CME Group

IThe business

CME Group sells the right to trade, and then guarantees the trade. Its exchanges list futures and options: contracts to buy or sell something at a fixed price on a later date. Farmers use them to lock in a corn price, airlines to lock in fuel, pension funds to cover interest rates, and traders to bet on any of it. CME is the home of the contracts on US Treasuries and short-term interest rates, the S&P 500 and Nasdaq-100 indexes, crude oil and natural gas, gold, wheat and livestock.² Its clearing house sits in the middle of every trade, becoming the buyer to every seller and the seller to every buyer, so nobody has to trust the firm on the other side.² CME reports one segment. It had about 3,875 employees at the end of 2025, which is $1.7M of revenue and $1.1M of operating income for each one C (calculated).²

It earned $6,520.6M of revenue in 2025: $5,281.1M of clearing and transaction fees, $803.1M of market data and $436.4M of other revenue, mostly connection fees. Operating income was $4,229.5M, a 64.9% margin. In the first nine months of 2026 revenue was up 9% and the company has said it was the best half-year in its history.²³¹⁶

Volume times rate

Almost everything follows from two numbers. The first is how many contracts trade in a day, called average daily volume or ADV. The second is what CME earns on each one, the rate per contract. Multiply them by the number of trading days and you have most of the fee revenue. In 2025 CME traded 7,060.4 million contracts, for $4,913.0M of fees, at an average of $0.696 a contract. Volume rose 6% that year and the rate rose 1%.² Contracts grew 6.7% a year over the nine years from 2016, from 15.6 million a day to 28.1 million. The rate fell from $0.754 to $0.696, or 0.9% a year C (calculated).⁶²

The rate falls for a reason that matters later. CME charges less for some contracts than others. An interest-rate contract pays about 49 cents. An energy contract pays $1.15 and a metals contract $1.33. Members of the exchange pay less than non-members, electronic trades pay differently from trades done in the pit, and big users earn volume discounts.³² In 2025, 85% of contracts were traded by members. So when volume grows fastest in the cheapest product, and in small “micro” contracts, revenue grows more slowly than contracts. In the third quarter of 2026 interest-rate contracts were 51% of contracts and 37% of fees C (calculated).³²

Why interest rates run the business

Interest-rate contracts are the biggest line, and they respond to how nervous people are about rates. When the Federal Reserve held its rate near zero in 2020, interest-rate volume fell 22%, from 10.3 million a day to 8.0 million, in a year when equity-index volume rose 63%.¹⁷ It rose to 12.5 million in 2023 and 14.2 million in 2025 as the Fed raised and then began to cut rates.¹⁹² Through 2026 it has run higher: 18.7 million a day in the first quarter and 15.1 million in the third.³ The Fed left its target at 3.75% to 4.00% on 28 October by a vote of 10 to 2, with two members wanting a rise, and the ten-year Treasury yield was 4.88% in Q4 2026 R (reported).²⁰²¹ Treasury futures and options alone averaged 9.4 million contracts a day in the third quarter, up 25% on a year earlier C (calculated), which CME’s filing ties to the Fed’s first rate rise in over three years.⁴ CME’s own annual report says its volume tends to rise in periods of economic and geopolitical uncertainty.² I read that as: the product is insurance, and people buy more when they are afraid.

Clearing, and the cash it holds

The clearing house holds cash and securities from the firms that trade, called performance bonds, to make sure each can pay. At 30 September 2026 that was $155.2B, shown on the balance sheet as both an asset and a liability because it is not CME’s money.³ CME invests the cash and passes most of the interest back to the firms. It kept $411M of it in 2025 and $368M in the first nine months of 2026 C (calculated).²⁴ That income depends on how much cash is posted and on interest rates, not on how many contracts trade in a day, and it was $54M in 2021 when rates were near zero.²² I treat it separately in my valuation.

Data, cash markets and the stakes in other companies

CME sells its prices. Market data revenue was $803.1M in 2025, up 13%, and $243M in the third quarter of 2026, up 20%, a record.²³ The 10-Q credits higher use of some products, demand from new kinds of customer and price increases for some products.⁴ About 53% of the 2025 revenue came from outside the US, and about 30% of it passed through two resellers, 25% in the first nine months of 2026.²⁴ It also runs two electronic markets for cash trading, BrokerTec for government bonds and repo and EBS for currencies, which earned $283.7M of fees in 2025 and $224.5M in the first nine months of 2026, up 4%.²⁴ In the third quarter BrokerTec’s US Treasury trading averaged $96.4B a day, up 15%, and US repo $396.0B, up 4%.⁴ CME bought them with NEX in 2018 for an equity value of $5.6B, paid with 16.9 million CME shares and $2.5B of cash, and took on $1.0B of NEX debt as well.⁶ The filings do not say what those businesses earn.

