Built by Grok 4.6 HighCME Group Inc. reported third-quarter 2026 results on Wednesday, October 21, 2026, before the open. This analysis was prepared Monday, October 26, 2026. Quarter ended September 30, 2026. CME is a calendar fiscal, so the period slug is 2026Q3. The next print is Q4 2026, expected about February 3, 2027.
CME | Earnings Review — Q3 2026
Verdict: the growth rate repaired, the level did not reset. Q3 revenue of $1,737.3 million grew +13.0% versus $1,537.6 million, and adjusted diluted EPS of $3.02 grew +12.7% versus $2.68. Those rates are +1,215 bps and +1,167 bps faster than Q2's +0.8% and +1.0%. They are still slower than Q1 2026 (+14.5% revenue, +20.0% adjusted EPS). This is a rebound off the Q3 2025 trough, not a new run-rate above the Q1 spike.
What drove it. Average daily volume was 29,445 thousand contracts, +16.3% versus 25,322 thousand, a third-quarter record and double-digit in all six asset classes. Average rate per contract fell to $0.679 from $0.702, −3.3%, so clearing and transaction fees grew only +12.2% to $1,378.1 million. Market data held a record $243.1 million, +20.0%, but that rate did not re-accelerate (Q2 was +20.2%). It is about 14% of revenue. Volume still sets the growth rate.
The beat is small because the volume was already public. Versus the print-day FMP bar (adjusted EPS $2.99, revenue $1,714.9 million, matching the FactSet bar cited by Dow Jones that morning), the surprise is +$0.03 / +1.0% and +$22.4 million / +1.3%. The late-September FMP vintage was $2.94 and $1,692.3 million. Scoring against that older bar would invent a +2.7% beat. Adjusted EPS has not missed in the last 12 quarters on the FMP series (11 beats, 1 in-line). Surprise magnitude is not improving: last-four average +1.3% versus +2.0% in the oldest four.
Margins followed volume. GAAP operating income was $1,141.4 million, a calculated margin of 65.7%, +245 bps versus the year-ago quarter after −186 bps in Q2. Adjusted operating income was $1,205.5 million, calculated margin 69.4%, +104 bps. Adjusted expenses of $531.8 million grew +9.3%, slower than revenue, but equity-complex license fees kept EPS growth in line with revenue rather than ahead of operating income. GAAP diluted EPS of $2.89 is not the Street test. Q4 2025 GAAP EPS of $3.24 is not a clean run-rate either. Follow adjusted EPS.
Guidance did not move. The caveat did. FY2026 adjusted operating expense excluding license fees is still $1.695 billion, capex about $85 million, and the adjusted tax rate 23.5%–24.5%, expected at the low end. Fourth straight print. The new sentence is that continued Q4 outperformance would lift performance-linked variable costs. Nine-month ex-license spend leaves about $450 million of room, only about +0.8% above last year's fourth quarter, while management said Q4 volume was running +17% versus last year's record fourth quarter. There is no revenue or EPS guide. FMP's Q4 revenue average of about $1,762 million is only +6.9% versus $1,648.7 million. That is the management-versus-street gap. It is a volume comment, not a formal guide, and Q3 already showed volume can outrun fees by a few points when the rate falls.
Tone, contradictions, catalysts. The call was more confident than July ("exceptional," record third-quarter ADV, "firmly on track" for another record year) and more careful about 2027 than about Q4. Five of 15 questions were deflected, including whether CME would list perpetuals if it loses, the 2027 open-interest bridge, and Treasury Link economics. Three conflicts are unreconciled: a $100 billion margin-savings headline that cannot sit beside the same call's $30 billion line, April's "perpetuals are against the law" claim that is now an open lawsuit, and a sports-event stance that fights CME's own book. Compute futures, Treasury clearing, and single-stock futures are not in the Q4 or FY2027 revenue algorithm (FMP FY2027 revenue +5.6% versus its own FY2026 number). Open interest did not confirm the volume.
