The FCA's Tokenized Gold Question and What London's 70% Bullion Share Actually Buys

The FCA has begun consulting on how tokenized representations of physical gold should be treated under UK financial promotion and custody rules — and it is doing so at a moment when the London bullion market still clears roughly 70% of global over-the-counter gold flow, a share the LBMA has cited in its public trading data submissions for years. The conventional read is straightforward. London writes the rules because London holds the vaults, the clearing, and the price discovery. Sitting on an Asia desk that watches Shanghai, Hong Kong and Singapore open before London wakes, we find that conventional read half-right and half-obsolete.

The Conventional View: London's 70% Share Makes It the Natural Venue for Tokenized Gold

Here is the story you will hear repeated at every conference panel from Zurich to Dubai this quarter. Tokenized gold is a technology wrapper around an old asset. The wrapper needs a rulebook. The rulebook should live where the asset lives. The asset lives in a small number of London vaults operated by clearing members of the LBMA, moved by armored trucks between roughly a dozen addresses in Zone 1, and settled through a system — loco London — that the rest of the world quotes against.

The 70% figure is doing a lot of work in that argument. It is invoked as if it were a market share in the way Amazon has a market share of e-commerce. Whoever writes the rule for the biggest venue writes the rule for the market. And so the FCA, sitting on top of that venue, gets to define what "tokenized gold" means for the purposes of financial promotion, custody, redemption rights, and the treatment of the underlying claim. Everyone else — MAS in Singapore, HKMA in Hong Kong, the JFSA in Tokyo — will follow, harmonize, or find themselves quietly out of scope for the largest counterparties.

You can see why this framing has legs. If you are an issuer of a gold-backed token and your target customer is a UK-regulated wealth manager, the FCA's rulebook is not one input among many. It is the input. The consultation matters. The comment letters matter. The final wording matters more than any technical decision about which chain the token sits on, because the chain is easy to change and the regulator is not.

Why This Is Actually True

Let me concede this properly, because if you skip the concession you miss why the conventional view survived for as long as it has. London is not just where the biggest pile of allocated gold sits. It is where the price gets made in a specific, mechanical sense. Loco London settlement means that when a Singaporean private bank buys a tonne of gold from a South African refiner, the entry that changes is a ledger entry at a London clearing member. The metal may never move. The price is quoted, cleared, and settled against a system that is physically and institutionally in the City.

This has a knock-on effect for tokenization. A gold token whose redemption right does not resolve to loco London delivery is, functionally, a different asset. It might be a better asset for some purposes — faster settlement, retail-scale denominations, transparent on-chain proofs of reserve. But it does not sit in the same fungibility class as the thing the LBMA price fix is fixing. And regulated buyers care about that fungibility class, because it determines whether a position can be netted, hedged, and reported against existing frameworks.

The FCA understands this. Its consultation is not, in the reading we have done, an attempt to relitigate the fundamentals of the bullion market. It is an attempt to draw a bright line between tokens that plug into loco London — same custody chain, same redemption mechanics, same auditability — and tokens that use "gold-backed" as marketing while the underlying arrangement is something else entirely. If you have ever tried to explain the difference between an allocated account and an unallocated account to a client who has been reading crypto Twitter, you know why this distinction is worth writing down.

So the conventional view is right about the direction of gravity. Rules written in London do bind the largest institutional flows. The 70% share is not fabricated. And the FCA is doing serious work on a real problem.

But here is what that framing misses entirely: the 70% is a settlement share, not a demand share, and the demand share has been moving east for fifteen years without asking anyone's permission.

Where It Breaks Down: The Asia Session Already Runs Its Own Gold Market

Sit at a desk in Singapore at 08:30 local time. Shanghai has been trading for an hour. Hong Kong is open. Tokyo is open. London is asleep. This is the window where a very large fraction of the world's physical gold demand — jewelry fabrication, central bank accumulation, retail bar-and-coin — is actually priced against domestic markers, not loco London.

The Shanghai Gold Exchange runs contracts that settle in yuan against physical metal held in mainland vaults. That is not a token wrapper on London. It is a parallel settlement system with its own vaulting, its own clearing, and its own price. Since the PBOC pushed the yuan through its 2015 devaluation and the managed-float evolution that followed through 2022, Chinese domestic gold demand has been quietly building an alternative price discovery layer that does not rely on the LBMA fix for its economic decisions.

Hong Kong plays a different role. Under the linked exchange rate that has been in place since 1983, HKD gold pricing has always tracked London closely — because the HKMA's mandate forces it to. But the vaulting infrastructure that has grown up around Hong Kong International Airport since the early 2010s is now a physical settlement location in its own right, used by mainland buyers who want a bonded delivery point outside PRC jurisdiction. That inventory is real. It is not on the LBMA's clearing rails.