It also owns 27% of S&P Dow Jones Indices, carried at $1.4B, the index company behind the S&P 500; 51% of a joint venture with FanDuel, the sports-betting company, which runs a prediction-markets app; and 33% stakes in two smaller exchange and broker ventures.² Together these are booked as equity earnings, $371.7M in 2025 and $302.9M in the first nine months of 2026. The filings do not say how much of that is the index company.²³ In October 2025 CME and S&P Global sold OSTTRA, a post-trade services business each held half of, to funds run by KKR, and CME booked a $306.1M gain.²

The new fronts

The 10-K lists the places management wants growth. Micro contracts, a smaller size, grew 32% in 2025, and the 10-Q puts the 2026 fall in the rate per contract largely down to them and to more trading by members.²⁴ Credit, Treasury-bill and mortgage futures are newer lines, and options volume is growing fast. The second-quarter release lists one-ounce gold contracts that trade 24 hours a day, seven days a week, among CME’s newer tools. In the third quarter bitcoin futures and options averaged 103,000 contracts a day, up 21%, and ether 72,000, down 70%.²¹⁶⁴ The joint venture with FanDuel launched in December 2025. Flutter, FanDuel’s owner, reported 66,000 average monthly players in the third quarter, $119M of volume and revenue it called not material, and said all sports contracts had moved to Crypto.com.²⁴ I could not find how much of the app’s trading still runs on CME.

The one that changes how CME is positioned is Treasury clearing. The Securities and Exchange Commission has set dates by which US Treasury trades must be cleared through a central clearing house: 31 December 2026 for cash Treasuries and 30 June 2027 for repo. CME’s new securities clearing arm was approved in December 2025, and the 10-K says CME expects to launch it later in 2026.² I found no launch date or revenue target in the filings. I come back to it in the next section.

The latest quarter

The third quarter of 2026 was strong. Revenue was $1,737.3M, up 13%. Operating income was $1,141.4M, up 17%, and earnings per share were $2.89 against $2.49. Adjusted for items the company sets aside, earnings per share were $3.02, and the adjusted operating margin was 69.4%, which is adjusted operating income of $1,205.5M over revenue C (calculated).³ Volume was 29.4 million contracts a day, a record for a third quarter, with growth in all six asset classes. The rate per contract fell to $0.679 from $0.702 a year earlier, which the 10-Q puts down to more micro contracts and more trading by members, partly offset by a fee increase that took effect on 1 April 2026.³⁴ Expenses rose 5.5% to $595.9M. CME returned nearly $1.0B in the quarter: $465M of regular dividends and $524M of buybacks.³ A wire copy of CME’s monthly release says October volume was another record for the month, 30 million a day, up 15%; I could not compare it with CME’s own page.⁵

IIThe moat

CME’s protection is not mainly a law. It is a habit that is costly to break, plus a handful of index licences nobody else can get. This is what makes it hard to dislodge, and it is also where the weak spots are.

One pool of trades, and one clearing house

A futures market is good when both buyers and sellers are always there. The 10-K lists the depth of its markets and the efficiency of its clearing among the reasons customers pick an exchange.² Once most of the world’s Treasury, SOFR and S&P 500 futures sit in one place, anyone who wants to hedge goes there, and that is the reason others go. A new venue has to give traders a cheaper or better market and persuade them to leave a deep one.

The clearing house adds a second lock. It nets a firm’s positions across six asset classes and calls for margin, the cash that backs a position, on the net risk. CME says its offsets saved clients an average of $72 billion of margin a day in 2025, more than $85 billion in the first quarter of 2026, more than $95 billion in the second and $100 billion in the third.²²⁵¹⁶³ A client who moves part of the book elsewhere gives up some of that saving. CME also cross-margins with the Fixed Income Clearing Corporation (FICC) and with the Options Clearing Corporation, and the 10-K says it has widened the FICC link to more products.² That is why a rival with a cheaper fee has not been enough.

Behind the guarantee sits real capital. In a default CME uses the defaulting firm’s own margin first, then $100M of its own money for futures and options (and $150M for interest-rate swaps), then the other firms’ guaranty fund of $8.3B, and then calls on those firms for up to 275% of their contributions in a single default. At the end of 2025 clearing firms had $347.3B of margin on deposit.² CME’s own $100M is a small part of the stack; its job is to be a credible counterparty.

The licences

CME does not own the S&P 500, the Dow Jones average, the Nasdaq-100 or the Russell indexes. It licenses them and pays the owners a share of revenue. The expense line called licensing and other fee agreements, which also holds broker rebates and revenue sharing, was $371.0M in 2025, 5.7% of revenue C (calculated).² The terms are what make it hard to copy. The S&P licence for futures and options is exclusive for as long as CME owns at least 5% of S&P Dow Jones Indices, and it then stays on, non-exclusively, for one more year. CME owns 27%.² The licence runs both ways: S&P Global says the index company earned $193M in 2025 and $154M in the first nine months of 2026 from its licence with CME, a share of the profits on CME’s equity-index products, and CME gets 27% of the index company’s income back.²⁶²⁷ The Dow Jones licence renews itself every five years while contracts are open, the Nasdaq licences are exclusive to 2039 and the Russell licences to 2037.² The 10-K is careful to say that CME cannot guarantee it will keep the exclusivity, and that others may try to create similar products.² I read the S&P exclusivity as lasting as long as CME keeps at least 5% of the index company, and CME has built its largest equity product on it.

The regulators

The Commodity Futures Trading Commission (CFTC) regulates CME’s exchanges and clearing house, and the Financial Stability Oversight Council has named CME a systemically important financial market utility. A clearing house must hold capital equal to at least one year of operating expenses and liquid resources equal to six months of them.² That is a cost and a barrier. I treat CME’s $2.3B of cash and securities as mostly spoken for, and I come back to that in the valuation.