What this is not. Not a price note. No stock price, market cap, or multiple is used. Bloomberg and Visible Alpha were not connected. Consensus is FMP, pulled 2026-10-26, and is labeled as such. SharePoint, OneNote, Outlook, Teams, and internal Excel models were unavailable and were skipped.
| Revenue | $1,737.3M (+13.0% YoY, +1.3% vs FMP) | Adjusted diluted EPS | $3.02 (+12.7% YoY, +1.0% vs FMP $2.99) |
| GAAP diluted EPS | $2.89 (+16.1% YoY) — not the Street test | ADV | 29.4M contracts (+16.3% YoY); all six classes double-digit |
| Rate per contract | $0.679 (−3.3% YoY); clearing fees only +12.2% | Market data | $243.1M (+20.0% YoY); rate stalled after +20.2% in Q2 |
| Adjusted operating margin | 69.4% (calc., +104 bps YoY) | GAAP operating margin | 65.7% (calc., +245 bps YoY) |
| FY2026 ex-license opex | $1.695B reaffirmed; Q4 outperformance can lift variable cost | Q4 FMP vs volume color | Revenue bar +6.9% / EPS +9.2%; volume running +17% (partial) |
| L12 / L4 EPS beats | 11/12 plus one in-line · 4/4 · magnitude not improving | Deflections | 5 of 15 questions; open interest did not rise with ADV |
CME is an exchange. Average daily volume times rate per contract produces clearing and transaction fees. Market data is the smaller, subscription-like second engine. There is no cost-of-goods line, so GAAP operating margin is the gross-margin substitute. Adjusted diluted EPS is the series the Street models. Margins below are calculated from the cited dollar lines. Basis points use unrounded ratios.
| Metric | 23Q4 | 24Q1 | 24Q2 | 24Q3 | 24Q4 | 25Q1 | 25Q2 | 25Q3 | 25Q4 | 26Q1 | 26Q2 | 26Q3 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ADV (000 contracts) | 25,513 | 26,360 | 25,941 | 28,289 | 25,503 | 29,768 | 30,217 | 25,322 | 27,353 | 36,231 | 29,843 | 29,445 |
| ADV YoY | +17.0% | -1.9% | +13.5% | +26.6% | 0.0% | +12.9% | +16.5% | -10.5% | +7.3% | +21.7% | -1.2% | +16.3% |
| Rate per contract ($) | 0.682 | 0.695 | 0.708 | 0.666 | 0.701 | 0.686 | 0.690 | 0.702 | 0.707 | 0.652 | 0.678 | 0.679 |
| RPC YoY | +4.8% | +4.7% | -2.2% | -5.8% | +2.8% | -1.3% | -2.5% | +5.4% | +0.9% | -5.0% | -1.7% | -3.3% |
| Clearing and transaction fees ($M) | 1,184.5 | 1,208.9 | 1,249.9 | 1,297.1 | 1,232.3 | 1,337.3 | 1,388.0 | 1,227.9 | 1,327.9 | 1,542.6 | 1,352.5 | 1,378.1 |
| Clearing fees YoY | +20.7% | +0.7% | +11.8% | +19.5% | +4.0% | +10.6% | +11.0% | -5.3% | +7.8% | +15.4% | -2.6% | +12.2% |
| Market data ($M) | 167.2 | 175.4 | 175.0 | 178.2 | 181.6 | 194.5 | 198.1 | 202.5 | 208.0 | 224.1 | 238.1 | 243.1 |
| Market data YoY | +9.1% | +5.8% | +7.3% | +6.3% | +8.6% | +10.9% | +13.2% | +13.6% | +14.5% | +15.2% | +20.2% | +20.0% |
| Total revenue ($M) | 1,439.3 | 1,487.9 | 1,532.5 | 1,584.4 | 1,525.3 | 1,642.3 | 1,692.0 | 1,537.6 | 1,648.7 | 1,880.1 | 1,706.2 | 1,737.3 |
| Revenue YoY | +19.2% | +3.2% | +12.7% | +18.4% | +6.0% | +10.4% | +10.4% | -3.0% | +8.1% | +14.5% | +0.8% | +13.0% |
| GAAP operating margin | 60.0% | 64.5% | 65.3% | 64.6% | 62.1% | 67.5% | 66.7% | 63.3% | 61.8% | 69.7% | 64.9% | 65.7% |
| GAAP OM YoY (bps) | +464 | +111 | +364 | +333 | +212 | +297 | +145 | -139 | -25 | +219 | -186 | +245 |
| Adjusted diluted EPS ($) | 2.37 | 2.50 | 2.56 | 2.68 | 2.52 | 2.80 | 2.96 | 2.68 | 2.77 | 3.36 | 2.99 | 3.02 |
| Adjusted EPS YoY | +23.4% | +3.3% | +11.3% | +19.1% | +6.3% | +12.0% | +15.6% | 0.0% | +9.9% | +20.0% | +1.0% | +12.7% |
| GAAP diluted EPS ($) | 2.24 | 2.35 | 2.42 | 2.50 | 2.40 | 2.62 | 2.81 | 2.49 | 3.24 | 3.18 | 2.88 | 2.89 |
The rate decline is mostly within product, not a shift into the cheap rates complex. Rates RPC was $0.489 versus $0.487. Equities fell to $0.605 from $0.652 (−7.2%), energy to $1.146 from $1.214 (−5.6%), agriculture to $1.377 from $1.423 (−3.2%), and metals to $1.326 from $1.505 (−11.9%). FX was the exception, $0.852 versus $0.841.
