What tokenizing an asset actually costs
Every page on this subject quotes a confident percentage saved. We went to the fee schedules instead. About 60% of the cost stack is public, exact and dated. The rest the market does not price openly — and we say which is which rather than filling the gap with a plausible number.
The three studies that measured this disagree
Every vendor page cites a savings percentage. None of them mentions that the central banks which actually ran the numbers reached different answers — and that the largest and newest found no issuance-cost saving at all.
Three measurements of the same question, plotted on one axis. They do not merely differ in size — they differ in sign.
- HKMA 2023 — significant only at the 10% level; the authors call the sample small
- BIS 2025 — found narrower bid-ask spreads, but no issuance-cost effect
- ECB 2026 — largest and newest; fees marginally HIGHER, not significant
Hong Kong Monetary Authority
202328 tokenised bonds, 23 issuers
Underwriting fees 0.22 percentage points lower (a 25.8% reduction) and yield spreads 0.78 points lower.
Both headline results are significant only at the 10% level. The authors write that the sample is small and that readers should interpret the results with caution.
Read the study → (opens in a new tab)Bank for International Settlements
202539 tokenised bonds
Bid-ask spreads about 37% narrower. Issuance costs do NOT differ systematically.
The BIS calls the gains modest and says the jury is still out on whether the yield effect reflects tokenisation or investor enthusiasm.
Read the study → (opens in a new tab)European Central Bank
2026183 tokenised bonds against ~200,000 conventional securities
Yield spread 0.14 points lower and bid-ask 0.05 points lower — but underwriting fees 0.04 points HIGHER, and not statistically significant.
The largest and most recent of the three. It finds no issuance-fee saving. The liquidity improvement also reverses for retail-accessible tokenised bonds — the opposite of the HKMA result.
Read the study → (opens in a new tab)The one thing all three share is an improvement in secondary-market liquidity, not in issuance cost. That is a real and interesting result. It is also not what the marketing material claims.
What we can state to the euro
Registry, regulator, exchange and central-securities-depository fees. Published, exact, dated. Every row links to the fee schedule it came from.
Note the shape of this: the entire statutory layer for a bond issuance lands in the low tens of thousands. Against a conventional issuance that is a rounding error — Adyen’s €849m IPO paid €25.2m to underwriters and €1.7m, or 0.2% of the offering, for listing, legal, administration, publication and tax combined. Tokenization touches the small component.
The handful of vendors that publish a price
Real numbers, but each reflects one provider's economics rather than a market rate.
| Item | Jurisdiction / vendor | Amount | Basis | Effective |
|---|---|---|---|---|
| White-label tier setup (opens in a new tab)Platform | DigiShares | €35,000 | one-off | 2026 |
| Platform subscription, by tier (opens in a new tab)Platform | DigiShares | €300–€1,500 | annual | 2026 |
| Pro tier and above, floor only (opens in a new tab)Platform | Fireblocks | $36,000+ | annual | 2026 |
| Self-service custody, 5bp per month on assets above $100,000 (opens in a new tab)Custody | BitGo | $60 | annual | 2026 |
| Per verification, basic to compliance plan (opens in a new tab)KYC | Sumsub | $1.35–$1.85 | one-off | 2026 |
| Protocol licence — open source (opens in a new tab)Smart contracts | ERC-3643 / T-REX | $0 | one-off | 2026 |
| Standard tiers, per month (opens in a new tab)Platform | Tokeny | €3,000–€5,000 | annual | 2026-08-03 |
| Annual management fee on the fund itself (opens in a new tab)Platform | Spiko | €0.25 | annual | 2026-08-03 |
What the market does not price publicly
We emit no number here. A calculator that returns a confident total for these is guessing, and the guess is the part you would be relying on.
Enterprise tokenization platform onboarding
no public priceEleven of fifteen platforms checked publish nothing: ADDX (fees explicitly at the provider's discretion), Taurus, Archax, Securitize, Superstate, Libeara, Centrifuge, tZERO, Dinari, Backed Finance and Zeconomy. Four do publish — Tokeny at €3,000 and €5,000 a month, Fireblocks from $999 a month, DigiShares, and Bitbond. Checked 2026-08-03.
Top-tier smart-contract audit
no public priceConfirmed by direct check that OpenZeppelin, Cantina and Quantstamp publish no rate card. The widely repeated figure of $25,000 per engineer-week traces to a single source nobody can locate. Only the contest platforms are genuinely public: Code4rena charges a 0% platform fee, Sherlock's pools run $20k–200k, Hats takes 10–20% of the payout.
Law-firm structuring fees
no public priceUniversally quoted privately in every jurisdiction checked. The only defensible public anchor is the European Commission's 2015 prospectus impact assessment — eleven years old and weighted to equity.
MiCA authorisation fees outside France
no public priceNo published fee schedule could be located for Luxembourg, Ireland or Malta. France is the only regulator publishing clean MiCA figures.
ERC-3643 deployment gas
no public priceNo published benchmark exists. The suite deploys six to eight contracts against ERC-20's one, so it is materially more expensive, but no sourced figure was found.
Four numbers you will meet, and what is wrong with them
These circulate widely enough that you will arrive having read at least one.
“$15–20bn annual infrastructure savings” and “35–65% post-trade cost reduction”, attributed to the World Economic Forum
The full 218-page report was retrieved and searched. Neither figure appears anywhere in it. The attribution appears to be an artifact of AI search summarisers repeating one another.
Quantstamp charges roughly $130,000 for a ten-week audit retainer
Contradicted by Quantstamp's own pages, which state that pricing is by custom quote only.
Roland Berger: €4.6bn in tokenised-equity trading-cost reduction by 2030
Appeared only inside a search-engine summary. A direct fetch of the Roland Berger page found no such figure, only qualitative language.
BCG and Ripple: 40–60% operating-cost reduction for investment-grade bonds
The figure is real and really published, but the report discloses no methodology, no baseline process, no sample and no formula. Every exhibit is captioned only “Source: Ripple and BCG”. Ripple has a direct commercial interest in adoption.
How we source and date all of this is on the calibration page. For the products themselves, see the Yield Board. None of this is investment, legal or tax advice — disclosure.