Where it is thin

There is no law that makes Treasury futures or crude oil trade at CME. Rivals exist and the 10-K names them. What is new in 2026 is a different kind of rival: venues that do not look like exchanges. The first is prediction markets. Kalshi raised $1B at a $40B valuation on 29 October and says it holds over 75% of prediction-market volume, claims I could not check.²⁸ The second is perpetual futures, contracts with no expiry date. On 29 May 2026 the CFTC approved a bitcoin perpetual contract on Kalshi as a futures contract. CME sued the CFTC on 18 June, arguing such products are swaps, and on 2 September the CFTC moved to dismiss the case for lack of injury, saying CME could list the same product itself.¹⁵ CME’s opposition was due on 2 October; I could not find the later court schedule. I can’t say who is right. What I can say is that the answer decides whether a regulated trader can get perpetual crypto exposure somewhere other than CME.

The third is the Treasury market itself. FICC is the established clearer. ICE’s Treasury clearing went live on 3 February 2026 and plans to clear repo from the fourth quarter.²⁹ FMX, owned by BGC, trades Treasury futures cleared at LCH, and BGC says it traded 112,000 contracts a day in September and more than 80,000 in the third quarter, against CME’s 9.4 million Treasury futures and options a day in that quarter. That is about 1% and the products are not identical, but BGC says FMX has grown nearly seven times in a year and holds 42% of cash Treasury central-limit-order-book trading.³⁰⁴ UBS said in a note reported on 22 October that the competitive pressure that had weighed on the stock earlier this year appeared to have faded.³¹

IIIThe competitive advantage

The last section says where CME is protected. This section is about what it does better than the people it competes with, and whether that is growing.

Costs that don’t move with volume

The 10-K says the majority of CME’s expenses do not vary directly with contract volume. Licensing fees do vary, with equity, energy and swap volume.² So extra volume mostly falls through to profit. In 2025 revenue rose 6% and expenses 4%, so the operating margin went from 64.1% to 64.9%.² In the third quarter of 2026 the adjusted margin was 69.4% C (calculated).³ It has not been a straight line. The margin was 53.2% in 2019 and 54.0% in 2020, after the NEX deal pushed the yearly write-off of purchased intangibles to $314.7M and $311.2M C (calculated).¹⁷⁸ By 2025 that charge was $223.4M.²

It can raise prices

CME put through a fee increase on 1 February 2025, which helped lift the rate per contract by 1% that year.² The 10-Q says another increase in certain clearing and transaction fees took effect on 1 April 2026, and CME told clients about further pricing changes on 2 February and 11 August 2026 in notices it furnished to the SEC. The filings do not give the size of any of these.⁴⁹¹⁰ It has not been enough to lift the rate per contract, which fell to $0.679 in the third quarter from $0.702 because micro contracts and member trading grew faster.³⁴ Market data is the cleaner example. Revenue rose from $406.5M in 2016 to $803.1M in 2025, 7.9% a year, and 20% in the third quarter of 2026. The 10-K credits the 2025 rise largely to price increases for certain products and more use of others.⁶² I treat the market data line as the best evidence of pricing power, because a customer who needs the live benchmark price mostly has nowhere else to go, though the 10-K does list vendors whose reference prices are an alternative.

Technology

CME signed a ten-year partnership with Google Cloud in 2021. Applications for clearing and market data have moved to the cloud, and the 10-K says CME was the first derivatives marketplace to offer live market data natively on Google Cloud.² Google Cloud is building a private cloud region and a co-location facility in Aurora, Illinois, where trading will eventually run. The 10-K says CME has advanced its preparations to move its markets there, and in the third-quarter release Mr Duffy said CME had made strong progress and was the only exchange offering clients a choice of co-location and cloud-managed services. I found no date for the move in the filings.²³ In November 2021 CME also sold Google LLC non-voting convertible preferred stock for about $1B in cash, which converted into Class A common stock on 5 March 2026; the Series G designation was eliminated on 24 September 2026.³⁴⁴³⁵ The cloud is a bet, and it has a cost. Technology expense rose 16% in the first nine months of 2026 to $240.8M, and the 10-K lists the cost of running old and new systems together as a risk.⁴²

Evidence the advantage is widening

• Market data revenue rose 20% in the third quarter and grew every year from 2018 to 2025.³⁸
• Daily margin savings rose from $72 billion in 2025 to $100 billion in the third quarter of 2026.²³
• Interest-rate volume set a third annual record in 2025, and Treasury futures and options rose 25% in the third quarter of 2026.²⁴
• Volume outside the US was a third-quarter record of 9 million contracts a day, and 31% of electronic volume came from outside the US in 2025.³²
• Micro products grew 32% in 2025, though each earns a lower rate.²⁴

Evidence it is narrowing

• The rate per contract is down 8% since 2016, from $0.754 to $0.696, and $0.679 in the third quarter.⁶²³
• FMX’s futures volume rose nearly seven-fold in a year and its share of cash Treasury trading reached 42%.³⁰
• The 2026 proxy lists competitive pressure in energy markets among the reasons the 2025 bonus target was set where it was.¹²
• Prediction markets and perpetual contracts are growing outside CME, and CME is in court over the second.²⁸¹⁵
• Technology spending is up as old and new systems run together.⁴

Where CME wins and where it loses

CME wins where a client needs the deepest market to hedge a large position: Treasuries, SOFR, S&P 500 futures, WTI crude. It also wins where the client wants one clearing house and one margin number. It is weaker with small retail traders who can use a broker’s own product, and in markets where a rival’s fee is much lower and the client cares only about cost. The rate per contract tells you which way those forces have netted out. Over nine years it has fallen, and volume has more than made up for it.

