How FinDaS designed Dravanti's tokenized commodity economy to survive three regulatory readings.
Dravanti arrived with the industry default in mind: a high-value, low-float token launch. What the business actually needed was an instrument architecture for physical metal reserves, one that Europe, Dubai, and the US would each classify differently.
TL;DR
- Client: Dravanti, a liquidity provider tokenizing high-purity industrial metal reserves, starting with 99.9998% purity copper powder
- Challenge: Select the right instrument (utility, security, or RWA token) when the same structure classifies as an ART in Europe, an asset-referenced virtual asset in Dubai, and a security in the US, and when the underlying metals are too heterogeneous for a single token to represent
- Approach: FinDaS evaluated multiple instrument directions across all three regulatory readings, designed two instrument classes (fixed-quantity spot commodity tokens and formula-issued basket security tokens), added a 1:1,000 ERC-20 wrapping layer for retail access, and valued the business through DCF and CAPM with listed commodity comparables
- Result: A complete three-phase token economy paper positioned closest to commodity treatment in the US; Dravanti has since moved its platform to a Nasdaq CSD equity-share structure
What problems did Dravanti have?
The brief arrived pointed at the wrong instrument
Dravanti came in with the industry default: a high-value, low-float token launch, where a small circulating supply meets a large fully diluted valuation and the early chart looks impressive. For a liquidity provider issuing claims on physical metal, that structure solves nothing. The token's value has to come from appraised reserves, not from float scarcity, and an engineered low float works directly against the one property an asset-backed token must have: a price that tracks its underlying. The first stretch of the engagement was spent widening the question from how to launch a token to which instrument the business actually needed.
The same design was a different asset class in every jurisdiction
Once the conversation moved to asset-backed instruments, the regulatory picture refused to settle. A portfolio of tokenized metals reads as an asset-referenced token (ART) under Europe's MiCA, as an asset-referenced virtual asset (ARVA) under Dubai's VARA rules, and as a security in the United States. Nothing about the token changes between those three readings; only the regulator does. Each classification carries different obligations for issuance, marketing, custody, and who may hold the token. FinDaS and Dravanti went through multiple rounds of design before settling on the structure that positioned the tokens closest to commodity treatment in the US.
Copper powder and nickel refuse to share an instrument
The underlying assets looked interchangeable on a slide and were nothing of the sort in practice. Copper powder of 99.9998% purity is a single, precisely specified asset: one grade, one form, one appraisal methodology. Nickel holdings, and future metal procurements, come in varying shapes, quantities, and qualities. A token representing a fixed amount of one exact grade cannot also represent a shifting basket of related materials. The team debated three structures: one instrument holding fixed amounts of each material, separate tokens per material, and portfolios composed of per-material tokens. Each option moved the regulatory classification and the issuance math at the same time, which is what made the debate long.
A NAV mechanism is only as good as its price feed
Every mechanism in the design ultimately references one number: the fair market value of the underlying. Redemptions pay out at NAV. Issuance mints against appraised value. Supply management triggers off the gap between market price and underlying value. For exchange-traded commodities that number is easy to source; for high-purity industrial metal powders it is not, and producing a defensible price meant combining independent appraisals with market data feeds. Getting this wrong would not degrade the token gradually. It would break every redemption and every issuance at once.
How did FinDaS approach the problem?
Protocol and Business Deep Dive
FinDaS started with the business, not the token. Dravanti operates as a liquidity provider: it procures metal, has it independently appraised, issues claims against it, and earns revenue from redemption fees and from selling tokens at or above the value of the underlying. The deep dive also mapped the regulatory geography, since Dravanti's structure touches Europe, Dubai, and US investor exposure simultaneously. Out of this phase came multiple candidate directions: a utility token, which Hristo advised against because the business had no genuine utility case to hang one on, security tokens, and RWA structures. Several rounds of back and forth followed before the mix was right. That iteration was the engagement, not a detour from it.
Token Utility and Value Capture Design
The instrument decision split the problem in two. For a single, precisely specified asset, a spot commodity token: 1 DSCCu token represents exactly 1 gram of 99.9998% purity copper powder, minted only against certified reserves and redeemable against the physical asset itself. For heterogeneous holdings, a security token: DSTCu is backed by a basket of varying copper forms and qualities, with issuance driven by a valuation formula that preserves the underlying value each existing token represents whenever new metal enters the basket. Two instruments, two issuance models, one boundary rule: nothing gets minted without an appraisal.
