The Perpetual Comes Ashore: Coinbase's Single-Stock Futures and the Anatomy of a Regulatory Stalemate

On September 22, a line item appeared in the CFTC's product registry that most crypto media scrolled straight past. "Single Stock Perpetual Futures Contract." Filed by Coinbase Derivatives. Status: Approval Pending. No ticker. No launch date. No funding rate schedule. No market maker named. Just three words in a government database that carry more structural weight than every token announcement this quarter combined. I've watched the perpetual contract travel a long way from where it started. BitMEX shipped the first one in 2016 — an instrument with no expiry, priced by a periodic funding payment exchanged between longs and shorts. It was originally a hack around the rollover friction of quarterly futures. It became the dominant derivative in crypto. Trillions in notional, most of it offshore, most of it invisible to the US regulatory perimeter. Now Coinbase wants to point that same mechanism at Apple. At an S&P 500 ETF. At individual US equities — cash-settled, traded Sunday 8 p.m. through Friday 5 p.m. ET. The ledger doesn't lie: the filing is real, the approval is not. What sits in that registry is a structure, not a product. So let me price the structure. The background matters here because most of the coverage treats this as a product launch. It isn't. It's a rule-change filing, and rule-change filings are the plumbing of exchange business models — invisible when they work, fatal when they stall. Coinbase Derivatives operates as a CFTC-registered venue, which means it lives inside the Commodity Exchange Act, not the Securities Exchange Act. That distinction is the entire ballgame. Anything listed on a designated contract market must answer to the CFTC. The moment the underlying reference looks like a security, the SEC's antenna goes up. That's why the cash-settlement design is not a footnote. It's the load-bearing wall. A contract that never delivers shares, that settles P&L in fiat, and that explicitly derives profit from the futures position rather than from holding and reselling the underlying asset — that construction is a deliberate attempt to keep the instrument on the derivatives side of the fence. It doesn't fully close the door. But it moves the conversation. I spent years auditing this exact class of decision from the other direction. In 2020, during the DeFi Summer, I manually walked through the first versions of Compound and Aave contracts looking for integer overflow paths that automated scanners missed. What I learned then applies here: the headline mechanism is rarely where the risk lives. The risk lives in the parameterization — the part that isn't in the pitch deck. And that's precisely the gap in this filing. The proposal does not standardize a funding rate or a funding interval. Instead, those parameters are delegated to a product appendix or market reference material — per-product, case by case. Read that carefully, because it tells you something the press release won't. A funding model that varies by underlying is a funding model that admits the underlying assets are not interchangeable. Apple has depth. A mid-cap growth name does not. Apply the same eight-hour funding cadence to both and you get two completely different price-anchoring behaviors, only one of which is stable. In crypto perpetuals, the funding rate is the entire anchor. It's the tether that keeps the perp trading near spot. When funding is well-calibrated and liquidity is deep, the perp and the underlying walk in lockstep and everything looks elegant. When liquidity is thin — and every new listing starts thin — funding swings violently, the basis widens, and the contract stops tracking the reference. I watched this happen repeatedly in the 2017 ICO era, when I was running triangular arbitrage across three pairs on early Uniswap forks and ShapeShift. The mispricings were real. So was the slippage. My system printed roughly $150,000 over four months before slippage costs ate the edge entirely. I pulled the capital. Post-mortem was simple: the liquidity wasn't there to support the mechanical spread I'd modeled, and the spread I was harvesting was the liquidity's absence, not a genuine inefficiency. Coinbase is walking into the same structural problem, except the reference asset is a US equity with a closing bell, and the trading window is not. That brings me to the deepest hole in the proposal: the 24/5 schedule. Sunday 8 p.m. ET to Friday 5 p.m. ET. Read that alongside the fact that US equities trade 9:30 a.m. to 4 p.m. ET, Monday through Friday, with pre- and post-market windows that are notoriously thin. A perpetual that runs 24/5 needs a reference price that runs 24/5. Where does it come from when the underlying exchange is closed? When Apple hasn't printed a quote since Friday afternoon and the futures contract is still live on Sunday night? The filing doesn't answer that. This is an important information gap, and I'd rather stare at it than pretend it isn't there. A perpetual contract with no expiry is only as good as its price anchor. If the anchor is a stale last-sale, the funding mechanism will be computing against an artifact. If the anchor is a composite of overnight futures and index quotes, then the issuer is quietly building a synthetic index — and the regulatory classification of that index becomes the real question, not the classification of the contract. I've seen this movie. In 2021, I ran statistical models tracking floor price deviations on OpenSea for CryptoPunks and Bored Apes, executing 42 large-volume trades during windows of extreme volatility. The