IVManagement

Who runs it

Terry Duffy has been a CME member since 1981, ran his own trading firm, TDA Trading, until 2002, and has been chairman of the board since 2002 and chief executive since 2016.¹² On 17 June 2026 the board said he will become Executive Chairman, and Lynne Fitzpatrick, 48, who joined in 2006 and has been chief financial officer since 2023 and president since November 2024, will become chief executive on the later of 1 March 2027 and the date CME files its 2026 annual report.¹¹ Mr Duffy stays as Executive Chairman to 31 December 2027. Jack Tobin, 63, the chief accounting officer since 2015, becomes chief financial officer in March 2027, having planned to retire on 30 October 2026, and CME hired Matthew Render from Citadel and will make him chief accounting officer, and promoted John Marchese to general counsel.³⁶ Charlie Carey is lead director.¹¹ The team is long-serving and mostly home-grown. That is a strength for continuity. It is also a company handing over its top job to its finance chief at the moment its volume is at a record.

How they are paid, and what that pays them to do

Mr Duffy’s pay in 2025 was $23.4M: $2.0M of salary, $13.3M of stock awards, $7.1M of annual bonus and $0.9M of other items.¹² Ms Fitzpatrick’s agreement for the chief executive role sets a $1.2M salary, a bonus opportunity of 200% of salary and long-term equity of 700% of salary, about $12M a year at target, and two years’ salary as severance if she is let go without cause.¹¹

The annual bonus pays on one number, cash earnings. It is net income, plus depreciation, plus stock pay, plus amortization of purchased intangibles after tax, less capital spending, adjusted for net interest expense.¹² The same measure sets the dividend: the regular dividend is aimed at 50% to 60% of the prior year’s cash earnings.²¹² Three things follow. First, stock pay is added back, so the bonus does not charge for it; I do charge for it. Second, net income includes interest on clearing-firm cash, which moves with the Federal Reserve and the size of the pool, not with management’s work. Third, there is no charge for capital, but CME needs very little: capital spending was $83.5M in 2025.

In 2025 cash earnings of $4.238B were 114.3% of target, close to the 120% cap, and the bonus paid 171.62% of target. In 2024 the figures were 117.3% and 186.32%.¹³¹² The proxy explains the 2025 target by pointing to hard comparisons with the year before, tariffs, the cost of moving to the cloud, expected lower collateral income and competitive pressure in energy.¹² Those are real. But the bar was set 4.1% below what the company had just earned, and the company then grew cash earnings 9.7% C (calculated). I read that as a goal that rewards continuing to do what the company was already doing.

The long-term awards are half restricted stock vesting over four years and half performance shares over three. Half of the performance shares pay on CME’s share return against the S&P 500 and half on net income margin, net income as a share of revenue.¹² Because net income includes interest on collateral and equity earnings, a rise in rates or in S&P Dow Jones Indices’ profits lifts the margin without any change in the exchange. That is my reading of the definition, not a statement by the board.

How much of their own money is in

The chief executive must own $10M of stock and other executives $1.575M.¹² Mr Duffy owned 55,594 shares on 16 March, worth $16.9M at $304.52 C (calculated), and after his September grant the latest Form 4 shows 62,442.¹²¹⁴ That is under one year of his pay. All 30 directors and executive officers together owned 912,542 shares, 0.3%.¹² The Forms 4 show a different pattern from the guideline. In 2026, 11 officers and directors sold 88,431 shares for $26.4M, and none bought on the market; the only purchases were small ones by one director through a broker’s dividend reinvestment plan, as his Form 4 says.¹⁴ Mr Duffy sold 35,000 shares on 15 May at about $298 and 15,000 on 10 September at about $274, $14.6M in all, after selling 25,000 in December 2025. The filings I read do not say whether the sales were made under a trading plan.¹⁴ Selling after a grant of stock is normal, and the sales are about 0.02% of the company. But 75,000 shares between December 2025 and September 2026, and no buying anywhere, says the people closest to the business are taking some of the money out.

Who else owns it

Vanguard held 36.2 million Class A shares, 10.0%, and BlackRock 28.9 million, 8.0%, according to the 2026 proxy statement.¹² Alphabet Inc.’s Form 13F for the quarter ended 30 September 2026, filed on 6 November, lists 3,484,020 CME Group common shares, which Alphabet valued at $912.7M, about 1% of the Class A shares outstanding. The count is the same in its 31 March and 30 June filings.^3738394 It is 1.1 million fewer than the 4,584,020 Series G preferred shares Google bought in 2021, which converted one for one into Class A stock on 5 March 2026. The March filing already shows the lower number, so the gap opened in March or earlier, and I found no filing that explains it. I report the 13F count and nothing more.³⁴⁴ A small number of Class B shares, 3,138 in all, are tied to trading rights and elect some directors; Class B holders have sued CME, and a jury found for the company in July 2025, post-trial motions were denied on 15 June 2026 and the plaintiffs have appealed.²⁴

What they did with capital over a full cycle

From 2016 to 2025 CME produced $26.2B of owner cash and paid $25.3B of dividends, 97% of it C (calculated).⁸² It bought no stock until 2025, when it spent $266M in the last two months of the year.²²³ The share count did not fall. Diluted shares went from 339.0M to 360.3M, mostly from the 16.9M shares issued for NEX C (calculated).⁶⁸ That purchase cost an equity value of $5.6B in 2018 and the cash markets business it brought earned $283.7M of fees in 2025. The other part of NEX, the post-trade business OSTTRA, was sold in 2025 at a $306.1M gain on a $1.2B book value.² The filings give no profit for either business, so I can’t say what the return on NEX has been. I can say it was the one large purchase of the decade, and that the owners have since been paid out of the exchange.

