Security-layer growth doesn't reach the governance token

Short answer: between 2018 and 2026, the dual model (a security RWA token plus a separate utility or governance token) helped the issuer and the holders of the security token, but not the holders of the second token.

We examined more than 20 projects across four archetypes: treasury RWA, real estate, STO platforms from 2018-2022, and private credit. By July 2026, every utility/governance token in the sample was trading 51-99.9% below its ATH by price. For comparison, over the same period BTC fell roughly 50% from its peak, and ETH roughly 62%. Meanwhile, the security layer at the best projects grew several-fold: the AUM of the Ondo platform rose from $125m in early 2024 to $3.43b in May 2026, while Maple's AUM grew from $516m to $4.59b over the course of 2025 alone.

The reason for the divergence is structural. Almost no governance token had a legal right to the cash flow of the security product: Centrifuge's holder revenue per DefiLlama equals $0, and as of May 2026 ONDO had neither a fee switch nor a buyback. The model is genuinely useful, first, to the issuer: it provides regulatory isolation and capital raising. Second, it is useful to the holder of the security token, who has an enforceable claim. The governance token holder gets crypto beta, unlocks, and a narrative. The only partial counterexample, Maple (SYRUP), points to the same conclusion. The token held up better than its peers only while an explicit buyback from revenue was operating, and that came after the merger into a single token.

Prices consolidated into a single table show the same outcome across different assets

The prices below are from CoinGecko, data pulled on October 15, 2026, unless otherwise noted. "July 2026" refers to the CoinGecko daily value on July 1 or July 15, 2026 (the date is specified). The change from ATH was calculated by us from the prices shown; it is not a quote from the source. The "Link" column describes whether the utility/governance token derives value from the security layer.

Price sources: CoinGecko ONDO, CoinGecko EDEN, CoinGecko REG, CoinGecko RIO, CoinGecko BST, CoinGecko EL, CoinGecko PROPC, CoinGecko TNGBL, CoinDataFlow OM, CoinGecko MANTRA, CoinGecko POLY, CoinGecko POLYX, Coinpaprika SWM, CoinGecko SMT, CoinGecko STBU, CoinMarketCap NEU, LiveCoinWatch NEU, CoinGecko CFG, CoinGecko SYRUP, CoinGecko GFI, CoinGecko TRU, CoinGecko CPOOL, CoinGecko HASH, Digrin TZROP, MEXC INX, SEC 6-K INX.

The sample excludes projects without a token pair: Propy (utility token without securities), Kasa and Lofty (securities without a platform token), Superstate, Midas and Hashnote/USYC (security tokens without a governance token). Plume (PLUME, ATH $0.2475 on 03.19.2025, $0.00962 on 07.01.2026, -96.1%) is mentioned only as context: it is an RWA L1, not a classic pair (CoinGecko PLUME).

Treasury RWA: the Regulated Product Grows, the Issuer's Token Lives Separately

Ondo: AUM Up 27x, ONDO Down 85% in Price

Ondo is built as a pure dual model. OUSG represents a stake in Ondo I LP, is sold under Reg D Rule 506(c) to qualified purchasers only, and transfers are possible solely between allowlisted addresses. USDY is structured as a note issued by Ondo USDY LLC for non-US persons with a 40-day transfer restriction (eco.com, OUSG, eco.com, USDY). Product fees accrue to Ondo Finance Inc., USDY LLC and the Ondo Foundation. ONDO confers governance only: there is no staking, revenue share, buyback or burn, and "revenue to token holders is zero" (MoneyCheck, 02.05.2026).

The security layer grew without pause. Ondo's tokenized assets stood at $125m at the start of 2024 (21Shares), TVL surpassed $1b in March 2025 (Markets Media) and $2b in January 2026 (The Defiant). As of 12 May 2026, platform AUM was $3.43b: USDY $2.10b, OUSG $285m, Ondo Global Markets $1.04b (Allium). Around August 2026, DefiLlama shows on-chain AUM of $4.119b (DefiLlama RWA Ondo). ONDO's price moved in the opposite direction over this period: an ATH of $2.14 on 15 December 2024, $0.81 in April 2025, $0.3740 on 31 December 2025, a low of $0.2428 on 13 April 2026 and a range of $0.3076-0.4156 in July 2026 (CoinGecko market_chart). Allium puts the ratio of ONDO's market cap to the platform's circulating AUM at 0.41x. That is 3-4 times more expensive than peers, while the token "does not receive protocol cash flow." Allium sees the main catalyst in a DAO vote on a fee switch in H2 2026 with a binary outcome (Allium). Whether it took place was not found in the sources.