All six classes are double-digit after a Q2 in which rates, FX, energy, and metals were negative or flat. Equities were already growing and accelerated further, which is why license fees rose. Metals' +114% / +130% spike in Q4 2025 and Q1 2026 has cooled to +13.3%. Do not extrapolate the gold and copper quarter. Financial products and commodities each grew about 16%, matching the prepared remarks.
The annual tape into FY2025 was still decelerating: revenue YoY +11.1%, +9.9%, +6.4%. Market data was the exception, +13.1% in FY2025. FMP's FY2026 revenue of $7,091 million is +8.7% versus $6,520.6 million, and FY2027 is only +5.6% on top of that. Nine-month 2026 revenue is $5,323.6 million versus $4,871.9 million, +9.3%. Holding the FMP full-year number implies a Q4 of about $1,768 million, near the quarterly FMP bar and slower than Q3.
| Annual | FY2021 | FY2022 | FY2023 | FY2024 | FY2025 | FY2026E FMP |
|---|---|---|---|---|---|---|
| Revenue ($M) | 4,689.7 | 5,019.4 | 5,578.9 | 6,130.1 | 6,520.6 | 7,091 |
| Revenue YoY | — | +7.0% | +11.1% | +9.9% | +6.4% | +8.7% |
| Market data ($M) | 576.9 | 610.9 | 663.7 | 710.2 | 803.1 | — |
| Market data YoY | — | +5.9% | +8.6% | +7.0% | +13.1% | — |
| EBITDA ($M) | 2,793.0 | 3,150.8 | 3,561.7 | 4,046.6 | 4,337.0 | — |
| EBITDA YoY | — | +12.8% | +13.0% | +13.6% | +7.2% | — |
| Adjusted diluted EPS ($) | 6.69 | 7.98 | 9.34 | 10.26 | 11.21 | 12.39 |
| Adjusted EPS YoY | — | +19.3% | +17.0% | +9.9% | +9.3% | +10.6% |
The three lines move together: volume, revenue, and adjusted EPS. Market-data YoY is the smooth climb that paused at +20%. Operating leverage is back because the cost base is largely fixed. It was not a pricing story. October 1 fee changes (0.5%–1% of clearing fees, annualized, on management's count) started after this quarter closed.
A small double beat of a bar the Street had already raised, on a year that re-accelerated. CME does not guide quarterly revenue or EPS, so there is no company-guide cell. GAAP EPS is shown so it is not mistaken for the miss. FMP's EPS actual is the adjusted number.
The print-day bar already implied about +11.5% revenue growth and +11.6% EPS growth. The company cleared a bar that had absorbed most of the volume. Of the +$199.7 million year-over-year revenue increase, clearing fees are about three-quarters and market data about one-fifth. The Street beat of $22 million is roughly a tenth of that year-over-year dollar increase. Audits of $4.7 million, a spoken figure, are too small to explain a +20% data line that was already running at that rate in Q2.
| Surprise vs FMP | 24Q4 | 25Q1 | 25Q2 | 25Q3 | 25Q4 | 26Q1 | 26Q2 | 26Q3 |
|---|---|---|---|---|---|---|---|---|
| Adj. EPS surprise | +2.4% | 0.0% | +1.7% | +1.9% | +0.7% | +0.6% | +2.7% | +1.0% |
| Revenue surprise | +0.7% | -0.2% | +0.4% | +0.5% | +0.1% | +0.3% | +1.4% | +1.3% |
Pattern: consistent EPS beater, compressed surprise. Last 12 quarters (2023Q4–2026Q3): adjusted EPS 11/12 (91.7%), plus one in-line, zero misses, average surprise +1.6%. Last four: 4/4, average +1.3%. Oldest four averaged +2.0%. Inside the last four the path is +0.7%, +0.6%, +2.7%, +1.0%. Q3 gave back most of Q2's re-widening. A 1% EPS beat is the run-rate, not evidence the Street has lost the model.
Revenue is mixed over 12 quarters (9/12) and 4/4 over the last four. Average revenue surprise stepped from +0.3% to +0.4% to +0.8%. The last two prints (+1.4% and +1.3%) are the widest dollar surprises in the window and are still about $22–24 million. The three revenue misses (2024Q2, 2024Q3, 2025Q1) are $0.9 million, $4.3 million, and $2.5 million. That is what a model looks like when ADV is known before the print.
