Lab-made gold: will gold share the fate of diamonds?

For centuries, alchemists searched for a way to turn base metal into gold. In October 2026, the San Francisco fusion start-up Marathon Fusion announced that it had done just that: mercury had been converted into stable gold, ten times faster than in the previous record set at CERN. Which raises a question: could lab-made gold crash the gold price the way it happened with diamonds? According to diamond industry analyst Paul Zimnisky, wholesale prices of lab-grown diamonds have fallen by 90–95% since 2015.

The answer is shorter than it might seem. The technology rests on real physics, but between a laboratory demonstration and the gold market stand several formidable barriers: feedstock, time, radioactivity and scale.

What exactly was announced?

Marathon Fusion was founded in 2023 by Kyle Schiller and former SpaceX engineer Adam Rutkowski. The company develops fuel systems for future fusion power plants. Its backers include Breakthrough Energy, the fund founded by Bill Gates, and Anglo American, owner of the diamond giant De Beers.

The company claims to have beaten CERN’s record. At the world’s largest particle accelerator, only a minuscule amount of gold was produced from lead, roughly 29 trillionths of a gram, and even that existed for just a fraction of a second.

How does mercury turn into gold?

Gold and mercury are neighbours on the periodic table, which is why one can be turned into the other. In a fusion reactor, mercury is bombarded with neutrons, and within a few days part of it turns into stable gold. It is exactly the same gold that is extracted from mines. Scientists call this process nuclear transmutation; for simplicity, this article refers to the result as lab-made gold.

However, not just any mercury will do, only one particular variety of it, or isotope. Natural mercury contains only about 10% of this isotope. According to the company’s calculations, a single power plant would need around 600 tonnes of it, which means several thousand tonnes of ordinary mercury would have to be processed. Physicists describe this kind of separation as technically complex, and the company itself acknowledges that a large-scale solution has yet to be developed. On top of that, the mining and trade of mercury are strictly restricted by the international Minamata Convention.

How much gold could be produced, and at what cost?

According to the company’s estimates, a single large fusion power plant could produce around 5 tonnes of gold a year. A hundred such plants would yield about 500 tonnes a year, roughly 14% of global mine production (3,672 tonnes in 2025, according to the World Gold Council).

The question of cost remains open. Marathon Fusion’s study models revenue, not costs. The company estimates that gold could raise a large plant’s annual revenue from ~USD 223 million to ~USD 482 million. In other words, gold would be a by-product of a fusion power plant that improves the economics of an energy project. It would not be a gold factory capable of operating on its own.

Note: both fusion gold cost figures were calculated by investor Michael Green. They are not based on data published by the company. Marathon Fusion has not disclosed its production costs.

Lab-made gold would need to be stored for almost 18 years

Irradiation also produces a small amount of radioactive gold, so the resulting metal cannot be sold straight away. According to the company’s study, it no longer requires radioactivity labelling after about 14 years, and after 18 years it is less radioactive than an ordinary banana.

This means that gold from a plant that started operating in, say, 2035 could reach the market only around 2050. Throughout that time, it would have to be stored, insured and financed.

So where is the risk for the gold industry?

It would be unfair to claim there is no risk at all. In the long run, several scenarios are worth watching.

A price “ceiling”. If the technology ever operated at costs well below those of gold mines, it could theoretically cap the rise in the gold price. That, however, would require substantial volumes.

The impact of supply on price. According to analysts’ calculations based on a World Gold Council model, every additional ~25 tonnes of annual supply lowers the price by about 1%. A hypothetical 500 tonnes a year would therefore mean roughly 20% downward pressure on the price. This is a modelled scenario, not a forecast.

Gold mining shares are more sensitive than gold itself. Mining companies’ profits depend on the gap between the gold price and their production costs, so their shares react more strongly to news like this. On 22 October, the day after the announcement, the gold miners ETF GDX lost 5.8%, while the gold-backed GLD fell 2.6%. Analysts, however, attribute the decline mainly to a broad stock market sell-off rather than to the fusion news.

Synthetic gold will be impossible to tell apart. Specialists can identify lab-grown diamonds. Once its radioactivity has faded, lab-made gold will be completely identical to mined gold. That is why no separate, cheaper market segment will emerge for it, as happened with diamonds. This gold would simply add to the overall supply.

Why there is no cause for concern yet

Limited feedstock. Karl Lackner, a physicist at the Max Planck Institute for Plasma Physics, estimates that even all the mercury available in the Earth’s crust could yield no more than ~600 tonnes of gold a year. In the near term, stockpiled mercury waste could yield around 20 tonnes a year.

Enormous production capacity would be needed. To produce as much gold as the world’s mines extract each year, more than a thousand large fusion power plants would be required. Marathon Fusion’s own experimental reactor is roughly a million times smaller than a single such plant.

There are no commercial fusion power plants yet. No fusion plant currently sells electricity. The first commercial plants are scheduled to start up between 2028 and the early 2030s, and Lackner believes that first-generation reactors will most likely not produce gold.

The price is set by existing gold stocks. Throughout history, humanity has mined around 222,600 tonnes of gold. Even 500 tonnes a year is only ~0.22% of that amount. The gold price is determined primarily by how much of the existing stock its owners wish to hold or sell, not by a given year’s new production.

Gold is not a diamond

For a long time, diamond prices were propped up by limited supply and powerful marketing. Once lab-grown diamonds could be produced cheaply and in large quantities, prices fell. Wholesale prices of lab-grown diamonds have dropped by 90–95% since 2015, and this has also affected the prices of natural diamonds. In 2018, De Beers launched its own lab-grown diamond brand, Lightbox (USD 800 per carat), but closed it in 2025 to focus on natural diamonds. That same year, the company’s operating loss reached USD 511 million. According to Bloomberg, Anglo American is discussing the sale of its 85% stake for around USD 1 billion. It is therefore telling that Anglo American is also an investor in Marathon Fusion.

Gold plays a different role. It is a monetary asset with a liquid global market, held by central banks in their reserves and still being actively bought by them. In recent years, the central banks of China and Poland have been among the buyers. The value of gold does not depend solely on how difficult it is to extract. It also rests on trust built up over thousands of years.

What does this mean for investors?

In the coming years, the gold price will continue to be driven by the same factors as before: real interest rates, central bank purchases, geopolitics and investor demand. Since its record high in January (USD 5,603 per ounce), gold has lost more than 27%. Analysts attribute this to rising real interest rates (the yield on 10-year US inflation-protected Treasury securities rose from 1.89% to 2.64% in 2026), not to the fusion news.

Lab-made gold is a long-term factor worth keeping an eye on, not an immediate threat. If the technology ever reaches commercial scale, it will happen gradually and be visible years in advance, because it would first require built power plants, specially prepared feedstock and decades of storage. Those who might feel the effects more keenly are high-cost gold mining companies, not holders of physical gold.

This story illustrates well why mining costs only partly determine the gold price. Diamonds lost their value when their scarcity turned out to be artificial. The value of gold rests on trust accumulated over thousands of years, and advances in physics will significantly affect it only if gold becomes truly abundant.

Sources:
Euronews,
The Oregon Group,
Lead-Lag Report,
The Wolf Den,
JCK,
CERN ALICE.