We run one of the platforms in this comparison, so read it with that in mind. What we can offer in exchange is that every number below was measured rather than repeated, including the ones that do not flatter us.
We run one of the platforms in this comparison, so read it with that in mind. What we can offer in exchange is that every number below was measured rather than repeated, including the ones that do not flatter us.
Two of those measurements reframe the entire category:
The most genuinely decentralised gold yield product we could find returns about 0.27% a year. We read its vault contract directly. That is what fully on-chain gold yield pays when nothing is subsidising it.
One widely cited "up to 11% APR" applies to a balance band of roughly $230 to $460. Above that, the rate drops to about 1%. It is a customer acquisition tier being quoted as a product rate.
Once you know both of those, the ranking question changes from "who pays most" to "what is generating the difference." Which is the only question worth asking here.
The spread in that first column is not a competitive ranking. It is a risk ranking.
0.27% is the control group. yoGOLD is a real ERC-4626 vault holding Tether Gold, open to anyone, with no off-chain credit arrangement manufacturing a return. We called its contract on 21 August 2026: its share price was 1.00232 XAUT after 315 days live. That is 0.27% annualised.
This matters because it is the honest answer to "what does gold yield on-chain, by itself."
Almost nothing. Gold produces no income, and a vault that does not introduce a payer cannot invent one.
0% is the reality in the big lending markets. Tokenized gold has been listed on the largest lending protocols for a long time and has paid essentially nothing throughout, because it was onboarded as collateral with borrowing switched off. Risk reviewers judged gold tokens most likely to be borrowed by short sellers.
A supply rate is paid by borrowers. No borrowers, no yield.
2 to 5% is what real economic demand pays. Jewellers, mints and refiners lease physical metal because it removes price risk from their inventory, and they pay rent in gold. That is a genuine payer with a genuine reason, and it is the benchmark every other row should be measured against. Its cost is access: accredited status, large minimums, and terms measured in years. We wrote that market up in full in Gold Lease Rates.
Anything materially above 5% is one of three things: a less conservative credit arrangement, an operator subsidy, or emissions of a token the operator issues. All three are legitimate designs. None of them is gold earning a return. Which one you are looking at is the single most useful thing to establish.
And watch the "paid in" column. Falcon Finance pays 3 to 5% but pays it in its own synthetic dollar, with a 180-day lock, so you are taking that issuer's credit on top of whatever backs the gold. A 4% yield paid in metal from a named lessee and a 4% yield paid in a platform's own liability are not points on the same scale.
The most cited high number in this category is a major exchange advertising "up to 11% APR" on tokenized gold. It is worth understanding what that means before it anchors your expectations.
The 11% applies to a balance band of roughly 0.05 to 0.1 XAUT. At recent gold prices that is about $230 to $460. Above that band the rate falls to roughly 1%.
It is not dishonest, exactly. It is disclosed in the terms. But it is a customer acquisition incentive on a few hundred dollars, quoted as though it were a product rate, and it gets repeated across comparison sites as "11% on gold" by writers who did not open the tier table.
The general lesson, which applies to us too: when you see a headline rate, find the balance band, the lock-up, the payment asset and the funding source before you believe it.
Any of the four can turn an attractive number into a different product.
We are not going to declare ourselves the winner of a table we are in.
Choose physical leasing if you qualify and can lock capital for years. It is the highest-integrity gold yield available: a real industrial payer, paid in metal, with decades of operating history. If you meet the accreditation threshold and do not need liquidity, this is the conservative answer and we would say so to your face.
Choose a bare vaulted token if you want gold exposure and no credit risk at all. PAX Gold and Tether Gold pay nothing, and that is the feature. Allocated bars, named custodians, published attestations, deep liquidity, and nobody who has to perform for you to get your metal back. If a yield is not worth a counterparty to you, this is the correct choice and it is not a compromise.
Choose an on-chain gold vault like yoGOLD if you want composability without added credit risk, and can accept that the return will be close to nothing.
Choose a custodial exchange product if convenience outweighs everything, and understand that you are an unsecured lender to that exchange.
Choose us if you want a gold-denominated yield at a small entry size, and you are willing to underwrite a named counterparty to get it. That is the actual trade.
Our entry point is a thousandth of an ounce rather than a full ounce, our yield is paid in gold rather than a stablecoin, and our accrual is verifiable on-chain rather than reported on a dashboard.
If those five are unacceptable to you, one of the other rows is your answer, and that is a reasonable place to land. The full risk picture is in Is Gold Staking Safe?.
Run these against any platform in this table, ours included. They take about twenty minutes and they will disqualify most of the category.
Who pays the yield, by name, and out of what business? If the answer is vague, stop. Gold produces no income, so somebody is paying, and a platform that will not identify them is telling you something. A named counterparty can be researched.
What are you paid in? Gold, a stablecoin, or the platform's own token. Being paid in a platform's own token stacks a second credit exposure on the first.
