The question came from a developer in Yaba during one of the twenty-four workshops I ran across Lagos in 2017, back when I was twenty-seven and certain that whitepapers could fix everything. He wanted to know why he could buy a token representing anything on earth, but not twenty dollars' worth of Apple. I had no good answer. I talked about KYC, correspondent banking, and the long arm of securities law, and watched his face close.
Nine years later I opened the CFTC's product registry on a Tuesday in late September and found something that would have made him laugh. Coinbase Derivatives has filed a framework for perpetual futures on individual US stocks and ETFs — cash-settled, no expiry date, funding payments, a trading window running from Sunday 8 p.m. ET to Friday 5 p.m. ET. The registry status reads: Approval Pending. Not approved. Pending. That gap is where this story lives.
Let me set the table, because the framing matters more than the headline.
A perpetual future is a contract with no maturity date. BitMEX popularized the structure around 2016, and its central trick is the funding payment: a periodic transfer between longs and shorts that tethers the contract price to spot without ever forcing settlement. On crypto exchanges it became the dominant derivative. On US equity markets it is essentially unheard of — CME's single-stock futures expire monthly or quarterly, so anyone who wants to stay in a position pays a roll cost every time.
Coinbase's proposal, disclosed through a rule change that drew a joint SEC notice, points that crypto-native mechanism at Apple, at ETFs, at whatever single-name equity the appendix eventually lists. It is explicitly cash-settled: profit and loss are paid in fiat, and no shares change hands. Coinbase has the operating record here — its existing crypto perpetuals run 10x leverage at a 0.02% fee, and its US500 product peaked near $104 million in volume. US retail appetite for this structure is real, not theoretical.
But the SEC notice confirms what the CFTC registry implies: approval has not come. Two agencies, overlapping jurisdiction, no precedent.
Here is what I keep circling back to, and it is not the headline. The filing sets no single funding rate and no single funding interval. Both are deferred to a product appendix and market reference materials that have not been published.
That is not a footnote. That is the product. Funding is what makes a perpetual behave like a future rather than a casino chip. Set it wrong and the contract drifts away from spot. Set the interval too long and the drift compounds before arbitrage can correct it. Set it too short and thin order books get whipsawed into liquidations. For a single stock — with its own volatility profile, its own borrow dynamics, its own depth — the parameters cannot be copied from a bitcoin playbook. Apple's underlying market and a small-cap growth name are different animals. Until that appendix exists, nobody can model the cost of carry, and nobody can price the risk of holding the contract through a quiet weekend.
Speaking of weekends, there is the trading window: Sunday 8 p.m. ET through Friday 5 p.m. ET. A 24/5 compromise that creates a problem the filing does not answer. What is the reference price when the equity market is closed?
I have spent enough years staring at oracle feeds to know how this ends. The anchoring index is the product's real single point of failure, and it will almost certainly be a centralized data feed wearing decentralized language. Feed latency is the weak joint of on-chain derivatives, and calling a network decentralized while it aggregates a handful of permissioned operators is a marketing position, not an architecture. Equity perpetuals inherit that weakness with a worse excuse, because equity reference prices genuinely originate from a duopoly of exchanges. If the funding calculation reads a stale or thinly-traded print on a Sunday night, longs are paid or punished by an artifact that nobody can actually trade against.
Cash settlement is doing double duty. Financially, it strips out the settlement machinery and the heavier margin regime that physical equity delivery would demand. Legally, it is the argument that this is a futures contract and not a security. Run the Howey test and the design intent is legible: money invested, common enterprise, expectation of profit — all yes. The contested prong is reliance on the efforts of others, and Coinbase is arguing that returns come from market prices and the funding mechanism rather than from an issuer's management. The entire filing is aimed at that fourth prong.
It is a clever construction, and it may not be enough. A perpetual future on a single stock has a security as its underlying, and the SEC does not lose jurisdiction because the wrapper says cash-settled. That is why the joint notice matters more than any press release. What we are watching is a jurisdictional negotiation conducted in public, slowly, in the language of appendices.
Underneath all of it sits clearing. Coinbase Derivatives would operate as the central counterparty, which concentrates counterparty risk into a single margin engine — an engine that has to run through hours when the underlying market cannot help it discover price. Every perpetual venue I have watched break, broke at exactly this layer. Not the contract. The liquidation engine during a gap.
Yes, this is possible. And here is the bug we must fix.
Then there is the liquidity bootstrap. Funding only anchors a market when enough participants sit on both sides to arbitrage deviations. On day one, they will not. Someone — fee rebates, incentive programs, Coinbase's own balance sheet — has to pay market makers to quote through the first weeks. The competitive map is strange: CME has depth but no perpetuals; offshore venues like dYdX already list equity-linked perpetuals with no CFTC blessing at all. Coinbase's differentiation is not the mechanism. It is the compliance wrapper around it.
Now the angle I have not seen anyone raise, and the one I think will matter most.
Equities pay dividends. Perpetual futures do not.
Traditional equity futures price dividends into the basis — the gap between futures and spot reflects cost of carry, which is the interest rate minus the dividend yield. A perpetual has no maturity, so that carry has to live inside the funding rate instead. Which means the funding rate on a dividend-paying stock must, at minimum, track the risk-free rate minus the expected dividend yield, and it must step correctly around ex-dividend dates, when the spot price mechanically drops.
If the product appendix does not specify dividend treatment explicitly, the funding rate will be systematically mis-signed for part of every quarter. Longs will overpay the true carry on some names and underpay on others; arbitrageurs will harvest the gap; everyone else absorbs it as unexplained bleed. This is the kind of thing that never appears in a regulatory announcement and always appears in a retail P&L statement.
Trust the process, but verify the code. Right now there is nothing to verify — only a framework pointing at a document nobody outside the building has read. The approval headline is the loud part. The dividend adjustment clause is the quiet part, and it is where retail outcomes actually get decided.
My friend in Yaba may eventually get his twenty dollars of Apple, wrapped in a contract that never expires and settles in cash. I hope he does. But I would tell him what I tell every developer who joins my Code & Coffee sessions: read the appendix before you read the announcement, and watch the funding rate for the first month before you trust it in the sixth. When the registry finally flips from Approval Pending, what will you check first — the ticker, or the parameters?