The buyback is new. After $266M in 2025, CME spent $1.76B in the first nine months of 2026, 6.43 million shares at an average of $272.89, and $1.0B of the $3.0B authorization is left.⁴ The third quarter’s 2.05 million shares cost $255.81 each, and the second quarter’s buybacks came to $695M.⁴¹⁶ The $1.6B CME received for its OSTTRA stake arrived in 2025, so these buybacks drew on that cash.² At my base value of $234, $272.89 buys 86 cents of value for each dollar. The 10-K says the annual variable dividend is set from the prior year’s results and expected cash needs, after other returns of capital including buybacks.² So a dollar spent on buybacks takes the place of a dollar of next year’s variable dividend. It does not add to what owners receive; it changes the form.

VFinancials

Ten years show a business whose revenue grew 6.8% a year, whose profit grew faster, and whose cash went almost entirely to its owners.

What an owner could really take out

Over ten years CME reported $27.9B of net income and $29.6B of operating income. The cash an owner could have taken, after capital spending and after stock pay, was $26.2B, or 94% of net income C (calculated). Take out the $1.5B tax credit of 2017 and the match is close to 100%. In 2025 owner cash was $4.10B against net income of $4.07B, and in the twelve months to September 2026 it was $4.23B against $4.42B of net income C (calculated).⁸²⁴ The reason the two are so close is that CME needs almost no capital. It spent $83.5M in 2025, which is 1.3% of revenue and less than its depreciation of $107.5M.²

Owner cash grew 11.2% a year from 2016, from $1.57B to $4.10B C (calculated). Revenue grew 6.8%. The gap has two parts I can name. The operating margin rose from 61.2% to 64.9%. And net interest on clearing-firm cash, which was $54M in 2021, added $411M in 2025 C (calculated).²²² I would not extend 11% a year into the future. Part of the gap is rates, and rates will not rise forever.

Where the cash goes

The company’s dividend policy has two parts. A regular quarterly dividend targets 50% to 60% of the prior year’s cash earnings. On top of that the board declares an annual variable dividend, set from the prior year’s results and CME’s expected cash needs.² For 2025 results the board declared $6.15 a share on 12 February 2026, paid 26 March, and the regular dividend is $1.30 a quarter. That is $11.35 a share for 2026 if the regular dividend holds, a yield of 3.7% at $304.52 C (calculated).²⁴ The company says it has paid out nearly $30B since the variable dividend began.⁴¹

In 2026 the payout has gone beyond what the business earns. In nine months CME paid $3.63B of dividends and spent $1.76B on buybacks, $5.39B in all, against $3.16B of owner cash C (calculated).⁴ Cash fell from $4.4B to $2.2B, because the payout drew on the OSTTRA proceeds received in 2025. The nine-month figure overstates the pace, since it holds the whole year’s variable dividend of about $2.2B but only three quarters of cash earned. Over the twelve months to September 2026 CME paid $4.09B of dividends and spent $2.02B on buybacks, $6.10B in all, against $4.23B of owner cash, 144% C (calculated). That is still more than the business earned, and it can’t repeat without a smaller cash pile or more debt, so I do not read the 2026 payout as a new run rate.⁴²

Quality of earnings

• Interest on clearing-firm cash. CME kept $411M of it in 2025, 8% of pre-tax income C (calculated). It depends on balances and rates, not volume, and it is the one part of earnings I would not expect to grow with the business.²
• One-time gains. 2025 net income includes a $306.1M gain on OSTTRA and $343.7M of investment gains in all, which the company’s adjusted figures take out. I take the gains out of my base.²⁴¹
• Stock pay. $94.8M in 2025, added back in the cash flow statement; I deduct it.²
• Amortization. $223.4M a year of purchased-intangible amortization is not a cash cost. In my reading it also does not cut cash taxes, because the $5.2B deferred tax liability suggests the assets came from share deals, so I tax earnings before the charge.²³
• Equity earnings. $302.9M in nine months of 2026. CME does not split it, but S&P Global, which controls the index company, reports $322M of Indices income attributable to the noncontrolling holder, CME, in 2025, up from $284M in 2024 and $241M in 2023. That is most of CME’s $371.7M of equity earnings in 2025 C (calculated), and OSTTRA, until its sale, and smaller ventures made up the rest. CME’s cash flow statement makes no deduction for these earnings, which suggests they arrive in cash, but CME does not say so. I keep the line to slow growth in my cases.³²²⁶
• Taxes. The effective rate was 23.6% in 2025, up from 22.4% on state apportionment, and 22.9% in nine months of 2026. I use 23%.²⁴
• Receivables. Up 23% to $788.1M since December while nine-month revenue rose 9%. I did not find the reason.³

One cross-check on the joint-venture line. S&P Global carries CME’s 27% of the index company at its estimated redemption value, $4.9B at the end of 2025 and $5.3B at 30 September 2026, because CME may sell at least 20% of its share to S&P Global at fair value. CME’s own book value was $1.4B at the end of 2025.²⁶²⁷² That redemption value is before tax. A sale would be taxed at 23% on the gain over book value, which leaves about $4.4B C (calculated). My base case values all the joint-venture income, not only this stake, at about $5.8B after tax E (estimate), so on this line I am not being cautious.

