Lab-made gold? A fusion startup puts scarcity back under the microscope

Marathon Fusion says it has turned mercury into stable gold. That is a long way from flooding the bullion market, but it raises an interesting question about what makes an asset scarce.

Turning ordinary metal into gold used to be an alchemist's ambition. Now a San Francisco fusion startup says it has done it.

Marathon Fusion announced on October 21 that it had converted mercury into stable gold, claiming a result ten times greater than a CERN benchmark. As Euronews and BeInCrypto report, however, the announcement did not establish the quantity produced, the production cost, or independent verification of the result. An intriguing claim, then, rather than proof that cheap gold is about to arrive.

For anyone who owns gold because it is difficult to produce, the distinction matters. Making a few atoms is one achievement. Manufacturing investment-grade bullion profitably is another.

The physics is real. The economics still need proving.

Creating gold through nuclear transmutation is possible. CERN has already documented the conversion of lead into gold nuclei at the Large Hadron Collider. During its 2015–2018 run, the four major experiments produced about 29 picograms of gold, which survived only briefly before fragmenting when it struck accelerator equipment. This was a physics result, not a new source of gold bars.

Marathon's proposed route is different. Its researchers describe using fast neutrons from deuterium-tritium fusion to convert mercury-198 into mercury-197, which then decays into stable gold-197. Their 2025 research paper models production of approximately two tonnes of gold per gigawatt of thermal fusion power each year. That is a simulation-based projection for a proposed plant configuration, not demonstrated annual output.

The commercial ambition is to produce gold alongside electricity and fusion fuel, giving a future power plant an additional revenue stream. Marathon presents that combination as a way to improve fusion economics. Whether an operating plant can deliver those benefits remains to be demonstrated.

Could gold follow lab-grown diamonds?

The comparison is tempting. Euronews points to the diamond industry, where manufactured stones have become much cheaper and have put pressure on the market for mined diamonds. But that history does not establish what would happen to gold.

Our view: any serious price argument needs evidence of repeatable production, competitive costs and enough saleable metal to matter. The announcement alone does not supply that evidence.

There is also a delay between producing gold and selling it. BeInCrypto reports that Marathon's paper discusses radioactive contaminants that could require years of storage, depending on the safety standard and the purity of the mercury feedstock. A successful reaction would not necessarily produce bullion ready for immediate delivery.

Calling this the end of gold would be premature. Dismissing the research simply because commercial production is difficult would be premature too.

A different kind of scarcity

For Bitcoin readers, the story invites a useful comparison.

Gold production can potentially change when technology changes. Bitcoin's issuance follows network rules: miners receive newly issued coins according to a declining schedule, with a supply limit of 21 million under the current consensus rules. More powerful equipment does not give a miner permission to award itself extra bitcoin. Nodes reject blocks that break those rules. Bitcoin.org's FAQ explains how issuance and validation work.

That does not guarantee Bitcoin's price or make it a risk-free investment. Scarcity alone cannot guarantee demand. It does explain why a better way to manufacture gold and a better way to mine Bitcoin have different implications for supply.

A laboratory could run Bitcoin mining equipment, of course. It would still have to play by the same rules as everyone else.

So, congratulations to the scientists working on lab-made gold. We're still waiting for someone in a white coat to manufacture bitcoin number 21,000,001.