ONDO's decline is explained by token supply, not by the product. On 18 January 2025, 1.94 billion ONDO unlocked, 134% of the then-circulating supply (Cryptoast). Circulating supply rose from 3.16 billion (April 2025) to 4.87 billion (October 2026) (CoinGecko). Between 1 January and 1 February 2026, ONDO's market cap grew from $1.14b to $1.70b, while the price fell from $0.3595 to $0.3492. This points to yet another January unlock (our conclusion based on CoinGecko data). Regulatory positives barely moved the price. The SEC closed its investigation, which examined whether Ondo's products and ONDO itself were securities, without charges. On the day of the news ONDO rose 3.7% to around $0.47 (Coin360), and by 31 December 2025 it was worth $0.3740.

OpenEden: Regulated TBILL and USDO Failed to Hold Up EDEN After Listing

OpenEden replicates Ondo's structure. TBILL is issued as a regulated BVI professional fund with BNY acting as manager and custodian, USDO is issued by Bermuda-licensed OpenEden Digital, and EDEN is issued by a separate BVI entity, OE Services Limited (OpenEden, 14.08.2025). EDEN's link to the products is limited to staking into xEDEN, discounts, and the fact that part of protocol revenue "may" go toward buying back EDEN. The source of staking rewards is not disclosed (The DeFi Saint). By August 2025, TBILL TVL exceeded $286m and USDO TVL exceeded $277m. EDEN listed on Binance and KuCoin on 30 September 2025 and set an ATH of $1.31 the same day (CoinGecko). Bitrue attributes the first-day drop to selling by airdrop recipients and an unclear unlock schedule (Bitrue). 80% of the airdrop vested over 120 days after TGE (The DeFi Saint). EDEN's price fell from $0.0694 in mid-January to an ATL of $0.0261 on 30 March 2026, and in July 2026 it fluctuated in the $0.039-0.046 range (CoinGecko market_chart). We found no data on TBILL and USDO AUM after August 2025, so growth of OpenEden's security layer in 2026 is unconfirmed.

An important detail for comparison: the largest treasury issuers operate without a governance token at all. Superstate, Midas and Hashnote/USYC have no such token (DIA, The Next Web), and USYC was acquired by Circle (The Defiant). It turns out that in the treasury segment a second token is needed neither to scale the product nor to exit.

Real Estate: the Worst Outcomes Come from the Asset Layer, Not from Tokenomics

RealT: $140m in security tokens versus REG market cap of $470k

RealT's timeline is a good illustration that a governance token offers no protection against a failure at the asset level. RealTokens were issued as interests in a Delaware Series LLC, and later as shares in a corporation. Under the 2019 memorandum, distributions are made at the discretion of the Managing Member, and token holders rank behind creditors in a liquidation (CrowdfundedWealth). REG launched on April 29, 2024 as the governance token of RealToken DAO, with no share of fees and no burn (CoinDesk REG). In July 2025, Detroit filed suit over 408 properties (Michigan Public). In February 2026, weekly rent distributions were suspended (Michigan Public, 20.03.2026), and on July 2, 2026 a voluntary liquidation was announced, with roughly $140m raised in total (Mogul). About $640,000 remained in escrow for the wind-down, roughly $45 per investor (Crypto Briefing). REG's price fell from an ATH of $1.99 (December 1, 2024) to $0.0739 on June 30, 2026 and $0.0168 on July 10 — a 77% drop over the ten days around the liquidation announcement (CoinGecko).

Tangible and MANTRA: reflexive collateral and market structure

At Tangible, a link between the tokens did exist, but it worked against the system. USDR was backed by 190 UK properties and DAI. When roughly $12m of DAI was withdrawn on October 11, 2023, USDR fell to $0.51 (Blockworks). Excluding TNGBL, the stablecoin was undercollateralized; including TNGBL, the collateralization ratio stood at 102% (The Block). A project's own token used as collateral creates reflexive risk: under stress, both the collateral and the debt decline. TNGBL later migrated into the RWA token of the re.al network with veRWA, which receives 100% of network fees (Everything Blockchain). The TNGBL listing on CoinGecko is outdated, while USDR was quoted at $0.4356 in October 2026 (CoinGecko USDR).

