A US fusion startup claims it has achieved something alchemists chased for centuries: producing gold from mercury. Marathon Fusion’s research paper describes how high-energy neutrons from a deuterium-tritium fusion reaction can irradiate mercury-198, producing a radioactive isotope that decays into stable gold-197. Crypto Twitter quickly branded the discovery gold’s “infinite mint bug,” and the question being asked this week is whether a metal revered for millennia for its scarcity can survive contact with modern nuclear technology, and whether Bitcoin, the asset marketed as digital gold with a hard cap of 21 million coins, inherits the upside.
The Claim and the Reality Check
The science is legitimate. Nuclear transmutation of mercury into gold has been demonstrated in particle accelerators, and Marathon says it produced stable gold at a rate 10 times the record set by CERN, where the ALICE experiment generated an estimated 29 picograms of gold nuclei during its 2015-2018 run. The catch is scale. Research engineer Jake Harter says the quantities produced so far remain microscopic, with the company hoping to reach macroscopic milligram quantities in 2027.
The larger numbers in Marathon’s paper are projections for a future fusion power plant, modeling roughly two tonnes of gold per gigawatt of thermal fusion power annually. For context, the World Gold Council reports that mines produced 3,672 tonnes of gold in 2025. A fusion plant that does not exist yet, built at a cost nobody has quantified, would add a rounding error to global supply.
The Lab-Grown Diamond Comparison Everyone Is Making
Coinbase chief executive Brian Armstrong argues the story plays out in Bitcoin’s favor, comparing net-new gold created through alchemy to the lab-grown diamond industry, where wholesale prices have collapsed by more than 90 percent since 2018 as production scaled. His conclusion, posted publicly: you cannot invent anything that will increase Bitcoin supply. There will only ever be 21 million.
Analysts are more skeptical about the gold parallel. Nagham Hassan, a market analyst at eToro, says the comparison is understandable but the economics are completely different, because diamond production became cheaper as factories scaled while fusion gold would require infrastructure that does not yet exist commercially. Nikos Kavalis of Metals Focus is blunter: we are nowhere near a stage where we can talk about production significance, and there is no demonstration the process is commercially viable, only theoretical revenue projections with nothing about reactor costs.
There is also an awkward detail the headlines skip: the fusion-produced gold would be radioactive and would need to be stored for a bit over a decade before dropping to banana-levels of activity. Marathon’s Harter does not see that as a major barrier since most gold is just stored as an investment anyway, which is either a clever observation or a commentary on the asset, depending on your view.
Why Bitcoin Holders Are Paying Attention
The deeper argument is about monetary properties. Fabian Schär, professor of distributed ledger technology at the University of Basel, confirms the basic economics: for any asset, a substantial increase in available supply would put downward pressure on its market price, and under a hypothetical assumption of inexpensive gold production at scale, gold would be no exception.
Bitcoin faces no such tail risk. Its supply schedule is enforced by consensus rules every node validates, and changing it would require convincing an ecosystem famous for rejecting even modest upgrades. That rigidity frustrates developers sometimes, but it means no laboratory, no government and no corporation can manufacture more of it. In a week when the market is digesting the largest exchange hack in history, the contrast is a useful reminder of why the hard cap narrative exists at all.
The Sober Bottom Line
With the first commercial fusion reactors estimated to be at least a decade away, and economists of every camp agreeing the economics remain unproven, nothing happening in a California lab is moving the gold price or the Bitcoin price this month. Gold bugs and their most famous advocate can breathe easy, and Bitcoin investors should treat the infinite mint bug framing as an interesting thought experiment rather than a catalyst.
What the episode does highlight is durable. Gold’s scarcity is a physical accident that technology keeps nibbling at, from deep-sea mining to asteroid ambitions to transmutation. Bitcoin’s scarcity is a social contract enforced by thousands of independent nodes. For a portion of institutional allocators weighing the two as inflation hedges, that distinction quietly accumulates value every time gold’s supply assumptions make headlines, even when the technology in question is a decade away and microscopic in scale.
two tonnes per gigawatt against 3,672 tonnes mined in 2025. the fusion plant doesn’t even exist yet and people are pricing gold to zero lmao
hard cap gang will spin anything but he’s right on this one. you cant transmute sats in a particle accelerator
you can transmute anything with enough neutrons, btc just lives on clocks and hashes instead of atoms. thats the whole pitch
exactly, the 21m cap is enforced by nodes not chemistry. golds cap was always vibes, this just proves it
Armstrong’s lab grown diamond comparison is doing a lot of work here. diamonds collapsed because factories got cheap, nobody has built a commercial fusion reactor to make gold
lab diamonds were chemistry getting cheap. turning Hg-198 into gold needs an actual working fusion plant, we are decades and billions out from that scale
golds infinite mint bug and the math is still 3,672 tonnes mined in 2025 vs picograms from a reactor that doesnt exist. scarcity discourse runs on vibes
3,672 tonnes a year from mines against picograms from a reactor that does not exist, and gold still traded scared. the headline did all the work
right, CERN ran 4 years for 29 picograms and Marathon is hyping milligrams by 2027. bitcoiners can relax for a few decades
milligrams by 2027 is a funding deck slide, not a plant. i will worry about gold when someone actually builds a gigawatt scale reactor
CERN ran four years for 29 picograms. even at 10x that rate you power a lab for a century before denting 3,672 tonnes of yearly mine supply