mispricings were caused by illiquidity, not by mispricing in the fundamental sense. The floor price was a number, but it wasn't a price. A stale quote is not a market. It is a hopeful estimate wearing a market's clothing. Now map that onto a single-stock perpetual running through a weekend. The contract is live. The funding clock is ticking. The reference is dark. Any quant who has ever traded an illiquid instrument knows what happens next: the perp drifts, the basis opens, and the first participants with real capital get to set the price. That's not a flaw. That's a business model. And it is precisely the kind of business model that makes regulators nervous about retail exposure. Let me contrast the structure against what already exists, because the differentiation is thinner than the marketing implies. | Dimension | Coinbase Single-Stock Perp | Traditional US Equity Futures | Crypto Native Perp | |---|---|---|---| | Expiration | None | Monthly / quarterly | None | | Settlement | Cash, in fiat | Mostly cash | Cash, in stablecoin | | Funding rate | Framework proposed, appendix pending | N/A | Typically 8-hour cycle | | Trading window | 24/5 (Sun 8 p.m. – Fri 5 p.m. ET) | Exchange hours | 24/7 | | Liquidation | Continuous window, needs high-throughput engine | Concentrated in session | 24/7 | | Regulatory boundary | CFTC (futures posture) | CFTC | Fragmented, mostly offshore | Look at that third row. Traditional equity futures have no funding rate at all — the term structure is embedded in the contract's price across expirations. The whole point of a perpetual is that it removes the term structure and replaces it with a periodic cash transfer. That's the innovation. It's also the innovation that doesn't exist in the traditional equity derivatives stack, which means there is no precedent for how a single-stock perp's funding should behave relative to a dividend-paying equity, a hard-to-borrow name, or a stock with a scheduled earnings date. Consider dividends alone. A quarterly-expiring equity future prices dividends into its basis. A perpetual, by construction, cannot. It just pays funding forever. So when Apple goes ex-dividend, what happens to the perp's fair value? Either the funding model absorbs a discrete drop, or the contract prints a step-function basis relative to spot. Neither is described in the filing. I'm not saying it can't be solved. I'm saying that if you're going to underwrite liquidity for this instrument, you need to know which one it is before you quote a spread. I don't trade narratives. I trade mechanics. So let me be blunt about the market-structure layer. A perpetual with no expiry imposes a continuous clearing obligation. In crypto, that obligation is borne by an automated liquidation engine running 24/7. Coinbase Derivatives will need an equivalent for the 24/5 window — a risk engine that can margin, monitor, and liquidate positions when the underlying reference is one hour from open or has been closed for eight hours. That is a fundamentally harder engineering problem than liquidating BTC perps, because BTC never closes. The underlier's session structure leaks into the derivative's risk model, whether the product team acknowledges it or not. This is where the ecosystem-position argument gets interesting. Coinbase isn't just listing a product. It's positioning itself as the compliance bridge between crypto-native capital and traditional securities exposure. Look at the dependency chain: [Upstream] CFTC and SEC rulemaking, reference price sources, clearing banks, custodian infrastructure → [Coinbase Derivatives] matching engine, margin system, liquidation engine, market surveillance → [Downstream] Coinbase retail app users, API algo traders, prop desks, hedge funds → [Competitors] CME, Kraken, Bakkt, offshore perp venues The unique ecological niche here is not technical. Every component — matching, margining, funding, liquidation — already exists inside Coinbase's crypto perpetual stack. The niche is jurisdictional. Coinbase wants to be the first venue that offers a CFTC-sanctioned equity perpetual, which means the product's real asset is its license posture, not its order book. That's a strategic bet, and it's a defensible one. But it also means the product's success is not determined by Coinbase's engineering. It's determined by how two US agencies decide to split a boundary that has never needed splitting before. Now for the contrarian part, because the prevailing read on this filing is wrong and I want to say so carefully. The consensus interpretation is: Coinbase filed, therefore Coinbase launches soon, therefore bullish for COIN, therefore bullish for crypto broadly. That's a three-step logical hop with no empirical support at any stage. Start with the registry status itself. "Approval Pending" is not a countdown. It is a description of an indeterminate state. Bitcoin ETFs spent roughly a decade traveling from first filing to actual approval. That's not an anomaly; it's the base rate for products that require agencies to resolve classification questions that have no clean precedent. A single-stock perpetual is exactly that: a hybrid instrument with a crypto-native mechanism applied to a security-class underlier. There is no prior ruling to build on. The absence of precedent does not accelerate approval; it decelerates it, because every internal office has to invent its position rather than cite one. There's a second, sharper problem. I hold a specific view on how agencies operate, formed over years of watching enforcement actions outpace rulemaking: regulation-by-enforcement is not a failure of understanding. It is a deliberate