Fuse Energy's Neutron Record: A Fusion First, Not Net Power

Fuse Energy Technologies says its FAETON-X machine produced 1.27 trillion fusion neutrons in a single shot — which would be the highest neutron yield ever publicly documented by a private fusion company, a range previously reported only by US national laboratories. The claim, published August 11 as a preprint that has not yet been peer-reviewed, is a real milestone for a seven-year-old startup — but it is a record in neutron output, not energy: run the numbers and the shot released roughly one joule of fusion energy from a million joules of stored electricity.

What Fuse Energy announced

FAETON-X is a dense plasma focus (DPF) — a pulsed-power machine, not a reactor. According to the company’s preprint, a 500-microfarad capacitor bank stores one megajoule at a 65-kilovolt charge, then dumps it through the device in a single pulse peaking above 4.5 million amperes. In its best shot at 10 torr of deuterium fill gas, the machine’s diagnostics recorded (1.27±0.27)×10¹² neutrons. The same document notes the device is “not yet fully conditioned” and typically produces around 5×10¹¹ neutrons per shot — the headline number is the peak, not the norm.

The company’s sharpest boast is efficiency. Fuse says the shot beat the neutrons-per-megajoule of the world’s largest government plasma-focus facilities — about 40% more than Lawrence Livermore National Laboratory’s MJOLNIR, a megajoule-class DPF built for flash neutron radiography. That comparison is the company’s own framing, but the underlying LLNL figures are public: MJOLNIR has reported up to 1.2×10¹² neutrons per discharge from 1.3 megajoules stored, which works out to FAETON-X claiming about 38% more neutrons per megajoule — the arithmetic checks, with the caveat that no independent party has benchmarked the two machines head-to-head, and MJOLNIR is an active program rated for 2 megajoules it hasn’t yet used in full.

Worth flagging plainly: the preprint is self-published research, its co-authors include Fuse’s founder and CEO J.C. Btaiche, and as of this writing we found no independent replication or physicist critique of the result. The company has some track record here — the paper on its smaller predecessor device, FAETON-I, later cleared peer review at Scientific Reports. Until this one does the same, treat every FAETON-X number as a company claim.

What a dense plasma focus actually does

The DPF is one of fusion’s oldest machines, invented independently in the early 1960s by J.W. Mather in the US and N.V. Filippov in the USSR, with roots tracing to 1954. A capacitor bank discharges across two coaxial electrodes, ionizing the fill gas into a current-carrying sheath; electromagnetic force accelerates that sheath down the barrel; and at the muzzle it collapses inward into a “pinch” — a fleeting, extremely hot and dense plasma column. With deuterium in the chamber, the pinch produces a nanosecond-scale burst of fusion neutrons.

The field’s founders hoped this would become a power plant. Scaling studies killed that hope decades ago: by the 1980s the DPF had settled into a different role — a compact, relatively cheap source of intense neutron and X-ray pulses. That is not a consolation prize. Pulsed neutron sources are how you test whether electronics, satellites and defense hardware survive radiation, and US national security agencies treat the capability as critical: the Nevada National Security Site fields its own Gemini DPF — roughly 8×10¹¹ neutrons per pulse from a 2-megajoule machine — for some of the highest-priority missions in its nuclear-arsenal custodianship program.

The math the headline leaves out

A trillion of anything sounds enormous, so it is worth doing the energy arithmetic the coverage skips. Deuterium-deuterium fusion has two equally likely outcomes, only one of which emits a neutron, and the two branches release 4.03 and 3.27 MeV respectively — call it 3.65 MeV per reaction on average. So 1.27 trillion detected neutrons implies roughly 2.5 trillion fusion reactions: about 1.5 joules of fusion energy, from a shot fed by 1,000,000 joules of stored charge. That is an energy gain on the order of one millionth — six orders of magnitude short of scientific breakeven, and the company, to its credit, claims nothing else.

For calibration, here is where the record sits among the machines it invites comparison with:

The table’s last row is the trap to avoid. The National Ignition Facility’s numbers describe energy gain from laser fusion — ignition, first achieved in December 2022 at 3.15 MJ out from 2.05 MJ in — a fundamentally different physics goal than a plasma focus chasing neutron counts. And even NIF’s gain is target-level, not plant-level: scientific breakeven counts only the energy delivered to the fuel, while the lasers delivering it are around 1% efficient at the wall plug. Zap’s FuZE-3, meanwhile, is the closest technical cousin to a DPF among funded startups — a pulsed Z-pinch — and its company-stated yields sit four orders of magnitude below the FAETON-X claim, though on a much smaller machine chasing sustained plasma conditions rather than single-shot output.

Neutrons that pay the bills

Why celebrate a neutron record at all? Because neutrons, unlike fusion electricity, are sellable today — and Fuse’s business is built on exactly that. The company, founded in 2019 by Btaiche and headquartered in San Leandro, California, sells radiation-effects testing to defense and space customers as its bridge to an eventual power play. It holds contracts with the US Air Force, supported by the Defense Threat Reduction Agency and the Air Force Nuclear Weapons Center, and told POWER it expects about $10 million in testing revenue over two years and projects over $300 million across the next decade from its next machine — projections that are, again, the company’s own. More neutrons per megajoule means a harsher, more realistic radiation environment per shot, which is the product.

That “sell the byproduct first” model is unusual in a sector otherwise selling promises of future power — the beat our Chips & Hardware hub tracks is full of machinery monetized long before its grand ambition. The contrast with the industry’s front-runners is stark. Commonwealth Fusion Systems raised another $1 billion in July — the sector’s largest-ever round, taking its total to $4 billion — while its revenue anchor is a 200 MW power purchase agreement Google signed for a plant targeting the grid in the early 2030s. Google’s appetite for frontier hedges is a pattern we’ve covered before in its $94 billion SpaceX position, and it is no coincidence the buyers of future fusion power are the same companies straining grids with AI data-center buildouts. Helion sold Microsoft 50 MW deliverable by 2028 — a date one plasma physicist told E&E News he would be “happily surprised” to see met. General Fusion, near collapse in 2025, became the first publicly traded fusion company in July via a SPAC. Against that backdrop, the Fusion Industry Association counts $14.24 billion raised across 56 companies — a record $4.48 billion of it in the year to July 2026 — while ITER, the intergovernmental flagship, works to a rebaselined schedule that dropped its 2025 first-plasma target and points to research operations from 2034.

What to watch

Three things will tell you whether this record matters in six months. First, peer review: the FAETON-X paper following its FAETON-I predecessor into a journal would move the yield claim from “company-reported” to “checked” — and any independent benchmarking against MJOLNIR’s next campaign would be better still, since LLNL’s machine has headroom to 2 megajoules that could retake the per-megajoule crown quickly. Second, conditioning: whether Fuse’s routine yields climb from 5×10¹¹ toward the peak — a testing business is priced on the shot you can sell every day, not the best shot ever fired. Third, revenue conversion: defense contracts turning into the projected testing backlog would validate neutrons-as-a-product the way specialized silicon found its market in AI inference — niche hardware winning by doing one job better than general-purpose giants.

The honest frame for readers: nothing announced this week brings fusion electricity closer on the calendar. The industry’s own survey consensus is grid power between 2030 and 2035, with $77 billion still to raise, and independent voices put a meaningful grid contribution a decade past that. What the FAETON-X shot does show is a private company matching national-laboratory hardware on a metric the national-security market pays for now — and in a field that has run on promissory notes for seventy years, a machine with paying customers is its own kind of breakthrough.