2026Q1 is the one quarter where the print-day wire disagreed with FMP ($3.34 retained, a small beat; some wires used about $3.37–$3.38). If that quarter is rescored as a miss, L12Q becomes 10/12 and L4Q becomes 3/4. Either reading is a beater. The Q2 2026 "$3.01 miss" in an older workspace was a pre-print aggregate. Print-day FMP scored Q2 a beat versus $2.91. It is not used as a miss here.
Management never said "beat." The remarks explain the year-over-year result: record third-quarter ADV, a lower rate, market data at +20% for a 34th straight quarter, and expenses up less than revenue with license fees on the equity complex. The +1% residual is what was left after monthly ADV had already moved the bar.
CME guides three full-year items only: adjusted operating expense excluding license fees, capex, and the adjusted effective tax rate. It does not guide revenue, EPS, margin, ADV, or rate per contract. The point estimates did not change. The new sentence is the release valve.
| Guided item | FY25Q4 | 26Q1 | 26Q2 | 26Q3 | Change |
|---|---|---|---|---|---|
| Adj. opex ex-license | $1.695B | $1.695B | $1.695B | $1.695B | Unchanged, fourth print. Now activity-contingent if Q4 stays hot. |
| Capex, approximately | $85M | $85M | $85M | $85M | Unchanged. Immaterial to the P&L. |
| Adj. tax rate | 23.5%–24.5% | 23.5%–24.5% | 23.5%–24.5% | 23.5%–24.5% | Unchanged. Still "low end." Q3 actual 23.8%. |
| Revenue, EPS, margin | Not guided | Not guided | Not guided | Not guided | No FY2027 P&L guide either. |
The ceiling is no longer slack. Nine-month non-GAAP operating expense was $511.7 million, $521.2 million, and $531.8 million. License fees were $106.8 million, $109.1 million, and $104.2 million. Ex-license year-to-date is about $1,244.6 million (arithmetic, not its own series). Room under $1.695 billion is about $450 million, against a Q4 2025 ex-license actual of about $447 million ($543.4 million minus $96.6 million). Versus FY2025 delivered ex-license cost of about $1,624.8 million, the guide is only about +4.3%, slower than the +8.7% revenue growth in the FY2026 FMP consensus — if the ceiling holds.
Q4 and the outer year are the Street's numbers, not CME's.
| Forward marker | Base | Company | FMP | Implied YoY |
|---|---|---|---|---|
| Q4 2026 revenue | $1,648.7M | Not guided | $1,762M | +6.9% |
| Q4 2026 adjusted EPS | $2.77 | Not guided | $3.03 | +9.2% |
| Q4 volume, early | 27,353k ADV | +17% vs last year's Q4, through the call | No ADV estimate | Color, not a guide |
| FY2026 revenue / EPS | $6,520.6M / $11.21 | Not guided | $7,091M / $12.39 | +8.7% / +10.6% |
| FY2027 vs FY2026 FMP | FMP FY2026 | Not guided | $7,485M / $13.20 | +5.6% / +6.5% |
FY2025 was the template for hitting a guide: the original ex-license figure was cut in-year and the year landed on the cut. FY2026 has been held for four prints. The Q3 caveat cracks that firmness toward higher cost if volume stays hot, not a demand cut. That is a high-quality problem only if incremental revenue still drops through. Tone versus July: confidence went up with the print, and the restraints are specific (the expense valve, open interest that is not growing with volume, and a refusal to put a number on 2027 new products). No guide was withdrawn. No CFO or CEO change was discussed on this call. Jack Tobin's move to CFO in March 2027, with Fitzpatrick becoming CEO, was announced September 9 and was not re-litigated.
The growth rate is a volatility cycle, not a smooth compounder. Across eight quarters, revenue YoY swung from −2.95% to +14.48%, and adjusted EPS YoY from 0.00% to +20.00%, with four reversals in the sign of acceleration. Acceleration below is the change in the year-over-year rate versus the prior quarter's year-over-year rate, in basis points. It is not a sequential change in dollars.