policy of withholding clarity so that the ambiguity itself becomes a tool. When the rule is undefined, the regulator holds optionality on every future case. A firm filing like this one is asking that optionality to be surrendered, in writing, on a public docket. Nobody surrenders optionality cheaply. That's why I read the joint SEC-and-CFTC activity in this matter as coordination being visible precisely because the boundaries are contested — not because a decision is near. Now apply the Howey lens, because that's the test everyone will run and almost nobody will run honestly. Money invested — yes, margin. Common enterprise — arguable, with Coinbase Derivatives as the common counterparty. Expectation of profit — yes, by design. Profits from the efforts of others — this is the contested element, and it's contested because the honest answer is "partially." The contract's return derives from market prices, which are nobody's effort. But the contract's existence, its pricing anchor, its funding calibration, and its liquidity provisioning all derive from Coinbase's effort. A legal team will argue the first framing. A regulator will look at the second. The design defense is cash settlement. By never delivering shares, by never putting a security into a customer's account, the issuer keeps the instrument inside the derivative perimeter. This is not a naive choice; it's a considered one. It's also incomplete. Classification follows the economic reality, not just the settlement mechanics. If a product's P&L is functionally a leveraged bet on a single equity's price, the fact that it settles in dollars rather than shares does not automatically end the analysis. It shifts the burden. And burden-shifting is a negotiation, not a resolution. Here's what bothers me most about the bullish read: it assumes the product, if approved, will be immediately liquid. That assumption has no foundation. New perpetual listings start illiquid by definition. The first participants are market makers operating under incentives and a handful of directional traders taking a view. Funding will be volatile. The basis will be wide. Retail users will get filled at prices that make the exchange's own marketing uncomfortable. This is the normal life cycle of a perpetual, and there's no reason a single-stock version would skip it. The prior data point that gets cited is the US500 product, which reportedly peaked around $104 million in trading volume. That's real, and it validates that US users will trade derivative exposure through Coinbase. But it also sets the ceiling for expectations in the near term. Generating institutional-scale liquidity in a brand-new instrument class takes quarters, not weeks, and it requires market makers who are willing to warehouse single-name equity risk inside a crypto venue's risk framework at 24/5. That's a narrow set of firms, and they will price their participation accordingly. Volatility is just unpriced fear wearing a mask. Right now, the fear in this product isn't in the price — there is no price. It's in the schedule. The market is not trading an approval; it's trading a rumor of a schedule it has no visibility into. Let me close the loop on what I'd actually watch, because the useful output of any analysis is the set of variables that would change your position. First, the registry line. Any transition from "Approval Pending" to an approval status is the single largest discrete catalyst in this entire narrative, because it resolves the biggest unknown. That is the level. Not the COIN chart, not the crypto beta — the government database entry. Set the alert there. Second, the reference price methodology. This is where I'd focus my diligence if I were underwriting market maker exposure. The exchange will eventually have to disclose how the contract is priced during non-session hours, and that document will tell you more about the product's real risk profile than anything in the current filing. If the anchor is a stale last sale, expect funding instability and wide spreads. If it's a constructed composite index, expect a second classification debate that will run in parallel to the first. Third, the funding parameter appendix. Once published, model the historical funding volatility of the closest existing single-name derivatives and ask whether the published parameters would have survived the 2020 or 2022 regimes. If they wouldn't have, the parameters are marketing, not risk management. Fourth, competitor motion. If CME publishes a similar roadmap, Coinbase's first-mover advantage compresses and the strategic premium in the filing evaporates. Fifth, the actual volume in the first month of any launch. Compare it to the US500 baseline. If it clears, the demand hypothesis holds. If it doesn't, the demand hypothesis was always a story about Coinbase's user base rather than about equity-perpetual demand. Silence is the only honest signal in the noise, and right now the signal is the empty space in the filing — the funding rate left blank, the non-session price source left unnamed, the launch date left nonexistent. Those gaps aren't oversights. They're the actual state of the product. Arbitrage waits for no one, and neither should you. But this particular arbitrage isn't between two prices — it's between the market's expectation of approval and the structural reality that approval is a multi-agency classification problem with no precedent and no clock. If you're positioned for a launch, you're positioned for the wrong event. The event is the ruling. And the ruling has no ETA. The floor isn't where the marketing says it is. It's where the registry entry still reads "pending."