Gold Can Be Made. Bitcoin Can't.

On Wednesday 21 October a San Francisco startup called Marathon Fusion posted a sentence alchemists spent two thousand years trying to write: "Last month, we demonstrated high-rate alchemy: turning mercury into real, stable gold, beating CERN's record by a factor of 10."

It landed in a bad week for gold. The metal set a record of about $5,600 an ounce on 29 January and slipped to around $4,000 this week, more than a quarter off its high. Put the two together and my timeline did what timelines do. Half of it decided gold was finished. The other half decided the whole thing was a hoax.

Neither is right, and for Muslims the question deserves more care than a hot take. Gold isn't just another chart for us. It was the dinar of the early Muslims, it's one of the six things the Prophet ﷺ named when he taught us what riba is, and it's the yardstick our zakat is measured against. When someone says they can make it in a machine, it's worth knowing exactly what that means.

So here's my read. This won't change gold's supply for decades, if it ever does. Gold falls for ordinary reasons and always has, and it's falling for ordinary reasons now. But one thing did change this week, and it matters to anyone holding gold or Bitcoin as savings. Gold's scarcity was always a cost, and costs come down. Bitcoin's scarcity is a rule.

What Marathon Fusion actually claimed

The method is real physics. A fusion reactor burning deuterium and tritium throws off neutrons at 14 MeV, fast enough to knock a neutron out of an atom of mercury-198 and leave mercury-197. Mercury-197 is unstable, and over the next few days (its half-life is about 64 hours) it decays into gold-197, the only stable form of gold there is. Atom for atom, it's the same gold as the gold in a wedding ring.

The company first laid this out in July 2025, in a paper called Scalable Chrysopoeia via (n, 2n) Reactions Driven by Deuterium-Tritium Fusion Neutrons. Chrysopoeia is the old Greek word for gold-making. That paper was a simulation and still hasn't been peer reviewed. This week's post claims the first experimental result.

Here's what they published: the claim, and gamma-ray spectra. Here's what they didn't: how much gold they made, how pure it is, what it cost, or any independent check. The physics is plausible. The numbers aren't public. Until they are, it's a claim.

And the record they say they beat is tiny. In May 2025 CERN's ALICE experiment reported that lead collisions in the Large Hadron Collider had produced about 86 billion gold nuclei between 2015 and 2018. That's 29 picograms, trillionths of a gram, and each nucleus lasted a fraction of a second before smashing into the machine. Ten times almost nothing is still almost nothing.

Two more details are worth knowing. The reactor Marathon Fusion is building in San Francisco targets 1 kilowatt of fusion power, while its gold projections assume a gigawatt, a million times more. And its $5.9 million seed round in July 2024 was co-led by Anglo American, one of the largest mining companies in the world. The miners are watching too.

None of this should surprise anyone who knows the history. Physicists at Harvard got radioactive gold out of mercury with neutrons in 1941. Glenn Seaborg's team at Berkeley turned bismuth into a few thousand atoms of gold in 1980. What's new is a company proposing to do it at a rate that matters, as a by-product of a power plant.

Why it won't flood the market

Start with how much gold there is. These are the World Gold Council's figures for 2025, next to the fusion projection:

Behind all of that sits the stock: about 219,891 tonnes of gold mined in all of history and still above ground at the end of 2025, as jewellery, bars, coins and central-bank reserves. Gold almost never gets used up. A whole year of mining adds less than 2% to the pile.

Now the fusion side. To match one year of mining at 2 to 5 tonnes a plant, you'd need somewhere between 700 and 1,800 gigawatt-scale fusion plants. That's more plants than there are nuclear reactors of any kind running on Earth today, and the number of commercial fusion plants running today is zero. The most optimistic timelines put the first ones in the 2030s.

Then come the catches:

• The gold comes out radioactive. Other isotopes ride along with it, and coverage of the paper puts the wait at 14 to 18 years in storage before the metal is safe to sell. Gold made in 2040 reaches a jeweller in the mid-2050s.
• The feedstock is scarce too. Mercury-198 is only about a tenth of natural mercury, so it has to be enriched first. The same post announced a record in isotope enrichment, which is the step this whole idea depends on.
• The physics is shown, the economics aren't. No yield, no purity, no cost. A process can be real and still never be worth running.

So even in the best case for Marathon Fusion, fusion gold arrives decades from now, a few tonnes at a time, into a market of 220,000 tonnes. That's a footnote to supply, not a flood.

Gold falls anyway, for ordinary reasons

If anyone tells you gold is down because of a startup in San Francisco, look at the calendar. Gold was falling before Wednesday, and it has fallen hard many times without any help from physics.

• 1980 to 1999. Gold peaked at about $850 in January 1980 and spent nearly twenty years sliding to around $250, down about 70%, as Paul Volcker pushed real interest rates up and inflation came down.
• 2011 to 2015. From about $1,920 to about $1,050, down roughly 45%, as rates rose, the dollar strengthened and fear of the financial crisis faded.
• 2026. A record of about $5,600 on 29 January, then a fall of as much as 9% the next day, when President Trump named Kevin Warsh as his choice to lead the Federal Reserve. This month the reasons in the reports are higher oil prices feeding inflation fears, traders expecting rates to stay high for longer, and rising Treasury yields.

What moves gold is real interest rates, the dollar, central-bank buying and the mood of investors. A few extra tonnes of supply barely registers next to any of those. Gold being down more than a quarter isn't a sign it's broken. It's a sign it has a price.

