators under byproduct-material rules, so health physics and materials licensing beat construction permits. Drilling operations, reservoir engineering and gas processing are specialised in ways fusion simply has no use for. Semiconductor process integration doesn't travel the way equipment engineering does. The fix isn't to hide those years. It's to re-weight the document so the transferable third of your experience owns the top third of page one, and the rest reads as supporting context.

If you lead with one thing, make it systems engineering. Fusion plants fail at the interfaces, not the components: magnets, blanket and heat-extraction loops, tritium processing, vacuum, pulsed power, diagnostics, controls and the turbine island each work acceptably alone, and the hard job is making them work together safely at high availability. Anyone who has run requirements management, interface control documents and verification campaigns across mechanical, electrical, thermal and software boundaries — in aerospace, process plants, fission, defence or semiconductor capital equipment — is selling exactly what's scarce. That's the strongest transferable skill in the fusion industry today, and it's almost never the thing career changers put in their opening bullet. Move it up.

Which certifications and retraining are actually worth the time?

Almost no fusion-specific certification exists yet, so the credentials worth your time are transferable engineering ones. In rough order of return: welding and fabrication qualifications, including nuclear-qualified procedures and working knowledge of the ASME Boiler and Pressure Vessel Code; radiation protection and health physics training; project controls credentials for schedulers and cost engineers; and short courses in vacuum technology and cryogenics, which are genuinely learnable in weeks rather than years. On the software side, employers name ANSYS, COMSOL, Python, MATLAB and LabVIEW in listing after listing, plus ASPEN for fuel-cycle process work. Don't collect all of them. Pick the two that recur most often in the specific jobs you want, and get demonstrably good.

The regulatory lane is the most underrated opening in energy career change 2026. In February 2026 the US regulator proposed a framework classifying fusion machines as a category of particle accelerator and licensing them under byproduct-material rules rather than the reactor regime, following the direction set by the ADVANCE Act of 2024. Comments closed in May, and draft programme-specific licensing guidance went out for comment alongside the rule. Here's the detail that creates jobs: of the 39 Agreement States, 38 have taken authority over that material category, so state radiation control programme experience may count for more than federal reactor experience. Tennessee had already issued Type One a fusion-specific byproduct material licence in August 2026.

What's rarely worth it: a second master's degree. Fusion companies are hiring against fixed schedules, and two years out of the workforce costs you the window you're trying to catch. Cheaper routes in are hiding in plain sight — co-op and internship programmes that explicitly accept candidates from bachelor's through PhD level, including one 26-week on-site cryogenics co-op this year; contract technician and assembly roles that convert to staff; national labs used deliberately as a training ground; and supplier-side positions in magnet winding, precision machining or cryogenic plant. Budget six to eighteen months for a deliberate switch: one quarter to re-map and rewrite your CV and assemble a small portfolio of relevant hardware work, then two or three quarters of targeted applications.

How should you structure a fusion CV with no fusion experience?

Fusion hires on hardware proof rather than job titles, so build the CV around equipment, standards and measured outcomes. Put a capability block directly under your summary naming the machines, systems and codes you've genuinely worked on — pump types, chamber pressures, helium plant, capacitor banks, PLC families, weld codes, simulation tools — because that block is what a screener and a parser both read first. Then let the experience bullets prove it with numbers: base pressure achieved, leak rate, uptime, cycle time, budget, headcount, weeks of schedule recovered. A title like "Process Equipment Engineer III" tells a fusion hiring manager nothing. "Qualified a helium leak-check procedure at the 10⁻⁹ mbar·l/s level across 40 flanged joints" tells them everything they need.

The hard part of a career-change CV is vocabulary, not substance. Fusion listings use words your current CV doesn't: forced-flow cooling, cryostat, stacked-tape conductor, pulse-forming network, ultra-high vacuum, tritium accountability, interface control. If you wrote "chilled water loop" where they wrote "thermal-hydraulic cooling distribution", an automated screen scores you as a miss even though you've done the work for nine years. Take one target posting, list every technical noun in it, and decide line by line whether you hold a true equivalent — then use their word, not your old employer's internal term. A CV review built for career changers does that translation faster, surfacing the transferable skills you've buried and the titles that misread in a new field.

Then check the file an applicant tracking system actually sees, because a beautiful CV that parses into fragments never reaches the magnet team. Single column, no text boxes or icons, real dates, standard headings — the two-column template that looks sharp in a design tool is the usual culprit. You can run a free CV analysis to get a score out of 100, five category scores covering experience, tech stack, impact and ownership, clarity and structure, and ATS compatibility, plus a visual layout check. Paste the fusion job description in and the analysis flags which of its skills it can't find in your document — which is exactly the gap a career changer needs mapped before applying, not after four silent rejections.

Frequently asked questions