The balance sheet

At 30 September 2026 CME had $2.2B of cash, $0.13B of marketable securities and $3.4B of debt, all fixed-rate, with a $2.3B credit line undrawn. Ratings are AA- at S&P and Aa3 at Moody’s, both stable.³⁴ Debt is 0.7 times the last twelve months’ operating income before amortization C (calculated). That is low, and CME could borrow more. The balance sheet also carries $155.2B of performance bonds as both asset and liability, and CME’s own first-loss money in the clearing house is $100M for futures and $150M for swaps.³² I give no credit for the cash. A clearing house must hold capital equal to a year of operating expenses, about $2.4B, and CME’s cash and securities of $2.3B are close to that.²⁴ So I count the debt in full and the cash as working capital E (estimate).

VIValuation

First, how this goes badly wrong

Before the numbers I asked what would most likely hurt an owner here, and worked back. Three things stand out.

Volume falls back from a record and does not return. Fees depend on how many contracts trade, and 2026 is on course for the most ever, 31 million a day, up from 28 million. If 2027 volume is 26.5 million, the 2024 level, and grows from there, my base falls from $234 to $211. Each lasting million contracts a day is worth about $6.5 a share E (estimate).⁴² There is a precedent for a quiet year: with the Federal Reserve at zero in 2020, interest-rate volume fell 22%.¹⁷ Now the Fed has held its range at 3.75% to 4.00% with two members wanting a rise, and the ten-year yield is 4.88%.²⁰²¹ When markets calm down, this business gets quieter.

The price per contract keeps sliding. CME earned $0.754 a contract in 2016 and $0.679 in the third quarter. The rate fell 0.9% a year over the nine years to 2025, rose 2.7% a year from 2022 to 2025 on fee increases, and fell 3% in the third quarter from a year earlier. My base has it rising 0.5% a year, which leans on the recent record, and that is a judgement the sensitivity table shows. If it falls 1% a year from here, not rises 0.5%, my base falls to $208, about $17 a share for every point a year E (estimate). FMX, brokers that route flow to their own venues, and venues for event and perpetual contracts all compete on price. CME has been raising fees, most recently on 1 April 2026, and the rate per contract still fell 3% in the third quarter from a year earlier, because the growth was in cheaper contracts.⁶³⁴

The clearing house fails, or its position is changed by rule. CME’s own money in a default is $100M for futures and $150M for swaps against $347.3B of margin on deposit.² A failure of a large clearing firm would be a real test; CME’s release on Mr Duffy’s move lists the 2008 crisis and MF Global among those it has already come through.¹¹ A change in law, such as a transaction tax or an end to the 60/40 tax treatment of futures that CME lists as a risk, would hit volume directly.² I can’t put a probability on either, and I don’t put one in the cases. They are why the bear case is a lasting lower level and not a collapse.

None of these needs anything exotic, and the first two are already visible in the numbers. They are why I value cash that reaches an owner and not earnings, and why I show a range.

What an owner could really take out

I value the cash after stock pay and after the capital needed just to stand still. For CME that capital is small, but there are two charges that reported earnings leave out. The first is stock pay: $94.8M in 2025, which is already inside operating income and which I do not add back, as the company’s bonus measure does. The second is the capital regulators require. A clearing house must hold capital equal to a year of operating expenses, so each year costs rise, CME has to keep that much more and cannot pay it out. I charge that growth, $84M in 2026 and $94M in 2030.² I treat the $2.3B of cash and securities as the base of that requirement and give no credit for it.

I also split owner cash into three streams, because they behave differently. The exchange (operating income before the non-cash amortization) depends on volume and price. The interest CME keeps on clearing-firm cash, $486M this year, depends on rates and balances and not on volume. The joint-venture earnings, $403M, depend on S&P Dow Jones Indices and others. I take out the $306.1M OSTTRA gain and the $343.7M of 2025 investment gains.²³ The interest on debt is not in the cash flows because I subtract the debt itself.

On this basis 2026 owner cash is about $4,486M, against net income of about $4,273M, so 105%. The difference is that I leave out the $174M of annual interest on debt and then deduct the debt, and I charge the growth in required capital. Over the last twelve months the cash from operations less capital spending and stock pay was $4,234M, within 6% of my 2026 figure, which is encouraging, since the two were built separately.

What the price implies

Put the other way, what does $304.52 need? With a 9.0% discount rate and 3.0% growth after fifteen years, the price needs owner cash to grow 6.8% a year for fifteen years, from $4,486M to $12.0B. Over the last nine years it grew 11.2% a year, but part of that was the rise in interest income. My base has 4.3%. Put in terms of volume, the price needs contracts a day to grow 7.2% a year from 2028, to 79 million by 2041, 2.5 times 2026’s record. Volume grew 6.7% a year in the nine years to 2025, so the price needs a little more than that for fifteen years.⁸²⁶ Or it needs a lower discount rate: my base is worth the price at 7.7%.

