Deepstate

Paid for the first quote.

Deepstate has done what most tokenized-equity launches never get to: there is a live NVDA/USDG book, trades settle onchain, and makers are competing for the best price. That is the starting line, not the finish.

DeepstateMoni Score844WebsiteDocs

The project is a fully onchain matching engine on Robinhood Chain, built by Joseph Delong, SushiSwap's former CTO and a former Ethereum core developer at ConsenSys. Its broader bet is simple: tokenized stocks do not become a DeFi market because someone lists them. They need quotes, inventory, price discovery, settlement confidence and capital that stays after launch rewards stop doing the work.

joseph.ethMoni Score15106Website

The code tackles a real constraint. Orders are packed into 32 bytes, the book uses a radix tree, and the contracts repository includes SMT proofs, Halmos, KEVM/Kontrol and invariant tests. That does not create liquidity. It makes one part of the liquidity problem less punishing: a maker needs to be able to change a quote without gas becoming part of the spread.

It pays for the one number an exchange sells first

DEEP pays a narrow service: holding the best bid or best ask in NVDA/USDG when the next second turns over. A large resting order behind the touch earns nothing. Neither does volume that never improves the executable price.

Take a simple book. One maker bids $210.00 for 100 NVDA. Another bids $210.01. The second maker earns the bid-side rewards because they offer the better price. If someone steps in at $210.02, the reward moves again. The same contest runs on the ask.

That is cleaner than paying passive TVL or raw volume. Deepstate subsidises the price a trader can hit now, not capital that merely sits nearby. But the limitation is just as clear: a reward can buy a touch. It cannot prove there is meaningful size behind it, that makers can survive a gap, or that capital remains when the subsidy no longer covers inventory risk.

The rewarder is a handoff mechanism. The NVDA/USDG program holds 1 billion prefunded DEEP for 395 days, split between bid and ask. Early rewards are richer; earning the full stream later requires more inventory. At some point a participant stops looking like a farmer and starts running a market-making book.

The category has already shown where it breaks

EtherDelta and OasisDEX kept the book in Ethereum state. They worked as contracts, but every requote was a storage write. Market makers had to charge a spread wide enough to cover gas while AMMs were sitting next door.

0x, IDEX and later dYdX made the opposite compromise: keep the book offchain and settle onchain. Serum showed a fully onchain book could work when blocks were fast and fees were negligible. Its collapse after FTX was a separate lesson: matching logic does not settle the operator question.

Deepstate still has to clear three hurdles: cheap quote updates, real depth, and governance that does not reduce to one operator. The engine and vault owner point to a Governor contract rather than a personal upgrade key. That is better than an EOA key. It does not tell us who can direct the Governor once the first votes arrive.

The early activity is real. The quality of it is still open.

At the time of writing, 25 August, the public protocol dashboard showed roughly 106.17M DEEP distributed, 199 reward-earning farmers, a 73.3% top-ten share of claimed DEEP, a $0.0050 DEEP print and about $282.3k float market cap.

The team separately reported 563 wallets using Deepstate, 182 makers rewarded for quoting NVDA and $87.6M of cumulative USDG volume as of 22 August. Those are not the same cohort. A wallet that touched the app, a reward-earning address and a professional market maker are different things.

There is a functioning book. That already separates Deepstate from a tokenized-stock landing page. It is too early to call it a durable market. The harder proof is spread quality, executable depth and maker retention when NVDA moves.

DEEP and STATE are not the same bet

DEEP starts as a reward for holding the best quote. It is prefunded in the rewarder, not freshly minted at each claim. A maker can then deposit DEEP into the STATE vault, burning the DEEP and receiving STATE.

STATE carries voting power and a pro-rata claim on assets in the protocol vault. It is not a passive dividend token. To realise that claim, a holder uses redeemValue for USDG or redeemAssets for a basket, and burns the STATE being redeemed. Withdrawing value also gives up the same portion of governance and future vault claims.

That distinction matters. Burning DEEP reduces DEEP supply. Burning STATE removes a claim on the vault; it does not buy or burn another DEEP. Fees may make STATE attractive enough for someone to acquire and burn DEEP, but the protocol has a DEEP sink, not an automatic buyback.

The fee split is also narrower than the headline token loop. The official frontend charges 10 bps to the STATE vault and another 10 bps to Deep State Incorporated. The protocol layer is permissionless; the company runs a user-facing venue and keeps its own fee. STATE is therefore not equity in the interface business. Its claim is protocol-vault economics and governance over the contract layer.

The launch distribution makes this immediately political. The top ten addresses account for 73.3% of claimed DEEP. That does not prove ten wallets control governance, but it shows where the first political gravity sits. The first governance proposal is expected to open 30 August, with the vote snapshot on 2 September. Voting power comes from DEEP burned into STATE, so the early claim distribution is only a few steps away from the first electorate.

The real stress event is not another dashboard milestone

The book runs continuously. Nvidia's reference market does not. A maker can earn while the reference market is closed, when the adverse selection that normally punishes an equity quote is muted. That is a valid bootstrap choice. It changes what early volume and reward-capture figures mean.

The useful test is a live equity session: quoted spread, executable size at the touch, cancels, fills and which makers remain after a sharp move. Nvidia reports after the close on 26 August. A gap will say more about the book than another quiet reward window.

The capital answer is a roadmap, not evidence yet

Deepstate says thin secondary DEEP liquidity is deliberate: it wants holders to earn the token by making markets, not buy a float and wait. That makes sense if the goal is to recruit operators. It also means enough capital must quote both sides, carry inventory and take adverse selection. A reward curve cannot manufacture that forever.

The proposed answer is an ERC-4626 vault for NVDA and USDG, where users deposit and a strategist makes markets for them. If it ships with transparent risk limits and real net performance, that could let passive capital fund a specialist activity. For now it is a plan. Who runs the strategy, what loss limits apply and whether deposits remain after the initial yield are open questions.

The leverage roadmap has the same status. The team says it is working with an undisclosed lender; liquidations would need the ability to mint and burn NVDA after KYB. That could improve inventory financing. Until the lender, liquidation path and risk parameters are public, it is direction rather than product.

The read today

Deepstate is past “interesting code, no users” and nowhere near an institutional exchange. It has a real engine, an incentive aimed at quotes rather than vanity TVL, early market activity and a roadmap that addresses the capital problem directly.

The next evidence is behavioural: can tokenized NVDA keep a market through a gap, can the protocol bring in capital without subsidising every dollar forever, and can governance fund the team without turning the venue into a concentrated wrapper? Those are the project questions. The rest is launch data.

Dashboard values use the latest Dune reading available at publication. The displayed DEEP price is the most recent filled DEEP/USDG order, not an oracle or mark price. Website volume is an official interface reading, not an independently audited figure.