MANTRA surged on the narrative of a $1b deal with DAMAC (Semafor). The legal regime of these assets was not found in the sources. On April 13, 2025, OM fell from roughly $6.14 to $0.52, and market cap lost more than $5b. The causes relate to exchange liquidations and tokenomics changes, not to the assets (Cointelegraph). In March 2026, OM was redenominated into MANTRA at a 1:4 ratio (KuCoin).

Small platforms: peaks on the March 2024 narrative and a 98-99% loss

At Realio, Blocksquare, ELYSIA and Propchain, the utility token carries no rights to rental cash flow or NAV. In Realio's case, RST represents company equity under Reg D/Reg S, while RIO is explicitly described as "not an investment in the Realio company" (MEXC). By May 2024, Blocksquare had tokenized $100m across 118 properties, but BST has no fee or burn mechanism (The Defiant). The ATHs of PROPC (March 27, 2024) and BST (March 28, 2024) are one day apart. Both peaks came during the wave that followed the launch of BlackRock's fund, which was succeeded by a 14-52% correction in RWA tokens over the 30 days to April 28, 2024 (FXStreet).

STO-Era Platforms: A Mandatory Fee Token Failed to Create Demand, Buyers Acquired Licenses

Polymath: 153 Issuances in Two Years Against a Supply of Roughly 1b POLY

Polymath tested the most direct linkage possible: without POLY, you could not issue a security token. A ticker reservation cost 250 POLY, and creating the token another 250 POLY (Ivey). The platform raised $57.8m in its January 2018 token distribution. By December 2019, 153 tokens had been deployed, and only 12 of them had five or more holders, while the price of POLY had fallen more than 95% to an ATL of $0.0186 (Coin Bureau). Even as the number of issuances grew (more than 120 in May 2019, 153 by December), the price declined. Token demand at 500 POLY per issuance is negligible next to a supply of roughly 1b. Polymath moved to its own Substrate chain, Polymesh. The company explained the move by saying Ethereum "was not built with regulation in mind," and that updating a whitelist cost up to $100 a day (CoinDesk). The POLY-to-POLYX bridge closed on February 29, 2024, by which point more than 89% of supply had migrated (Polymesh). POLYX set an ATH of $0.7488 on March 31, 2024, the same week as SMT, and by July 2026 it was worth $0.0362 (CoinGecko POLYX).

Fee Sharing and Discounts Didn't Help Either: Neufund, Swarm, Stobox

Neufund's NEU conferred the right to revenue from ETO fees (Crowdfund Insider). But over its entire lifetime the platform raised roughly €20m and shut down in January 2022: BaFin granted neither approval nor prohibition (CoinDesk, Crowdfund Insider). Swarm abandoned SWM (delisted from Bitfinex in 2021) in favor of a new token, SMT, offering discounts and rewards (Swarm docs), and in December 2025 the company was acquired by Inveniam (DL News). Stobox prices its services in USD, starting at $12,000 for setup, and does not require payment in STBU (Stobox). The company is still operating, while STBU has lost 99.8% from its ATH.

The Control Group Without a Utility Token Preserved More Value

Companies with a single security layer and no utility token fared better. Harbor was acquired by BitGo (The Block), Securrency by DTCC (Ledger Insights), and Apex took a majority stake in Tokeny (Unlock). Buyers paid for licenses and technology, not for tokens. After the Republic takeover, INX Token holders received a share in a cash fund of roughly $34.3m (SEC 6-K). Not a single utility token holder in the sample received anything comparable. TZROP, from an STO that raised $134m (Overstock/Beyond), was still quoted at around $1.98 OTC in October 2026 (Digrin). Liquidity, however, was minimal: daily volume was below $15,000 in 2019 (The Block).

Private credit: defaults zero out tokens slowly, while buybacks buy time

Centrifuge and Figure: TVL and IPO grow, CFG and HASH fall

Centrifuge's TVL grew from $50m to more than $1b over 2025 (Centrifuge Q4 2025) and to $1.7b by the end of Q1 2026, while JTRSY alone exceeded $1b (Centrifuge Q1 2026). Over the same period, the new CFG fell from its ATH of $0.4004 (August 24, 2025) to $0.0672 (February 6, 2026), a drop of 83% (CoinGecko). According to DefiLlama, Centrifuge's fees for 2025 amounted to $25.89m, protocol revenue $2.22m, and holder revenue $0 (DefiLlama). The project itself states that "the fee switch is on," but DefiLlama Token Rights shows the fee switch as off, with no buyback and no dividends (DefiLlama Token Rights). The discrepancy is resolved as follows: revenue goes to the foundation via CP162, not to CFG holders (Centrifuge forum). The rebound to $0.2676 on June 1, 2026 coincided with listings on Binance, Upbit, and Bithumb, not with fee growth.