Singapore is the most under-appreciated piece of this. Since MAS put its wholesale market framework in place in 2008 and the government pushed to make Singapore a precious metals hub — GST exemption for investment-grade bars, freeport vaulting near Changi — the city-state has grown into a serious secondary vaulting jurisdiction. Not London's scale. Not trying to be. But large enough that a tokenized gold product designed for Asian retail wealth channels can plausibly custody in Singapore and never touch a London vault.

*The MAS licensing regime distinguishes between digital payment tokens and capital markets products. Tokenized gold sits in an unresolved seam between the two. Ask three lawyers in Singapore, get three answers.*

Now overlay Japan and Korea. The JFSA's 2005 FX law framework, and Korea's 2009 retail forex restrictions, both created legal traditions in which "regulated" and "tokenized" are not natural neighbors. The regulators there will not defer to the FCA on tokenized gold rules — they have their own precedents about how retail exposure to foreign-denominated assets should be structured, and those precedents pre-date the crypto conversation entirely. A UK-approved gold token does not automatically clear a JFSA suitability review.

So the FCA is writing rules for the 70% share of clearing that already exists. The demand growth is happening in the 30%. If you are building a tokenized gold product today for anyone other than a UK institutional client, London's rulebook is a constraint, not a home.

The Rule I Use Instead: Follow the Vaulting, Not the Clearing

Listen — I want to give you the actual mental model I use when a client asks me where a tokenized gold product "belongs" from a regulatory perspective. It is not "which regulator is biggest." It is "where does the metal physically sit, and what are the redemption mechanics from that location."

Here is the math, worked slowly, because this is where most analysts skip a step. Take a hypothetical tokenized gold product that issues 100,000 tokens, each backed by one gram of physical metal. That is 100 kilograms of gold. At a spot price around USD 65 per gram — pick your own number, the arithmetic works the same — that is USD 6.5 million of underlying inventory. Now the vaulting fee. Institutional allocated storage in London runs roughly 12 to 18 basis points per year on stored value; in Singapore's freeport it has historically run tighter, call it 10 to 15 basis points; in Zurich, 15 to 25. So on that USD 6.5 million, London vaulting costs somewhere between USD 7,800 and USD 11,700 annually. Singapore, USD 6,500 to USD 9,750. Zurich, USD 9,750 to USD 16,250.

Those look small. They are not small when you divide them across the token float. USD 10,000 of annual vaulting on 100,000 tokens is 10 cents per token per year — call it 15 basis points against a USD 65 unit. If your management fee is 40 basis points and 15 of those are eaten by storage, you have 25 to cover custody insurance, audit, chain fees, and issuer margin. Move the vaulting to Singapore and you free up 3 to 5 basis points. Move it to a jurisdiction the FCA does not recognize as equivalent for custody purposes and you have just excluded yourself from selling into UK wealth channels — which may or may not matter depending on your target book.

*A single institutional-grade audit cycle on 100kg of allocated inventory is a two-day physical inspection. The auditors count the bars. They read the serial numbers. They cross-check against the assay certificates. There is no on-chain shortcut for this step.*

That is the rule. Follow the vaulting, because the vaulting determines the audit trail, the insurance policy, the redemption mechanics, and — this is the part the London-centric view misses — the regulator whose written opinion actually matters to your buyer. The FCA writes rules for London vaulting. The MAS writes rules for Singapore vaulting. The SGE writes rules for mainland vaulting. A token whose reserves sit in Singapore is a MAS conversation first, whatever the marketing site says about "regulated by."

When the Old Rule Still Wins

I want to give the conventional view its due one more time, because there is a specific book of business where London's centrality is not just historical inertia — it is the correct answer.

If your target buyer is a pension fund, a sovereign wealth desk, or a large UK or European private bank allocating institutional-scale positions, the fungibility with loco London matters more than any storage cost saving. Those buyers hedge in the London forwards market. They mark against the LBMA fix. They need a redemption right that resolves into the same clearing system their custodian is already plugged into. For that book, a tokenized gold product custodied outside the London vaulting network is a taxonomy problem — the ops team will not know how to book it, the risk team will not know how to net it, and the compliance team will not know which regulator to escalate to. The friction wins.

The FCA's consultation is writing the rulebook for exactly that buyer. It will be a good rulebook, in the sense that it will be internally coherent and enforceable. It will not be the global rulebook, because the demand growth is not in that buyer's book anymore. Both statements can be true at the same time, and if you can hold both without picking a side, you will make better product decisions than the analysts who cannot.