| YoY rate | Q4'24 | Q1'25 | Q2'25 | Q3'25 | Q4'25 | Q1'26 | Q2'26 | Q3'26 |
|---|---|---|---|---|---|---|---|---|
| Revenue YoY | +5.98% | +10.38% | +10.41% | -2.95% | +8.09% | +14.48% | +0.84% | +12.99% |
| Revenue acceleration (bps) | -1,246 | +440 | +3 | -1,336 | +1,104 | +639 | -1,364 | +1,215 |
| Adjusted EPS YoY | +6.33% | +12.00% | +15.63% | 0.00% | +9.92% | +20.00% | +1.01% | +12.69% |
| EPS acceleration (bps) | -1,278 | +567 | +363 | -1,563 | +992 | +1,008 | -1,899 | +1,167 |
| Inflection | What the rate did | What was underneath it |
|---|---|---|
| Q3 2025 trough | Revenue -2.95% (-1,336 bps). Adjusted EPS 0.00% (-1,563 bps). Only negative revenue print in the window. | ADV 25,322 thousand, -10.5% versus 28,289 thousand. Rate was not the problem: +5.4%. |
| Q1 2026 peak | Revenue +14.48%. Adjusted EPS +20.00% (+1,008 bps). Fastest EPS growth in the window. EPS outgrew revenue. | ADV 36,231 thousand, +21.7%. Rate $0.652, -5.0%. Operating leverage on a volume spike. |
| Q2 2026 stall | Largest single-quarter drop in the rate. Revenue +0.84% (-1,364 bps). Adjusted EPS +1.01% (-1,899 bps). | ADV 29,843 thousand versus 30,217 thousand, -1.2%. A hard year-ago record, not an air pocket. |
| Q3 2026 re-acceleration | Revenue +12.99% (+1,215 bps). Adjusted EPS +12.69% (+1,167 bps). The two rates met again. Leverage did not return. | ADV +16.3% off the trough. Absolute ADV (29.4 million) is in line with Q2, not above the Q1 spike. Rate -3.3%. Market data +20.0%. |
Summing the three reported 2026 quarters matches the year-to-date comment: revenue +9.3% and, on the call, adjusted EPS about +11%. That is a high-single-digit year with the quarterly rate whipping around it. Absolute Q3 activity is a record third quarter and is healthy. It is not a step-up from the 29.8 million contracts a day in Q2 2026 or the 30.2 million in Q2 2025. Rate per contract has been a year-over-year drag in every 2026 quarter fetched here (−5.0%, −1.7%, −3.3%).
The next print is the test of whether +13% sticks. FMP's Q4 bar is about 610 bps slower on revenue and 345 bps slower on adjusted EPS than the Q3 rates. If Q4 revenue lands near that bar, Q3 will have been a one-quarter rebound off an easy comp. Q4 2025 ADV of 27,353 thousand is a firmer base than Q3 2025, so a sustained +17% volume pace would be a stronger statement than the Q3 comparison. The +17% is twenty days, not a finished quarter. One reading of the materials puts quarter-to-date ADV near 29.7 million, which would be only about +8.5% if it were the full-quarter pace versus the Q4 2025 average. Do not annualize the partial compare.
Open interest says the quarter was velocity. Duffy: quarter-end open interest 128 million contracts, and over 133 million on the morning of the call, the highest end-of-September open interest on record. Ken Worthington's figures, which management did not confirm as a filing, were equity open interest +0.2%, rates +1.2%, and energy −3.6%. The answer was that shorter-dated and weekly options expire positions that used to sit as open interest. Read the growth-rate rebound as activity against a weak Q3 2025 comp, not as a bigger positioned book.
Q4 consensus underwrites about +6.9% revenue. Adjusted EPS is +9.2% on the analyst-average file ($3.03, $3.026 unrounded, versus $2.77). The FMP earnings endpoint for the February 3 print is $3.01, which is +8.7% on the same base. Revenue agrees across the two files. It gives essentially no credit to Compute futures, Treasury clearing, or single-stock futures, and no debit to an approved S&P perpetual on a competing venue. Where the call and the model diverge is the volume trajectory, not the product list. Fitzpatrick refused to raise the old "~2% of growth from new products" algorithm. The model agrees with her on the dollars.