Figure provides the most striking divergence. YLDS is registered with the SEC as a face-amount certificate (Markets Media). Figure's IPO on September 10, 2025 gave the company a market cap of $5.29b (IPOScoop), while active AUM in YLDS and FGRS stood at $568.67m (DefiLlama RWA Figure). HASH set an ATH of $0.06015 three to four days after the IPO and by July 1, 2026 had lost 85% (CoinGecko HASH). The value went to FIGR shareholders. No link between HASH and YLDS was found in the sources.

Maple: a merger into a single token and buybacks delivered the sector's only gain

Maple merged MPL into SYRUP at a 1:100 ratio in November 2024 (Kraken) and, starting in January 2025, directed 20% of protocol revenue to a monthly SYRUP buyback (MIP-013) (Maple forum). SYRUP rose from an ATL of $0.0852 on April 6, 2025 to an ATH of $0.6532 on June 25, 2025, a 7.7x increase. Over the same period, DefiLlama TVL grew from $267m (end of 2024) to $1.65b (mid-2025) (DefiLlama Maple). It is the only token in CoinGecko's top-7 RWA list that gained from January 1, 2025 to March 31, 2026 (+28.6%) (CoinGecko 2026 RWA Report). Then the link weakened. Maple's AUM reached $4.59b by the end of 2025 (+767% year over year) (Maple 2025 Data Review) and held at around $4.55b in August 2026 (Maple Memo), but by July 1, 2026 SYRUP had declined to $0.1369 (-79% from its ATH). Holder revenue for 2025 was $2.88m against a market cap of $485.1m as of November 1, 2025, roughly 0.6% per year (our calculation). Such a buyback can, at best, provide a weak floor. The new MIP-021 (July 2026) tied the buyback to a 10/20/30% scale of net monthly revenue. The first buyback in August was executed at an average price of $0.1604 (MIP-021), and by October 1 SYRUP was trading at $0.2257.

Goldfinch, TrueFi, Clearpool: a credit event kills TVL, and the token follows

The direct effect of a default on price is noticeable but limited: MPL fell 24.7% within 24 hours after Orthogonal's $36m default in December 2022 (The Block). The main damage comes from the slow loss of TVL and trust. Goldfinch wrote down $5m on Tugende and roughly $7m of $20m on Stratos (The Defiant, The Defiant). In June 2026, the project approved a wind-down (GIP-87) (CoinPaprika), and GFI lost 62% in a month, from $0.1281 to $0.0492 (CoinGecko GFI). The first losses were borne by Backers, not GFI holders. The token suffered not through direct participation in the losses, but through capital flight. TrueFi's TVL fell from $813.7m (August 2021) to $21.7k (DefiLlama TrueFi). Clearpool resumed its CPOOL buyback on October 20, 2025 (Clearpool), but in August 2026 the token set an ATL of $0.01597. An announced value accrual without a disclosed size failed to hold the price.

Five price drivers, and AUM growth is not the main one among them

Crypto beta sets the backdrop but does not explain the scale of the decline

Utility-token peaks cluster around BTC/ETH cycles and narratives. The January-February 2018 peaks (POLY, SWM, NEU) came before the start of the window. Next come the 2020-2021 cycle (RIO, EL, STBU, TRU, CPOOL, GFI, the old CFG), March 2024 (BST, PROPC, SMT, POLYX), December 2024 (ONDO, REG) and 2025 (OM, PLUME, SYRUP, CFG, HASH, EDEN). By July 2026 BTC and ETH were far from their peaks, but governance tokens had fallen much harder. ONDO lost 85% from December 2024, while BTC over the same period lost about 45% (from $108,388.88 to $59,961.45, our calculation).

Source: daily candles from Coinbase Exchange, data pulled October 15, 2026 (Coinbase BTC-USD, Coinbase ETH-USD).