Is the rate fixed, target, or floating, and who can change it? A "fixed" rate on a variable source means somebody absorbs the difference, and you should know who and from what reserve. A target rate adjustable by an operator is a policy, not an obligation enforceable by you.
What is the exit, numerically? Notice period, minimum, and the maximum the operator could extend it to. Then check whether a secondary market exists. Verify the notice period in the contract rather than the marketing, which for our vault means calling withdrawalDelay().
What does the audit actually cover? By whom, over which contracts, on what date, and does the deployed code match the audited version. Then read the acknowledged findings, not just the resolved ones. And remember a code audit says nothing about whether reserves exist.
When was the last reserve attestation, and who performed it? Anything older than a year should be treated as unverified backing. An attestation is a point-in-time examination, not a full audit, and it does not tell you whether the metal is pledged elsewhere.
Can you actually get out at size? Check real market depth, not market capitalisation. In tokenized gold these diverge wildly: we found tokens with market caps in the hundreds of millions trading four figures a day. We measured the whole category in The Tokenized Gold Market in 2026.
There is no single answer, because the options differ on risk rather than quality. Physical leasing pays the most reliable gold-denominated yield and is largely closed to retail. Fully on-chain vaults are open and pay almost nothing. Credit-based platforms pay more and add a counterparty. Pick by which risk you are willing to hold.
Headline rates in this category reach double digits, but the highest published numbers frequently come with conditions such as small balance bands, long locks, or payment in a platform's own token. The highest rate genuinely available on an unrestricted basis is much lower than the highest advertised rate.
Supplying them to the largest DeFi lending markets currently earns essentially nothing, because they are listed as collateral with borrowing disabled. Some exchanges offer custodial earn products at low single digits. To use them on StakeMyGold you would convert to GGBR first, since PAXG and XAUT staking is on our roadmap rather than live.
It is achievable, and it is not gold earning it. Gold produces no income, so a double-digit rate is credit risk, a subsidy or a token emission. Whether it is sustainable depends entirely on which, and on the quality of whatever is behind it.
On documentary evidence, a bare vaulted token from a regulated issuer with monthly attestations carries the least counterparty risk, precisely because nothing is paying you. Safety and yield trade off directly here, and any platform claiming otherwise deserves scrutiny.
Physical leasing pays in metal. Kinesis pays in metal from platform fees. We pay in GGBR, which is gold-denominated. Several others pay in a stablecoin or their own token, which is a materially different product.
It varies enormously, from a few dollars for small-denomination tokens to accreditation thresholds and six-figure minimums for institutional leasing. Denomination is the main driver: a full-ounce token needs an ounce's worth of capital to buy one unit.
For an on-chain vault, read the contract. Compare the share price or exchange rate today against a past value and compute the realised change yourself. That is a measurement. Anything on a dashboard is a claim.
The useful ranking in this category is not by rate. It is by what is generating the rate.
0.27% is what gold yields on-chain when nothing is subsidising it. 0% is what the biggest lending markets pay. 2 to 5% is what real industrial demand pays, in metal, to people who qualify. Everything above that is credit, subsidy or emissions, and the platform owes you a straight answer about which.
We are in the credit column. We think the trade is worth it at a small entry size with a gold-denominated payout and verifiable accrual, and we have set out above exactly where we are the weaker choice. If those weaknesses matter more to you than the rate, one of the other rows is the right answer, and you should take it.
This article is for informational purposes only and is not financial advice.
name(), symbol(), asset() returning Tether Gold, totalAssets(), totalSupply(), and convertToAssets() returning 1.00232 XAUT per share, 315 days after deployment| Platform | Approx. rate | Paid in | Access | Structure | Verified how |
|---|
| Monetary Metals | 2 to 5% net | Gold | Largely accredited only | Physical leases to jewellers and refiners | Published investor yields |
| yoGOLD (YO Protocol) | about 0.27% | XAUT | Open, on-chain | ERC-4626 vault over Tether Gold | Read the contract directly, 21 Aug 2026 |
| Aave, Morpho, Compound, Fluid | 0% | n/a | Open, on-chain | Lending markets, gold listed collateral-only | Pool data, 347 days at 0% |
| Kinesis | about 0.5% | Gold | Open, own platform | Share of platform transaction fees | Published yield figures |
| Theo (thGOLD) | about 2% | See issuer | Not yet trading | Tokenised fund lending to jewellers | Issuer materials |
| Falcon Finance | 3 to 5% | USDf, a synthetic dollar | Open | 180-day lock, gold as collateral | Issuer materials |
| CEX earn products | 1% typical | Token | Open, custodial | Unsecured loan to the exchange | Published tiers |
| StakeMyGold (us) | Floating target, see staking page | GGBR, so gold | Open, on-chain | Credit arrangement against in-situ reserves | Contract read, 91-day accrual series |