A San Francisco fusion startup says it has turned mercury into stable gold. Marathon Fusion announced on X on October 21 that it had done it “last month,” at a rate it says beat a CERN record “by a factor of 10.”
The company hasn’t said how much gold it made. It hasn’t published the experiment’s data, and no one outside the company has confirmed the result.
The physics behind the claim is real. The open question is the scale, and the CERN comparison doesn’t settle it.
What Marathon Fusion says it did
“Last month, we demonstrated high-rate alchemy: turning mercury into real, stable gold, beating CERN’s record by a factor of 10,” the company posted on X.
Marathon Fusion was founded in 2023 by Kyle Schiller and Adam Rutkowski, a former SpaceX engineer, Euronews reported. Alongside the post, the company shared gamma-ray spectra that it says show the signature of mercury-197 decaying into gold, Business Today reported.
A spectrum like that can show a reaction happened. It can’t show how much gold came out. The company hasn’t disclosed the yield, the purity, or what device produced the neutrons.
How fusion neutrons could turn mercury into gold
The method comes from a 2025 paper by three Marathon Fusion researchers, posted on arXiv and not peer reviewed. It has three steps:
- A deuterium-tritium fusion reaction releases fast neutrons.
- A neutron hits mercury-198, a stable form of mercury, and knocks out two neutrons. That leaves mercury-197, which is radioactive.
- Mercury-197 decays into gold-197, the only stable form of gold, within days.
The paper modeled the process inside the blanket of a future fusion power plant. It estimated about 2 tonnes of gold a year for every gigawatt of the plant’s heat output. No such plant exists yet.
There are catches even on paper. Mercury-198 makes up only about 10% of natural mercury. Freshly made gold would also carry radioactive impurities and could need up to 18 years in storage before sale, according to the company, Euronews reported.
What did CERN do, and is “10 times” a fair comparison?
CERN’s ALICE experiment reported in May 2025 that the Large Hadron Collider turns lead into gold. When two lead nuclei narrowly miss each other, their electric fields can knock three protons out of one of them, leaving gold.
The numbers were tiny. The LHC’s four big experiments made about 86 billion gold nuclei from 2015 to 2018, or 29 picograms, CERN said. That’s 29 trillionths of a gram.
The collider now makes gold at up to about 89,000 nuclei a second. Each one exists “for just a tiny fraction of a second” before it hits the machine and breaks apart.
The two results aren’t the same thing. CERN’s gold was gold-203, an unstable form that falls apart almost at once. Marathon Fusion says it made stable gold-197, which lasts.
Marathon Fusion also hasn’t said which CERN number it beat by 10 times: total mass, nuclei per second, or something else.
Even the generous reading is small. If the claim means 10 times CERN’s top rate, that’s about 890,000 gold atoms a second.
Run nonstop for a year, that adds up to roughly 9 billionths of a gram, our calculation shows. An ounce of gold weighs about 31 grams.
We covered the CERN result in our 2025 science roundup.
Could lab-made gold ever move the gold market?
Not soon. The company’s own big numbers assume gigawatt-scale fusion plants. No fusion power plant yet supplies electricity to a grid, as projects like China’s “artificial sun” show.
For scale, mines produced 3,672 tonnes of gold in 2025, Euronews reported, citing the World Gold Council. Gold traded at $4,062 an ounce on Thursday, more than 27% below its January record. Central banks such as China’s have kept buying.
The comparison some investors reach for is lab-grown diamonds, whose prices collapsed once factories scaled up. “If lab-grown diamonds disrupted diamond prices, what could lab-grown gold do to gold prices?” Tushar Jain of Multicoin Capital asked on X, according to Business Today. That depends on a fusion industry that doesn’t exist yet.
What else is Marathon Fusion claiming?
Gold isn’t the company’s main business. It’s building fusion technology, and it says the same neutrons can make medical isotopes. It says it made copper-64 and copper-67, which are used in cancer imaging and treatment, by firing neutrons at zinc, Business Today reported.
In the same X thread, it said it had enriched isotopes in “the world’s most powerful plasma centrifuge,” beating a 40-year-old Soviet record. It also said it’s building a reactor in San Francisco. None of these claims has been independently checked either.
What would prove it?
The test is data. That means the gold yield from a known amount of mercury, the neutron source and dose, and measurements other labs can repeat. A peer-reviewed paper or an outside lab confirming the numbers would turn the claim into a result.
We found no independent scientist who had publicly reviewed Marathon Fusion’s data by Friday. Until that happens, the gold is a company claim, and a small one.