Unlocks, regulation, credit events and the shift in value accrual

The second driver is unlocks. ONDO on January 18, 2025 (+134% to circulating supply) and the vesting of the EDEN airdrop weighed on the price more heavily than AUM growth supported it. The third driver, regulation, works in both directions. The Kik (2020), Telegram (2020) and LBRY (2022) cases showed that the "utility" label does not protect a token sold with an expectation of profit (The Block, SEC, MoFo). The SEC's 2019 framework explicitly named buybacks and burns by the issuer as a hallmark of an investment contract (SEC). That is why RWA projects kept governance tokens "clean" for years, and value leakage was built into the design. The regime shift in 2025-2026 (the GENIUS Act of July 18, 2025, Project Crypto, the joint SEC-CFTC taxonomy of March 17, 2026) reduced this risk (Jenner & Block), but coincided with a market downturn. The fourth driver consists of credit events and asset failures: Orthogonal, Goldfinch, RealT, the USDR depeg. The fifth driver relates to changes in value accrual. The effect of an announcement is positive but short-lived: UNI rose as much as +44% within 24 hours after the fee switch proposal and by July 31, 2026 was worth $4.345 (Coin360, Coinbase UNI-USD). AAVE added 7.8% on the buyback proposal (The Block). A cosmetic redenomination without new economics does not work: by October 2024 only about 10.7% of MKR holders had migrated to SKY (Blockworks).

A Divergence in the Numbers: Tokenized Asset Market +256.7%, Six of Seven RWA Tokens in the Red

A CoinGecko report documents a divergence at the sector level. From January 1, 2025 through March 31, 2026, the market cap of tokenized RWAs grew from $5.42b to $19.32b (+256.7%), yet among the top-7 RWA tokens (Ondo, Maple, Centrifuge, Mantra, Clearpool, Chintai, Goldfinch) only SYRUP appreciated (+28.6%). The remaining six fell by 44.7-98.8%, with ONDO down 80.6% from $1.35 to $0.26. The report's conclusion: "the value created on-chain does not necessarily accrue to token holders" (CoinGecko 2026 RWA Report). Within this market, the AUM of tokenized treasuries per rwa.xyz grew from $3.9b at the start of 2025 to more than $15b by the end of April 2026 (Cointelegraph).

Sources: Allium, 21Shares, Maple, The Defiant, Centrifuge Q1 2026, IPOScoop, DefiLlama RWA Figure, OpenEden, CoinGecko 2026 RWA Report.

Theory explains the mechanism as well. Malinova and Park show that a "generic" token without carefully designed cash-flow rights is inferior to equity, while a properly designed token can replicate equity payouts (Management Science). In the dual model, the cash-flow right is deliberately assigned to the security token or the regulated fund, which means the governance token remains "generic" by construction.

The dual-token model works only when the link is explicit and quantified — here are the conditions for founders

The model solves the issuer's problems: it separates out the regulated product, opens up exchange liquidity for the second token, and provides early capital (Polymath $57.8m, Neufund $11.6m). It also protects the security-token holder: they hold an enforceable claim, as demonstrated by the INX payout and the dollar price of TZROP. For the governance-token holder, the model does not work unless the link is defined by contract and measured in dollars. The legal constraint has not gone anywhere. Under the SEC-CFTC taxonomy of March 17, 2026, a "digital commodity" must not confer rights to future income or profits (Jenner & Block), while in the EU a hybrid token carrying dividend rights qualifies as a financial instrument (BDO Malta). By granting the governance token a right to revenue, a project moves it closer to security status and recreates the very problem that the token split was meant to solve.

Conclusion

The key takeaway for tokenomics: the dual-token model contains no mechanism that transfers value from the security layer to the governance token. It has to be designed separately, with a legal basis and worked-out economics. In eight years, not a single project in the sample has done this in a way that let the token outperform BTC. Maple came closest, and only for as long as the buyback was visible against the pressure of supply. For a founder, the question "do we need a second token" reduces to the question "what dollar of revenue, and under what legal right, will reach its holder." If there is no answer, the market will price the token as crypto beta with unlocks, which is exactly what the 85-99% drawdowns show. At a minimum, a utility token must have justified demand — in other words, it should not redistribute income from the project's assets or revenues, but carry a concrete function within the ecosystem.

The second takeaway concerns risk. The worst outcomes (RealT, Goldfinch, Tangible) came from the real-asset layer and the operating partner, not from the token design. An audit of an RWA project should begin with the legal and credit structure of the security layer, and only then move on to the emission schedule of the second token. The shift in the U.S. regulatory regime in 2025-2026 has, for the first time, made adding value accrual to a governance token legally less risky. The next cycle will show whether an explicit fee switch — for example, Ondo's expected vote — can reverse the divergence that has been the rule until now.