What I can’t judge

• Whether 2026 is a new level. Volume was 36.2 million a day in the first quarter and 29.4 million in the third. A wire copy of CME’s monthly release puts October at 30 million.³⁵ I can’t tell whether that is the new normal or a fearful year.
• How long rates keep moving. The Fed’s next steps decide interest-rate volume, half of all contracts, and I can’t forecast them.
• How long the index company keeps growing. By S&P Global’s figures it is most of the joint-venture earnings, and the income attributed to CME’s stake rose 18% in 2024 and 13% in 2025. I assume 4% a year and can’t judge more than that.²⁶
• Whether event and perpetual contracts help or hurt. They may bring customers to CME or train them elsewhere, and CME is in court over it.¹⁵
• What Treasury clearing earns. The 10-K says it launches later in 2026. I found no revenue target in CME’s filings.²
• What the cloud move costs, and whether it saves. Technology expense is up 16% so far in 2026 and the filings give no date for the move or a saving.⁴

Because I can’t know these, I show a range and the price’s position in it, and I don’t treat $234 as more exact than it is. A fair reading of my work is that the base is somewhere from $200 to $260.

The discount rate

I use 9.0%, the rate on my Mastercard and Visa pages, so the three can be compared.⁴²⁴³ The textbook method gives CME much less. With the ten-year Treasury yield at 4.88% and Damodaran’s implied equity risk premium of 4.20%, CME’s beta against the S&P 500 gives 6.2% on ten years of monthly returns, 5.7% on five and 4.7% on three C (calculated).^2144145 CME’s beta is low, which fits a business that gets busier when people are afraid, and that is a real property. I do not trust it enough to value the business at 6%, for two reasons. The cash flows are not low-risk at all: they fall if markets are calm, and they depend on regulators allowing the products. And the next best use of the dollar, the 10-year Treasury at 4.88%, is only 4.1 points below my rate. At 8% my base is $285 and at 10% $198. The verdict depends on the rate: the price is the base value at 7.7%.

Three cases

Bear

Volume returns to its 2024 level of 26.5 million contracts a day in 2027 and grows only 1.5% a year, because volatility fades and rivals take the new product lines. The rate per contract falls from $0.665 by 0.8% a year as FMX, brokers and platforms push prices down; data growth fades to 3%; CME keeps $300M on clearing cash. Costs are cut to grow 3% to 4%. Owner cash stays near $3.8B for fifteen years. Value $125.

Base

Volume falls back from 2026’s record to 30.0 million in 2027, about the pace since the first quarter and above 2025’s 28.1 million, and then grows 4.5% a year for five years, 4% for five and 3% for five. The rate per contract is $0.680 and rises 0.5% a year, because fee rises are partly offset by the cheaper rates and micro products that grow fastest. Market data grows 11% falling to 6%. Interest kept is $450M. Costs grow 5% falling to 3.5% as the cloud move ends. Owner cash grows from $4.5B to $8.4B by 2041. Value $234.

Bull

2026 turns out to be the new floor: 31.5 million in 2027, growing 6.5% a year as Treasury clearing, event contracts, round-the-clock trading and Asia add to a base that already grows faster than the economy. The rate per contract rises 1% a year, data grows 14% falling to 7%, and costs grow 3% once the cloud move ends. Owner cash reaches $12.2B by 2041. Value $331.

Other uses of the same dollar

The best comparison on this site is Mastercard, another toll on other people’s money, with Visa next. At today’s prices my Mastercard cash flows are worth the price at 8.4%, Visa’s at 8.2% and CME’s at 7.7%, against the 9.0% I ask of each.⁴²⁴³ By my numbers none of the three clears its bar, and CME misses by most: the price is 30% above my base, against 13% for Mastercard and 18% for Visa. CME offers more cash today and less growth. Its owner cash yield on enterprise value is 3.8%, against 2.7% and 3.2%, and the dividend alone is 3.7%. The Treasury pays 4.88% with no risk, and CME’s owner cash yield is lower than that, so a buyer is paying for growth that has to be real.²¹ The company’s own alternative is its buyback. At $272.89, the average price in the first nine months, I get 86 cents of value for each dollar spent, and at $304.52 I get 77 cents E (estimate).⁴

What the market paid before

The stock has traded between 22.7 and 34.0 times earnings at the end of each year since 2018, and at 25.7 times my 2026 estimate it is now in the lower part of that range.¹⁸ It was 34.0 times at the end of 2019, when earnings per share were $5.91, and 22.7 times at the end of 2022. Earnings per share have grown about 10% a year from 2016 to my 2026 estimate, and the price has risen 10.5% a year before dividends, so the multiple has not moved much over the decade. The price is not at an extreme for this stock. The question is whether growth stays at the rate a 25-times multiple assumes.

The margin of safety

I set it as I did for Mastercard and Visa: my base value less a quarter, which is $176. My $125 bear value is 29% below it. The price is 1.7 times that. The stock last closed below $176 on 10 March 2023, and its lowest close in the last year was $218.58.¹ I am not saying the price will fall to $176. I am saying that is the level at which I would be paying for the franchise as it is today and getting growth as a gift.