And to be straight with you, Bitcoin hasn't been a calm place to hide this year either. It peaked at about $126,000 in October 2025, closed at $58,526 on 30 June, and was trading around $81,500 on 23 October. Anyone telling you to swap your gold for Bitcoin because Bitcoin is the steadier asset is selling you something.

The word that changed is "can't"

So what did this week change? Not the supply. The argument.

Every case for gold as money rests on one fact: nobody can make more of it at will. That was never quite a law of nature. It was a fact about cost. Gold is rare in the earth's crust and expensive to dig out, and that expense is what kept it scarce. Seaborg made gold in 1980 at a cost nobody would ever pay. CERN made it as a side effect of a particle collider. Now a startup with a mining company among its backers says it can make it as a side effect of a power plant. Each step is cheaper than the last. Nobody knows whether the line ever reaches the point where it pays, but "impossible" has become "expensive", and expensive is a number that moves.

Gold's supply already answers to its price. When gold rises, miners dig poorer ore that wasn't worth digging before, and people sell their old jewellery: recycling rose 3% last year. A higher price brings out more gold. That's how a commodity behaves, and there's nothing wrong with it. It just isn't fixed scarcity.

Bitcoin works differently, in a way you can check.

Its supply is set by its rules, not by what it costs to produce. If more miners join, or machines get faster, or energy gets cheaper, the network makes mining harder every 2,016 blocks, roughly every two weeks, so new blocks keep arriving about every ten minutes. The reward for each block halves every 210,000 blocks, and since April 2024 it has been 3.125 bitcoin. The total stops at 21 million, and more than 20 million already exist. A better machine wins a miner a bigger share of the same reward. It never makes more bitcoin.

You don't have to take anyone's word for that either. Every Bitcoin node checks that no block created more than the schedule allows, and refuses any block that did.

I'll admit the weak spot, because there is one. Bitcoin's limit is a rule that people enforce by running software, and in principle people could change it. In practice that change would need nearly everyone who holds bitcoin to agree to dilute their own savings, and every node that refused would carry on under the old rules. Gold's limit is geology plus cost, and one clever company can lower a cost. Bitcoin's limit is a rule plus everyone who checks it, and that can only change if nearly all of them agree.

Will the fear push money into Bitcoin?

What follows is my read, not a forecast.

Headlines like this one don't move supply. They move feeling. Plenty of people hold gold because nobody can make more of it, and this week they heard that someone can, at least in a lab. Most will shrug, and for now they're right to. Some will start asking a sharper question: which store of value can't be inflated by anyone, whether that's a central bank, a miner or a reactor? That question is the best case Bitcoin has, and every story like this one asks it for free.

If that doubt turns into demand for Bitcoin, I'd expect it to happen slowly, over years, not in a single candle. I wouldn't put one trade on it. Bitcoin's price this year has answered to the same rates and the same dollar that pushed gold down, and an X post won't change that by Monday.

What it means if you're a Muslim who holds gold

Your gold's ruling hasn't changed. It's still wealth in the most classical sense there is. Owning it is permissible, and exchanging it for money still has to be hand to hand, because gold is one of the six things the Prophet ﷺ named: "Gold for gold, silver for silver, wheat for wheat, barley for barley, dates for dates and salt for salt, like for like, equal for equal, hand to hand" (Sahih Muslim 1587). Here's how that applies to buying gold today.

Gold made in a reactor would still be gold. If it ever reaches the market, it will be gold-197, atom for atom the same as gold from a mine. The rulings on riba and zakat attach to gold itself, so I'd expect them to apply to it in full. I haven't seen a scholar rule on it, and I'm not a scholar. It's a question to put to them well before the first reactor-made bar is for sale, and there are years to do it.

Your nisab is measured in grams, not dollars. The gold nisab is 20 mithqal, about 85 to 87.5 grams. If gold ever did become cheap, the weight wouldn't move but its value in your currency would fall, so more people would cross the line and owe zakat, not fewer. Most contemporary scholars already prefer the lower silver nisab for cash, and our zakat calculator shows you both.

If you want to act on a fall, don't short it. A gold CFD never delivers any gold at all. A gold future puts off both the gold and the payment, which is exactly what the hand-to-hand condition forbids, and nearly all of them are closed before any delivery. Both usually come with leverage. A swap-free account doesn't fix that. If you're bearish, there are honest ways to be bearish without shorting, and the simplest one is holding less.

If you hold gold digitally, hold the real thing. PAX Gold and Tether Gold both pass our screen, because each token is a share of a specific, allocated bar. A custodian is still a counterparty, though, which a coin in your own hand isn't.

What I'm doing

Nothing, and I mean that as a position.

I hold Bitcoin as savings because its supply is a rule I can check. I'm not buying more because of an X post, and if gold is part of your savings, an X post isn't a reason to sell it either. My plan is the same one I use every week: spot only, money I own, entries decided before the price gets there, and a look at the Supply and Demand Index before I act on anything. That's my plan for my money, not a signal for yours.

If gold's scarcity is the reason you own it, it's worth knowing that argument picked up a footnote this week. Bitcoin's didn't.

People have been trying to make gold for two thousand years, and this week someone may have got a little closer. Nobody is going to make the 21,000,001st bitcoin, not with a reactor and not with a printing press. That isn't a price call. It's the difference between a cost and a rule.