Ondo Stocks: first by balance, halved by share, and the cost of the RFQ route

TL;DR

Ondo is still the largest issuer by on-chain value ($984M as of 2026-10-16), and it is losing the lead. Its share fell from 58% in February 2026 to roughly a third — not because it shrank, but because the market doubled while it stood still.

The real problem is not the size table but turnover. On the same Circle exposure, CRCLON turns over under 1% in a day; the bStocks version does 27% and the xStocks version 29%. Three tokens tracking one company, about thirty times apart in activity.

My read: Ondo's $984M looks like a custody balance rather than live money. It leads on compliance narrative and instrument coverage, and neither substitutes for distribution.

The legal structure: the token and the share are different assets

The issuer is Ondo Global Markets, registered in the British Virgin Islands, renamed Ondo Stocks in July 2026.

Official material describes the tokens as fully backed by the corresponding share or ETF. The Broadridge partnership announcement draws the boundary more clearly:

• The Ondo token and the share or ETF held as collateral are different assets
• Token holders do not directly hold shareholder rights in the underlying
• Some governance works by holders expressing views to the issuer, which is not the same as voting the share

That framing is considerably more honest than "1:1 backed", and closer to the truth. "Fully backed" describes a collateral relationship, not an ownership one.

Ondo's secondary-market terms spell out several conditions worth reading one by one:

• The issuer gives no guarantee that any secondary market continues to exist
• No guarantee that any particular token has sufficient liquidity
• Secondary market prices need not equal the issue price
• Redemption requires customer due diligence and KYC
• The issuer may refuse redemption where it suspects a transaction fails legal or internal control requirements
• Holder recourse may be limited mainly to the relevant collateral and what the offering documents provide

The second-to-last point is substantive. "May refuse redemption" means redemption is a service, not a right. Grading has to separate that from "has a redemption channel" — it is what pins Ondo at B rather than A.

The share narrative: the company says 70%, the data says a third

This is the claim to be most careful with.

Ondo states publicly that it holds over 70% share among tokenized equity issuers. On CoinGecko's basis the real figure is roughly a third.

The timeline:

Over the same period the market went from roughly $1B to $3.5B (rwa.xyz basis), while Ondo moved from the $1B TVL it reported in June to $984M now — the market more than tripled and it barely moved.

That 70% should be treated as a marketing figure and kept out of any research conclusion. Numbers like it usually rest on a narrow definition of "issuer" — counting third-party wrappers only, excluding exchange-issued tokens. When you meet one, go back to a neutral source.

Balance is not turnover

This is the most informative comparison in the file. All three issuers run a Circle-exposure token:

One underlying company, three versions, roughly thirty times apart.

NVDA's on-chain activity has the same shape over 30 days:

A single bStocks NVDA token sees 29 times Ondo's transfer count.

Solana DEX volume is more extreme still over 30 days: NVDA, xStocks $37.7M against Ondo $0.04M; TSLA, xStocks $30.8M against Ondo $0.002M.

What the RFQ route costs

None of this means Ondo lacks liquidity. Its flow runs through RFQ and mint/redeem channels, quoted by market makers when a user initiates, rather than sitting in DEX pools.

The design is defensible: RFQ often executes large orders better than a shallow AMM pool, and avoids impermanent loss and MEV. Jupiter's screener shows RFQ for some Ondo tokens instead of a fixed dollar liquidity figure.

But the cost is real:

1. DeFi composability is close to zero — a token not in a pool cannot be an AMM pair, and integrating it into lending markets is harder
2. Off-hours quote quality is uncertain — RFQ depends on a market maker choosing to quote; spreads and price impact can widen on weekends and while the underlying is shut
3. Ordinary AMM metrics do not apply — you need quote response time, fill rate and slippage at increasing order size, and none of that is public

And "can sit in a pool, can serve as collateral" is the only differentiation tokenized equities have over an ordinary brokerage. Coinbase's tokens total just $7.5M, yet they already trade on Uniswap and Aerodrome, work as Aave collateral and can be borrowed against on Morpho. On that axis $7.5M has more optionality than $984M.

The only issuer seriously returning voting rights

Credit where it is due: Ondo is the only large issuer working the voting problem properly.

Its 2026 custodial version will support proxy voting through Broadridge — the incumbent proxy-voting infrastructure in traditional securities, not a crypto-native improvisation.

It is also expanding geographically, working with SBI to bring Japanese equities on-chain.

If proxy voting actually ships, Ondo's legal-structure grade needs redoing. Robinhood has published a plan for 1:1 redemption plus voting; Ondo is taking the Broadridge proxy route. Whichever lands first may reset the pass mark for the whole category.

A slice of the book

2026-09-19, a Saturday, with AAPL last closed at $335.19:

AAPLON / Gate: roughly $94.7K of 24h volume, a 5.65 bps spread, about $16.2K of bid depth against $12.0K of ask depth, and +0.21% versus the close.

This is the most balanced of the four AAPL books — bid to ask at about 1.35:1, against 6.4:1 for AAPLB on Binance and 0.10:1 for AAPLX on Bybit.

Balance is a virtue, but the scale is small. Adequate for small limit orders; not for size without live RFQ.

A size figure that is easy to misread

Ondo's protocol TVL is around $3.56B, but the overwhelming majority sits in Treasury products like USDY and OUSG, not stock tokens. The equities line is currently a strategic loss leader bought with share.

By the same logic, comparing Securitize's $3.88B TVL to Ondo on "stock TVL" is a mismatch — that figure is mostly funds such as BUIDL. Lock the methodology before comparing, or you end up counting a Treasury fund as an equities moat.

The ONDO token itself: $0.368, around $1.79B market cap, roughly $103M of 24h volume. The market is pricing an RWA platform, not a category leader's premium.

Risks

What I will watch

1. Whether absolute TVL recovers above $1B — below $800M confirms the lead has changed hands
2. When Broadridge proxy voting is genuinely usable — the only lever that re-rates its legal structure
3. Whether CRCLON turnover escapes <1% — if it stays there, "leader" means a custody ranking
4. Whether it starts seeding DEX liquidity — abandoning the pure RFQ route would be a clear strategic turn
5. Whether more venues follow Bitget in delisting Ondo tokens — losing distribution is more dangerous than losing size

Bottom line. Ondo leads on regulatory path, instrument coverage and the voting-rights question, and those are real long-term assets. But a fall from 58% to a third says one thing plainly: asset coverage and a compliance narrative do not substitute for distribution. The balance is there; the money is not moving through it.

For now I grade it B: issuer and underlying are clear, but redemption is gated by KYC and issuer discretion, and secondary liquidity is the weakest among the major wrappers. If Broadridge proxy voting ships, that grade needs rebuilding. Until then, first by size is not first by quality.

Figures as of 2026-10-16 (size, share, leaderboard) and 2026-09-19 (the AAPL book slice); transfer and DEX data on a 30-day basis. Structural terms from Ondo's secondary market restrictions and the Broadridge announcement. Research, not investment advice.