Two different questions

The first is how good the business is. By the numbers it is very good: a 65% operating margin, owner cash of 101% of earnings, almost no capital needed, an AA- credit rating, and volume and revenue at records while new rivals grow from a small base.²³³⁰ The second is how good the price is. At $304.52 the price is 30% above my base value and 8% below my bull, which needs 2026’s record volume to keep growing. Those are different answers. A great business can be a poor purchase at the wrong price. This looks like that case, and the market may be seeing something I am not: that CME’s cash is safer than I allow, and that a 7.7% return on it is enough.

VIIThe case for the price

What most people miss

Four facts cut in CME’s favour. First, the S&P 500 futures licence lasts as long as CME owns at least 5% of S&P Dow Jones Indices, and CME owns 27%, so it has no set end date.² Second, market data revenue grew every year from 2018 to 2025 and hit a record $243M in the third quarter of 2026; its price is set mainly by the number of devices, not by contracts traded.⁸³⁴ Third, CME says its margin offsets saved clients $100 billion a day in the third quarter, up from $72 billion for 2025.³² Fourth, S&P Global carries CME’s 27% of the index company at $5.3B before tax, against $1.4B on CME’s books at the end of 2025. My values already count that income.²⁷²

The bull case

I found no named investor with a checkable view on CME, so I give the case as the company and one broker make it. CME’s chairman said in the third-quarter release that clients came to CME in record numbers as uncertainty grew.³ In February he called it a “risk-always-on” environment.⁴¹ On 22 October UBS raised its price target to $310 from $270, in a note reported by Investing.com. It said CME was making progress on single-stock futures, compute contracts and Treasury Link and that competitive pressure had eased, but that regulatory uncertainty could still be a concern.³¹

While rates and politics stay unsettled, people keep hedging, and CME has the deepest markets, a margin saving that locks clients in and data that rises in price. If so, 30 million contracts a day is a floor, not a peak.

Against reality

If volume holds at 2026’s level and grows 4% to 5% a year, and the rate per contract holds, the price is possible. That is close to my bull case, $331, and the price is 8% under it. History helps: volume has not fallen by more than 1% in any year since 2016, and after the 18% jump of 2022 it rose 5%, 9% and 6% in the next three years.¹⁷¹⁹² It does not settle it: after the jump of 2018 volume was flat for two years, and my base has it fall from about 31 million to 30.0 million in 2027, more than any year in that record. To make my base worth the price I would need 2027 volume of 41 million a day, or 7.2% growth a year for fifteen years.

It helps to multiply it out. The price needs five things: volume at or above 2026’s level growing 4% to 5% (about 45%); a rate per contract of $0.68 or better (70%); double-digit market data growth for several more years (75%); no rival or ruling that takes a large slice of rates, stock indexes or crypto (85%); and costs growing 5% a year or less (85%). Multiplied, that is about 17%, one chance in six. They are guesses, not independent, and not used in the values E (estimate).

What a 10x would take. Ten times $304.52 is $3,045, or $1.09 trillion of market value on today’s shares. At 25 times earnings, about today’s multiple, that needs $43.7B of net income, 10.2 times this year’s estimate of $4.27B. At a 60% net margin that is $73B of revenue, 10.3 times 2026’s, or 26% growth a year for ten years, and about 321 million contracts a day at today’s rate if all of it came from volume. Revenue grew 6.8% a year from 2016 to 2025, and the price rose 2.7 times in the last ten years, before dividends. CME says its market value is up more than 8,000% since the 2002 listing. That took 24 years, and the starting base was far smaller.¹¹ A tenfold price in ten years is not plausible for a company worth over $100 billion, and I would say so about any exchange this big.

Where I come down

Three things moved me: volume held near a record (October is reported at 30 million a day); market data keeps growing; and Treasury trading is busy, with Treasury futures and options up 25% in the third quarter.⁴⁵ The price did not move me. It still needs about the bull case, and I won’t raise my base on a record year, because a record year is what a fearful market produces, and my base is about the years after it. What would raise my values is volume holding at 31 million a day or more through the first half of 2027 with the rate per contract at $0.68 or better, or fees from Treasury clearing and event contracts that are not in my cases. Until then the price sits between my base and my bull.

VIIIRisks, and what would change my view

The ways this goes badly wrong are at the start of Valuation. These are the rest.

• Taxes and rules. CME lists a tax on futures transactions and an end to the 60/40 tax treatment among its risks, and it is regulated by the CFTC, the Federal Reserve and others abroad.²
• Technology. An outage, a cyber attack or a failed migration would stop trading, and the cloud move puts more of CME’s systems with one supplier, Google Cloud.²
• One firm and two resellers. One clearing firm paid 12% of fees in 2025, and two resellers carried about 30% of data revenue. CME believes customers would move to other firms, and the 10-K says so.²
• The licences. The S&P term is tied to CME’s stake. The Nasdaq licences run to 2039 and Russell’s to 2037, and the 10-K says CME cannot promise renewal on terms as favourable.²
• Leadership. A chief executive who has led for ten years steps back in March 2027 and a new chief financial officer arrives with her.¹¹³⁶
• Litigation. The Class B suit is on appeal after CME won at trial. CME has booked no accrual.⁴
• Insiders. Eleven officers and directors sold stock in 2026 and none bought on the market. That alone says little, but the people closest to the business were net sellers in a record year.¹⁴
• Capital return. Buybacks of $1.76B in nine months came on top of dividends, and $1.0B of the authorization remains. The price paid, $273 on average, is above my base value.⁴