Economic Modeling
The economics were modeled in Google Sheets: issuance schedules, redemption flows, treasury releases, and revenue projections across a five-year horizon. Where assumptions needed external grounding, FinDaS used observed market behavior; redemption rates, for example, were calibrated to the measured burn rates of live tokenized commodities such as Paxos Gold, using rwa.xyz data. The models are deterministic and only as good as their inputs, which is why every load-bearing assumption is named in the paper rather than buried in a cell.
Stress Testing and Valuation
Valuation followed the FinDaS framework: value the operating business, not the token float. A discounted cash flow model priced Dravanti's projected redemption-fee and token-sale revenue, with a CAPM discount rate derived from listed commodity-trading comparables (Archer-Daniels-Midland, Bunge, Glencore) rather than crypto tokens, because Dravanti's risk profile is a commodities business with a token wrapper, not the reverse. Revenue then ran through a beta-distribution sensitivity model spanning 50% to 200% of plan, producing a valuation range instead of a single number. The full valuation methodology and its outputs live in the internal-only sections of the paper, by design.
Documentation and Launch Readiness
The deliverable was a full token economy paper, structured by audience: a public executive summary suitable for the whitepaper, detailed monetary and fiscal mechanics for developers and technical reviewers, and internal-only valuation, sensitivity, and benchmark sections. It covers all three phases of the platform: spot commodity tokens, basket-backed security tokens, and a later service layer of wrapped portfolios, lending integrations, and a collateralized stablecoin. The paper closes with implementation notes and an explicit recommendation to bring token economy monitoring in house after launch.
A portfolio of tokenized metals is an ART in Europe, an ARVA in Dubai, and a security in the United States. The design had to survive all three readings at once.
What did FinDaS design?
1. Two instrument classes instead of one
The core architectural decision. Spot commodity tokens (the DSC series) carry a fixed-quantity claim on one exact asset grade and support physical redemption. Security tokens (the DST series) carry a proportional claim on a managed basket, with formula-driven issuance: before any new mint, the value each existing token represents is recalculated from fresh appraisals, and new tokens are issued only in the quantity the newly procured metal supports at that value. Basket rebalancing is bound by the same rule, preserving the fair market value each token represented when the rebalancing began.
Outcome: Each asset gets the instrument its physical properties can support, and no issuance event can dilute existing holders below appraised value.
2. A wrapping layer that trades reach for compliance, deliberately
ERC-3643 enforces who may hold the token at the protocol level, which satisfies the accredited-investor requirement and caps the audience. The wrapping contract converts ERC-3643 tokens into standard ERC-20 tokens at a 1:1,000 ratio, opening DeFi collateral, lending, and swap venues to retail holders. Wrapped tokens cannot be redeemed with the issuer; only an accredited investor can unwrap them. That asymmetry is the point: KYC-ed arbitrageurs are the single class of actor able to cross the boundary in both directions, and their profit motive is what keeps the retail market priced against the accredited one.
Outcome: Retail exposure without retail custody of the regulated instrument, and a price link between the two markets that no one has to actively operate.
3. A NAV corridor instead of a peg
Holders redeem at NAV, minus a 10% fee covering warehousing, shipping, and liquidation costs, paid in cash or against the physical metal within 28 days of notice, processed in FIFO queues. On the other side, Dravanti manages supply: repurchasing tokens when they trade at a discount to the underlying and releasing tokens at market prices when they trade at a premium. Neither mechanism fixes the price. Together they bound how far it can drift.
Outcome: A two-sided corridor around underlying value, enforced by redemption arbitrage from below and issuer supply from above, with no peg to defend.
4. Bridging under a hard supply invariant
The wrapping contract extends to cross-chain bridging: tokens are burned on the origin chain and reissued on the destination chain, in both directions, for spot and wrapped tokens alike. The invariant: total supply across all chains and standards never exceeds current circulation. 100,000 tokens on one chain and a 10,000-token transfer means 90,000 and 10,000 afterward, verifiably.
Outcome: Multi-chain reach without the classic bridge failure mode of duplicated supply.
5. A valuation the client could defend in a diligence meeting
Asset-backed projects get challenged on their numbers more than their mechanisms. The DCF and CAPM work gave Dravanti a valuation with named comparables, a stated discount-rate derivation, and an explicit sensitivity range. That is the difference between a number an investor can interrogate and a marketing multiple.
