Is Lab-Grown Gold Real?

Why gold is not like lab-grown diamonds

A lab-grown diamond is carbon atoms arranged into a crystal. That is chemistry, and factories do it cheaply at scale.

Gold is an element. To make it you have to change the nucleus of another atom, which takes a particle accelerator or a nuclear reactor. No chemical process can do it, and the gold that comes out of a nuclear process starts out radioactive.

What CERN did (May 2025)

On 8 May 2025 CERN reported that its ALICE experiment had measured lead turning into gold in near-miss collisions at the Large Hadron Collider. The paper is in Physical Review C.

Across the LHC's 2015 to 2018 run, about 86 billion gold nuclei were made at the four main experiments. That is 29 picograms, or 29 trillionths of a gram. CERN's own summary: "The gold exists for just a tiny fraction of a second."

This is a measured, peer-reviewed physics result. It has nothing to do with producing gold you could hold.

What Marathon Fusion claims (2025 and October 2026)

In July 2025 the San Francisco startup Marathon Fusion posted a preprint (arXiv 2507.13461, last revised 3 Aug 2025) proposing that a fusion power plant could turn mercury-198 into gold-197. Fast neutrons from the reactor would knock a neutron out of the mercury, and the result decays into gold. It has not been peer reviewed.

The paper estimates "about 2t/GW_th/yr", meaning about 2 tonnes of gold per year for each gigawatt of reactor heat. The company's website, as quoted by Kitco, puts it at five thousand kilograms per gigawatt of electricity. Both numbers come from computer simulations of reactors that do not exist yet.

Kitco reported that the gold would need storage for seven to seventeen years before its radioactivity falls, citing the paper. It also reported the company had raised $5.9 million plus about $4 million in U.S. government grants, citing the Financial Times.

On 21 Oct 2026 the company said on X that it had turned mercury into stable gold using fusion neutrons. Runtimewire, reporting the same day, said the thread included no experimental data and the claim had not been independently confirmed. The amount of gold was not given. Until data is published, treat it as a claim.

It has been done before, at absurd cost

In 1980 Glenn Seaborg's team at Lawrence Berkeley fired carbon and neon nuclei at bismuth foil and made trace amounts of gold. Seaborg's estimate, as reported by Scientific American: "more than one quadrillion dollars per ounce".

Making gold has been possible for decades. Making it cheaply, in quantity, and safely is the part nobody has done.

Could it affect gold prices?

Not any time soon. Gold mines produced 3,672 tonnes in 2025 and total supply including recycling was 5,002 tonnes, according to the World Gold Council.

Our rough arithmetic, not a source's estimate: one 1-gigawatt fusion plant making 5 tonnes a year would add about 0.14% to mine supply. It would take dozens of such plants, none of which exist yet, before the extra supply would matter. There is also far more gold already above ground (the 219,891 tonne estimate used on our gold vs Bitcoin market cap page) than any new source would add in a year.

If it ever worked at scale, the gold would be chemically identical to mined gold. A melt value calculator or an assay cannot tell the difference, so buyers would not pay less for it.

What "lab-grown gold" jewelry really is

Some jewelers now use the phrase. The ones we checked describe starting from recycled or responsibly sourced gold, which is normal refining with new marketing. If a seller says "lab-grown gold", ask what the feedstock is and check the karat stamp like any other piece.