| Catalyst | Timing | What management said | What consensus has |
|---|---|---|---|
| Q4 volume vs the +6.9% revenue bar | Through Oct 20, inside the quarter FMP is modeling. Report about Feb 3, 2027. | Volume 17% ahead of last year's record fourth quarter; year-to-date volume 12% ahead. Partial quarter. Not an ADV guide. | $1,762 million revenue, about +6.9% versus $1,648.7 million. Coherent only if RPC keeps falling the way equity (-7.2%) and metals (-11.9%) did. |
| October 1 fee card | In effect for all of Q4. First full print is February. | 0.5%–1% of clearing and transaction fees, annualized, on crypto, some grains, interest-rate tiers, and micro equities. Annual pricing freeze is being dropped. April changes are already in. | No fee bridge. A few tens of millions annualized on the $1,378.1 million clearing-fee run-rate. Inside the rounding of a $1.76 billion quarter. |
| Rates complex and the Fed | Q3 already includes the Sept 16 hike to 3.75–4.00%. October and December meetings are ahead. | No funds-rate call. Activity framed as shifting central-bank policy, geopolitics, and volatility. Rates ADV 15,121 thousand, +13.0%, after -6.1% in Q2. | The +6.9% revenue bar assumes the +13% rates-ADV pace cools. A contested December decision is a volume catalyst. It is also the variable under clearing-house interest income. |
| Single-stock futures | Listed July 27. Q4 is the first full quarter. | A little over 350,000 contracts and about $3.5 billion notional since launch (call). Live at 25 of the top retail brokers. Not a prediction contract. | No revenue line. A few thousand contracts a day against equity ADV of 7,714 thousand. Well under 0.1% of the equity complex. Cannot explain +22.9%. |
| Compute futures | CFTC review extended. Comment window ends on or about Nov 9. Not a Q4 launch unless the agency moves immediately. | Clients are calling (neoclouds, data centers, banks, funds, chip providers). Timing is the government's. No hyperscaler lift was identified inside energy ADV. | None. A post-November approval is a 2027 option inside FY2027 revenue growth of about +5.6%. |
| Treasury Link and securities clearing | Link in a test environment the week of the call. Securities clearing guided for Dec 7, ahead of the Dec 31 cash-Treasury mandate. Repo mandate June 30, 2027. | Guaranteed cash-versus-futures spread, strong demand, no revenue size. Launch has already slipped from an initial Q2 2026 expectation. | Three weeks of Q4 cannot move a $1.76 billion quarter. FY2027 +5.6% does not look like a clearing windfall. No BrokerTec fee model. |
| Perpetual-futures case | Motion to dismiss opposed. Merits expected before the judge in December. | Confident on swaps-versus-futures. Would not say CME lists equity-index perps if it loses. 94% of volume called institutional (call). | Not in the $3.03. Did not show up as a Q3 fee hole. Treat as unresolved franchise competition, not a Q4 earnings catalyst. |
| 24/7 metals vs withdrawn oil | Gold and silver live. 10-barrel WTI filing withdrawn Oct 2. | Weekend share about 20–30% of ADV in contracts that are already 24/7 (deck, not disputed). Commercials did not want a weekend oil desk. Primary contracts wait on bank infrastructure. | The metals RPC decline of 11.9% is the part of 24/7 already in the income statement. No weekend-oil revenue. |
| Expense ceiling | February check. | $1.695 billion holds unless Q4 outperformance lifts variable cost. License fees move with equity volume. | Q4 EPS of $3.03 (+9.2%) is a decelerating growth rate with a stable cost guide, not a margin-expansion story. |
Retail acquisition is already inside equity ADV and the non-professional third of market data: tenth straight quarter of double-digit client-acquisition growth, and well over 100,000 new retail traders, +40% versus the year-ago quarter (call, not a Daloopa series). Event contracts and NHL FutureSports indices are not the bridge to $1.76 billion. International ADV was described as 9 million, +22%. That geographic cut is management.
The useful version of the setup is specific. Management keeps repeating client demand for products the street is not putting in the model, then refuses to raise the new-product growth algorithm. The disagreement that can hit a number before 2027 is whether October's volume pace broadens at a stable rate, or whether RPC keeps absorbing it. A +17% volume start can still land near a high-single-digit revenue rate. It is too low a revenue bar if the pace holds and the rate stabilizes. It is not a license to model Compute, clearing, or perpetuals into Q4.
Fifteen questions on the October 21 call. Ten were well answered (67%). Five were deflected (33%). The tape did not relitigate the print. The street pressed durability: is the volume re-acceleration real positioning, or faster turnover, and will CME list the products partners are using to peel off flow.
Well answered, and backed with numbers: volume across all six asset classes; adjusted margin +104 bps as expenses (+9.3%) lagged revenue (+13.0%); market-data mix in three roughly equal buckets with audits of $4.7 million; October pricing at 0.5%–1%; single-stock distribution; equity margin savings split, with about 60% of the 2026 increase tied to new positions rather than mark-to-market (call); capital return in-lane, OSTTRA repurchase finished. Refusing to invent an undisclosed S&P / Cboe step-up is the right answer, not a deflection. The step-up starts in 2027.
