Posted July 13, 2026 — two weeks before CME's July 27 single-stock futures launch.

Full report PDF (16 pages)Addendum: The Basis Is the Debit Rate (8 pages)

On July 27, 2026, CME lists cash-settled futures on 55 US stocks — NVIDIA, Apple, Tesla, Meta, SpaceX among them — the first US security futures since OneChicago closed in 2020. In the same twelve months, perpetual-style equity index futures went live on a US-regulated exchange (Coinbase's AI10, China10, Defense10, Tech100), single-stock perps launched offshore at Coinbase International, Kraken, and Hyperliquid — whose builder-deployed markets cleared $62B in monthly volume in May — and the SEC and CFTC opened rulemaking consultations on the exact plumbing — portfolio margining, product definitions, 24/7 — that protects the bilateral swap business.

The question this piece takes seriously, from both sides: does the listed single-name complex disintermediate the equity swap desk and the prime brokerage franchise — the largest line in global banks' equities divisions at $34.5B in FY2025, +21.5% year-over-year (Coalition Greenwich)? The answer is segmented, not binary — and the segments are the interesting part.

The case study that anchors it: the leveraged single-stock ETF swap stack

The clearest public window into single-name swap economics is the 2x single-stock ETF complex — Direxion's TSLL (~$4.6B), GraniteShares' NVDL (~$4.1B), the T-Rex suite (the REX Shares / Tuttle Capital JV), and Defiance. These funds need 200% of NAV in daily-reset exposure and their N-PORT filings show how they get it: total return swaps, with counterparties and rates named. At the late-2024 peak, the two MicroStrategy 2x funds needed roughly $8B of MSTR swap exposure. The bulge bracket didn't write it — Cantor Fitzgerald, Marex, and Clear Street did, at OBFR +13% to +17% on notional. Compare that to the ~1.5% embedded swap cost GraniteShares disclosed on its mega-cap funds: the revenue is in the tail, and the tail exists because dealer capacity is constrained. In November 2024, Bloomberg reported Matt Tuttle asked his prime brokers for ~$100M of additional swap exposure into a close — and was offered $20M.

That episode cuts both ways, and it is the crux of the whole analysis. The swap desk's pricing power is real — that's what OBFR +17% is. And a listed future doesn't break it, because listing a contract doesn't create hedge capacity: the market maker quoting an MSTR-class SSF faces the same borrow scarcity and balance-sheet cost that produced the swap pricing, and the futures basis would embed a comparable spread. Meanwhile CME listed the easy names — the liquid mega-caps where spreads were already thin. Listed competition attacks the commodity end of the book first.

The prime brokerage moat in one number

A $500M/$500M long-short book in a portfolio-margin PB account nets: shorts finance longs, borrow is sourced internally, OCC TIMS (±15% stress) recognizes offsets, one negotiated spread prices the package. The same book rebuilt in single-stock futures posts 15% initial margin per leg — the statutory floor, set by the options-parity requirement — with no customer-level cross-margining against equities, options, or swaps. Cash variation margin daily. No short proceeds. No borrow rebate. At 15%, the SSF is no cheaper than portfolio margin; the offshore perp at 5-10% is cheaper, which is exactly why that demand went offshore. This margin math — not liquidity — is what killed OneChicago, and it is why the June 2026 SEC-CFTC portfolio-margining harmonization RFC is the single most load-bearing regulatory document for this thesis. The CME-FICC Treasury cross-margin extension to customer accounts (April 2026, savings up to 80%) proves customer-level cross-margining is no longer hypothetical.

The precedent that already ran to completion

At the index level, this movie already played. When UMR and SA-CCR made bilateral swaps capital-expensive, index TRS flow futurized: Eurex's EURO STOXX 50 Total Return Futures OI surpassed the conventional future's OI in March 2023, and CME's AIR Total Return futures hit a record $365B notional OI in September 2025, ADV +80% y/y. Three lessons transfer to single names: capital rules, not client preference, drive migration; dealers keep the revenue in a different shape (market-making, basis, clearing); and transparency compresses the commodity flow. One lesson does not transfer: the borrow. An index TRF has no stock-borrow leg. A single-name future on a hard-to-borrow does, and no listed structure manufactures lendable supply — the $15.3B securities-lending pool stays with whoever holds the inventory.

What about narrow-based index futures and baskets?

Conspicuously absent from every roadmap. A narrow-based index future is a security future — 15% floor, joint jurisdiction, no confirmed 60/40 — so exchanges engineer around the line instead: Coinbase's Mag7+Crypto future is 10 components, equal-weighted 10% each, which passes the broad-based tests and stays CFTC-only with 60/40 treatment. That's the listed wrapper aimed most directly at basket TRS and dispersion structures: an engineered index perp plus SSF overlays replicates most of a basket swap on-exchange. The June 2026 definitions RFC is where that boundary gets re-litigated.

The perpetual variable: Coinbase's stock perps and the access question

The sharper long-run competitor to the swap book may not be CME's quarterlies at all. Coinbase now runs single-name perpetuals — just not where US institutions can touch them. On Coinbase International Exchange (non-US only): perps on AAPL, MSFT, TSLA and SPY/QQQ, USDC-settled, 24/7, up to 10x, with $5.5B+ of institutional notional by Q2 2026. Onshore, Coinbase Derivatives lists only the engineered-broad-based index perpetuals (AI10, China10, Defense10, Tech100) — no US single-stock perp exists or is filed.