Outcome: Every figure in the paper traces to a stated assumption, and the assumptions most likely to move the valuation are the ones stress-tested explicitly.
What were the results?
The engagement delivered what it was scoped to deliver: a complete token economy whitepaper covering all three platform phases, with the instrument architecture, issuance and redemption mechanics, supply management rules, valuation, and sensitivity analysis behind them. The design positioned Dravanti's tokens closest to commodity treatment in the US, which was the deliberate outcome of the regulatory mapping that shaped instrument selection from the first weeks.
Dravanti has since taken its platform in a different structural direction, registering its offering as Nasdaq CSD equity shares. The instruments as designed in the paper have not launched, so this case study claims no market results: no price, no volume, no holder counts. What the engagement demonstrates is the part of asset tokenization that happens before any of those numbers exist: the instrument selection, regulatory positioning, and issuance math that determine whether the numbers can ever be good.
Key takeaways.
For asset-backed tokens, instrument selection is the tokenomics. Whether the token is a utility, a security, or an RWA claim determines more of the outcome than any emission schedule or vesting curve.
Regulatory classification is relative, not absolute
The same structure can be an ART in Europe, an ARVA in Dubai, and a security in the US at the same time. Design for the strictest plausible reading and the other jurisdictions come along; design for the friendliest and the strictest one will eventually find you.
Heterogeneous assets need heterogeneous instruments
A fixed-quantity token works for one exact grade of one material. The moment the backing varies in form or quality, issuance must be formula-driven against fresh appraisals, or existing holders get silently diluted.
Compliance and liquidity are bridgeable when someone profits from the bridge
A restricted ERC-3643 instrument and an open ERC-20 wrapper only stay priced together because KYC-ed arbitrageurs earn a spread for crossing the boundary. Design the actor, not just the contract.
The price feed is part of the token design
NAV redemption, appraisal-gated issuance, and supply management all reference the value of the underlying. For assets without an exchange-traded price, sourcing that number is a first-order design problem, not an operational detail.
What's next.
The paper's closing recommendation to Dravanti is the same one FinDaS gives every asset-backed client: a token economy is a living system, and once real issuance and redemption data exists, someone in house should own the monitoring and recalibration. The instrument-selection framework built during this engagement, evaluating utility, security, and RWA structures against jurisdiction-by-jurisdiction classification, now runs at the start of every FinDaS asset tokenization engagement.
Bring your project to a real token economist.
Free, no obligation, no juniors. We'll answer every question about your token design and tell you candidly where the weak spots are.
The questions we keep getting.
How do you tokenize a physical commodity like copper?
Start with the asset, not the token: the metal must be procured, independently appraised, and held before a single token is minted. Dravanti's design issues tokens only against certified reserves, with each spot commodity token representing a fixed quantity of one exact grade, and redemption at NAV anchoring the token to the physical asset.
What is the difference between a spot commodity token and an asset-backed security token?
A spot commodity token represents a fixed quantity of one specific asset grade, such as 1 gram of 99.9998% purity copper powder, and can be redeemed against the physical asset. A security token is backed by a basket of varying forms and qualities of the asset, uses formula-driven issuance to preserve per-token value, and settles in cash rather than metal.
How can retail investors get exposure to ERC-3643 security tokens?
Directly they cannot: ERC-3643 restricts holding to verified accredited investors at the protocol level. Dravanti's design solves this with a wrapping contract that converts ERC-3643 tokens into freely transferable ERC-20 tokens at a 1:1,000 ratio, with KYC-ed arbitrageurs wrapping and unwrapping to keep the two markets priced together.
What keeps an asset-backed token trading close to its net asset value?
Two mechanisms working from opposite sides. Redemption at NAV puts a floor under deep discounts, because holders can exit at underlying value, while issuer supply management closes premiums by releasing new tokens at market prices. Neither is a hard peg; together they form a corridor.
Is a tokenized commodity a security, an ART or an ARVA?
It depends on the jurisdiction reading it, and the same structure can be all three at once: an ART under Europe's MiCA, an asset-referenced virtual asset (ARVA) in Dubai, and a security in the United States. The practical approach is to pick the strictest plausible classification and design the instrument to survive it.
How do you value an asset-backed token project?
FinDaS values the operating business, not the token float: discounted cash flow on redemption fees and token sales, with a CAPM discount rate built from listed commodity-trading comparables. Revenue assumptions then run through a sensitivity model spanning 50% to 200% of plan, so the valuation is a range with named assumptions rather than a single number.