| # | Analyst | What was withheld | Why it matters |
|---|---|---|---|
| 3 | Alex Kramm, UBS — net interest | No dollar bridge from a roughly 2% sequential drop in clearing balances to the interest move he saw. Spread described as stable at about 33 bps. Average balances $147 billion versus about $150 billion in Q2 (call). | Gross investment income was $1,418.1 million, -8.4% YoY, and mostly not fee revenue. Cannot tell the keep from the pass-through. |
| 6 | Craig Siegenthaler, BofA — list perps if lost | "I would not make a decision one way or another." Case opposed; expected fully before the judge in December. | The franchise question. A win protects the moat. A loss with no CME product cedes the retail wrapper. Neither branch was planned in public. |
| 10 | Ken Worthington, J.P. Morgan — open interest into 2027 | Weekly expiries offered as the reason. His +0.2% / +1.2% / -3.6% open-interest figures were not confirmed. No 2027 factor. | ADV re-accelerated to +16.3% without a confirmed rise in open interest. If that is turnover on a flat stock of risk, the acceleration does not carry. |
| 11 | Ben Budish, Barclays — Treasury Link economics | Product description and "strong demand" only. No market size. No collateral-versus-BrokerTec split. | The strategic rates project into the clearing mandate is unmodelable after this answer. |
| 12 | Brian Bedell, Deutsche Bank — leverage caps | Overseas leverage of 25x to 250x was the risk he meant. He will not opine on a U.S. cap until it is in writing. No approval odds. | Same hole as the listing question. Retail growth makes the refusal on product design more conspicuous, not less. |
Duffy's repeated line is that clients stay because the institutional liquidity pool and the margin offsets cannot be copied by a retail platform. He answered the philosophy. He did not answer the two numbers that would decide it: how much of the +16.3% ADV is velocity that fades if open interest stays flat, and what CME does with perpetuals if the court says they are futures. Fitzpatrick does not see a fundamental change in the new-product contribution. 368 new products year to date do not change that.
Three unreconciled conflicts. Two are high severity, one is medium. The operating print is internally consistent: revenue rank, adjusted EPS, margin, market data, and the six ADV lines hold. The conflicts are a KPI management leads with, a legal claim, and a product policy. Labeled updates (the end of the annual pricing freeze, holding the 24/7 oil filing, Treasury Link still in test) are not counted.
Duffy, prepared remarks, October 21: record capital efficiency, "an average of over $100 billion a day in margin savings." He repeats it as the company-wide number. Mike Dennis, later on the same call, cites "$30 billion in margin savings across futures, options, swaps, and cash." Tim McCourt then puts equity savings alone at just under $47 billion ($42 billion base futures-and-options offset, a little over $5 billion OCC cross-margin). Equity alone is already larger than the $30 billion attached to "futures, options, swaps, and cash."
The pair is three calls old. The smaller number steps $25 billion (Q4 2025) to $27 billion (Q2 2026) to $30 billion. The larger number steps $80 billion to $95 billion to over $100 billion. Only one earlier phrase, limited to the interest-rate complex, makes the smaller series a subtotal. Nobody on the Q3 call says the $30 billion is rates-only. Duffy's "just $40 billion not too long ago" also fails: the February call already had a company-wide average of $80 billion.
Use over $100 billion a day as the Q3 company-wide average. Treat $30 billion as the unlabeled continuation of the rates series, not the firmwide total. The labeling has not been fixed. That is a management-quality flag on a KPI they lead with, not a speech slip.
April 22, Duffy: perpetuals "are against the law in the United States of America" and "are not allowed under the Commodity Exchange Act of 2000." July 22, McCourt quantified a live Kalshi Bitcoin perpetual (about $270 million a day versus CME crypto at $4.5–6.5 billion). The same call, Duffy: "our litigation will show that these are swaps, not futures," and CME could list them in existing clearing. October 21, general counsel John Marchese: a motion to dismiss has been filed, CME has opposed it, and the merits should be fully before the judge in December. Duffy would not say whether CME lists perps if they are deemed futures.
A present-tense ban does not coexist with a live U.S. contract, a quantified open interest, and a December merits brief. Do not model a statutory ban. Model an open swaps-versus-futures case, CME confident, and no public decision on listing. The April sentence is still on the record and was not walked back. The statute name in that sentence is also wrong. That error does not make the ban compatible with the later calls.
February 4: if the federal government calls sports event contracts swaps, "we will list them" and "we will participate." July 22: sports prediction markets are gambling and "not something that we want to be a part of participating in right now." The same July call reports about 525 million event contracts since launch, of which about 48 million (about 9%) are market events, and says sports growth is welcome. October 21 does not repair it. An analyst's premise that CME had stepped away from sports betting was left standing. FutureSports indices, a statistical team-performance future, are a different instrument and do not reconcile the two policies.
No P&L number breaks. Treat sports event contracts as listed and as the bulk of the reported event-contract count through Q2. Do not describe CME as having stepped away. Q3 gave no new event-contract mix. The retail narrative depends on which sentence is live, and the two sentences cannot both be the policy.