And there is no legal rail to bridge that gap from a prime broker's seat. The Coinbase/Deribit template — a US FCM intermediating an offshore venue as "foreign futures" under CFTC Part 30 — does not extend here: Part 30 covers foreign futures, not security futures, and the CFTC's May 2026 perp relief expressly excluded equity securities. The securities-side path is worse: the 2009 SEC order on foreign security futures lets QIBs trade foreign single-stock futures only where the underlying is a foreign private issuer whose primary market is outside the US. An offshore perp on Apple fails that test by construction. Access for US persons requires fresh SEC action — which is on the harmonization agenda, and explicitly not done.

What that means for PB today is leakage at the edges, not disintermediation: offshore-domiciled funds can trade the venue directly, and every dollar of USDC margin they post there is collateral outside the prime broker's netting set — a fragmentation cost that cuts both ways, since it also raises the value of consolidated PB margining for everything else. The Deribit precedent shows the endgame if relief ever extends to equities: access gets intermediated through FCMs, meaning the bank/broker keeps the client relationship and adds clearing revenue — the TRF pattern again, not replacement.

Are perps the better instrument for leveraged ETFs? Structurally, half yes: a perpetual is the closest listed replica of a TRS — open-ended, no quarterly roll footprint, financing accruing continuously — and on those dimensions it beats a dated future for a daily-reset fund. But three things block the migration case. Access: no onshore single-name perp exists, and a '40 Act fund can't use an offshore venue that excludes US persons. Funding economics: the perp's financing rate is set by the crowd, not negotiated — in retail-long-skewed names funding runs structurally positive (the crypto precedent is funding averaging high single to double digits annualized in bull phases), so a 2x long fund would swap a locked dealer spread for a floating retail-leverage premium. (The mirror image is interesting: inverse funds would collect that funding.) And tracking: funding accrual plus perp-to-spot gap adds daily noise against a fund whose product promise is 2x the official close-to-close return, which the TRS delivers contractually. If a US single-stock perp ever lists, the classification question decides everything: deemed a security future, it inherits the 15% floor and the SFP regime and the advantage evaporates; carved into a new category, it becomes the first genuinely retail-scale challenger. That is, once again, the June 2026 definitions RFC.

The counterargument: "my margin is lower, and my shorts carry no debit rate"

Every swap trader hears this one, and it deserves a precise answer rather than a dismissive one — it gets a full addendum (The Basis Is the Debit Rate, 8 pages). The short version: both halves confuse where a cost is printed with whether it exists.

Margin is not funding. The 15% initial margin secures performance against the clearinghouse; it finances nothing. The market maker on the other side hedges by carrying 100% of the stock on a dealer balance sheet — and charges for it in the basis, exactly as a swap desk charges for it in the spread: F = S × (1 + (r + s − b) × t) − PV(dividends). Same terms, different line item. And the 30–45% collateral seen in leveraged single-stock ETF swaps is a counterparty-risk price for 2x daily-reset funds on violent names — SPAN margin on those names would sit far above the 15% floor too.

The debit rate doesn't disappear; it's netted into the entry price. Short a 6%-borrow name for 90 days, spot at 100, SOFR 4.30%: the no-borrow forward is 101.075, but cash-and-carry arbitrage prices the actual future at 99.575. The short sells 1.50 points below the no-borrow forward — exactly 6% × 90/360 — and with spot unchanged loses $42,500 per $10M as the future pulls to spot. The PB stock short earns the rebate (SOFR − 6%) on proceeds: −$42,500. Identical to the dollar, except the swap reprices privately and continuously while the futures short reprices at four public rolls a year. Run the numbers yourself in the walkthrough tab below.

Drive the model

Four tools: the all-in cost of the same exposure across five channels (watch the ordering flip as you move from mega-cap to MSTR-class names), the step-by-step short-carry walkthrough behind the debit-rate argument, a revenue-at-risk model over the Coalition prime pool, and the segment-by-segment threat map.

The moats, ranked by durability

• Stock borrow & inventory — most durable. A perp funding rate can proxy borrow cost but cannot deliver the share.
• Netting & portfolio margin — durable but under direct review (the June 2026 RFC is the erosion path).
• Negotiated financing & balance sheet — cyclical. In December 2024 the listed channel was the expensive one: S&P futures implied financing peaked above 140bp over SOFR (D.E. Shaw). Listed is not automatically cheaper.
• SFP regime frictions — 15% floor, dual registration, no confirmed 60/40. Actively targeted by the harmonization agenda; CME launching anyway is a bet these get fixed.
• The bundle — capital intro, custody, cross-asset margin. Durable for institutions, irrelevant to the retail flow already leaving offshore.
• Section 871(m) / QDD plumbing — single names and narrow baskets carry dividend-equivalent withholding for offshore holders; qualified broad-based indexes don't. Full phase-in January 2027; it burdens offshore perp venues too.

Bottom line

Near term: low-single-digit erosion of the financing wallet — basis-point compression on liquid mega-cap spreads, new listed markets for dealers to make, minimal impact on hard-to-borrow and capacity-constrained flow. Medium term: contingent and dated. If customer cross-margining and 60/40 clarity arrive, the majority of vanilla synthetic longs comes into play and the desk's defense shifts from structural moats to service, borrow supply, and balance sheet. The asymmetry worth remembering: the most profitable swap flow is the least contestable, because listing a contract creates neither hedge capacity nor lendable shares. The franchise erodes from the commodity core outward — slowly and visibly, unless the cross-margining wall falls quickly, in which case the repricing is abrupt.

Full sourcing — CME/SEC/CFTC releases and orders, N-PORT filings, Coalition Greenwich, EquiLend, OFR, D.E. Shaw, FIA — in the full report; the cost-structure rebuttal is worked line by line in the addendum. Not investment advice.