No competing exchange was named. ICE, Nasdaq, Tradeweb, Cboe-as-rival, Coinbase, Kalshi, Polymarket, DraftKings, FanDuel, and Robinhood were not on this call. Do not assign the comments below to those tickers. The rates read-through is the one that fights the volume print.
Gross investment income was $1,418.1 million, −8.4% versus $1,548.7 million, while interest-rate ADV rose +13.0%. Earnings from cash performance bonds were $1,391.7 million, −8.1%. That gross number is not what CME keeps. Other non-operating expense was −$1,264.0 million versus −$1,396.6 million. The residual of those two lines is roughly flat, about $154 million versus about $152 million (arithmetic). Lynne's own spread was "about 33 basis points," similar to Q2, on average clearing cash of about $147 billion versus about $150 billion, plus about $27 million on higher corporate cash (call). Period-end cash performance bonds and guaranty fund contributions were $155,179.7 million, +4.1% versus $149,041.2 million. Balances are higher than a year ago. The gross yield on that cash is lower. She explained the sequential balance dip. She did not explain the year-over-year decline in gross investment income. Do not invent a funds-rate path to fill that gap. Total non-operating income was $213.4 million, +4.4%, helped by the joint-venture line.
| Who | Relationship | What was said | Read-through |
|---|---|---|---|
| S&P Global | JV partner. CME is on the licensor side of the S&P index license. | 25-year Cboe extension. CME benefits from escalations starting in 2027, not 2032. The dollar amount is not yet disclosed. Equity in earnings of unconsolidated subsidiaries was $102.8 million, +6.7% versus $96.3 million. | Better for S&P Global and for CME's JV income starting in 2027, not this quarter. Do not add a 2027 dollar. The +6.7% is the pre-step run rate, and the line is not disclosed as 100% that venture. |
| Cboe | Licensee on the S&P complex. Not discussed as a competing exchange. | The escalation CME receives is an escalation Cboe pays, starting in 2027 rather than 2032. Separately, CME's own license fees rose because equity ADV was +22.9%. | Worse for Cboe from 2027 if the undisclosed schedule is an increase, which Duffy said it is. Neutral for Cboe's competitive position against CME. Do not net CME's higher license bill against the 2027 venture income. |
| Google Cloud | Technology supplier for the cloud migration. | Dallas disaster-recovery site opened for client testing in June. Market migration begins late 2027. Livestock relocation to Dallas is no longer necessary. | Better for Google as a multi-year infrastructure win. Not a 2026 revenue item for either company. The go-live slipped, and the livestock move was cancelled rather than pulled forward. |
| AI-infrastructure lenders and operators | Prospective Compute customers. None named. | Banks' exposure to AI infrastructure was called an unhedgeable risk. Compute "sits so nicely" next to the natural-gas franchise. Energy ADV of 2,665 thousand, +16.1%, was not attributed to hyperscalers. | Better, later, for lenders who cannot hedge that exposure today. Not in Q3 energy ADV. The gate is the CFTC, not demand. Do not assign this to a named neocloud, hyperscaler, or chip ticker. |
| Commercial energy users | Customers with a veto on the trading calendar. Unnamed. | They did not want to staff a weekend desk. CME withdrew the 24/7 10-barrel WTI filing and will wait. International energy volume was described as up 40%. | Worse for a near-term 24/7 crude listing. Not a statement that energy volumes are weak. The 40% geographic cut is management. |
| FICC and OCC | Cross-margin partners. | About 24 FICC clearing members registered. September FICC savings about $2.3 billion a day (call). OCC cross-margin a little over $5 billion in the quarter, inside equity savings of just under $47 billion. | Better for both rails as the Treasury mandate approaches. The dollars are margin saved by members, not revenue. Most equity savings are still CME's own offset. Registered is not live. |
| U.S. banks, as funding infrastructure | The reason primary contracts cannot go 24/7. No bank named. | Duffy: nobody can prefund a $76 trillion U.S. equity market on a Friday. The figure is his, not a filing. | A constraint, not a credit comment. Worse for anyone modeling a near-term 24/7 S&P or Treasury future. Gold, silver, and crypto can run the weekend because they are small. |
There was no CPI, wage, tariff, or consumer-spending comment. "Retail" on this call means futures traders. 94% of the business was described as institutional, and retail "does not carry much open interest." Acquisition can keep growing and still not be what holds the book. OSTTRA proceeds have been returned through the repurchase program. That cash is no longer dry powder. No acquisition target was named. BrokerTec Treasury Link is CME pulling its own cash pool onto Globex. Testing interest is not revenue, and the economic model was not given a number.