StoneX Group Inc. SNEX

Key data

The business

StoneX is a financial plumber. It sits between the real economy and the world's exchanges, taking the other side of an order or routing it, clearing it, financing it, and keeping a spread. A grain elevator in Iowa hedging next season's corn, a Brazilian sugar mill, a bank moving dollars into Nigeria, a retail speculator in currencies: each touches one of four arms. Commercial hedging and physical commodities is the core. Institutional execution and fixed income, the retail platforms, and a global payments business that wires money into corridors big banks find unprofitable round it out.

The reported revenue line, $132 billion last year, is a fiction of gross flow and tells you nothing. What matters is what the firm keeps after the cost of filling those orders, about $2.6 billion, and that is what compounds. The moat is unglamorous and real: clearing scale and a payments network into frontier corridors that would cost a rival years and licenses to rebuild. It earns four ways, commissions, spreads, the interest it keeps on billions of client cash parked on its books, and, increasingly, physical commodity margins from the businesses it keeps buying.

The numbers

Look at what the firm keeps, not what flows through it. The last five quarters, oldest to newest:

The June quarter is the inflection. Net revenue of $1.23 billion was up 89% on a year earlier and net income doubled it. That step-up is the O'Brien physical-commodities book landing on the income statement, and so far it has landed clean, no credit event and no counterparty blowup, which is the question that hung over the deal when it closed. The tell sits in the last column: earnings per share rose 23% while profit doubled, because the diluted share count rose by two-thirds to fund the purchase.

The five-year picture, with the year to date:

Money kept has gone from $562 million in fiscal 2021 to $2.58 billion in fiscal 2025, roughly 46% a year. Net income compounded near 27%. But diluted earnings per share grew only about 11% a year, from $2.55 to $3.93, and that gap is the whole story. The firm compounds beautifully at the top and middle and leaks a large share before it reaches the owner. Return on equity of 20% says the capital works hard. The 12 times earnings says the market has decided the leak is permanent. What this memo believes the market underrates is that O'Brien scale, if the cost base holds, widens the incremental margin StoneX has been building for four years. One print settles it: whether operating margin on the combined firm holds or expands in the September full-year report, the first clean look with O'Brien inside it for four quarters.

Management

The record is a builder's and the selling is a seller's. Insiders bought once for $161,000 and sold 39 times for $134 million over the past year, about 1.6% of the company, concentrated in Sean O'Connor, the executive vice-chair, who took $35 million off the table across three sales in May and June. His fiscal 2025 package was $45 million, almost all of it a $41 million stock award, and he began selling into the very quarter the stock made new highs. The bundle does not separate scheduled 10b5-1 sales from discretionary ones, so read the concentration, not the intent. Sitting chief executive Philip Smith's $5 million package is modest against $306 million of profit. No dividend, no buyback since a token $12 million in 2021. Capital goes to acquisitions, $400 million last year, which is the strategy stated plainly.

How it fails or surprises you

The dilution eats the compounding. Net income doubled year on year and earnings per share rose 23%, because the share count ballooned to buy O'Brien. If StoneX keeps paying for growth in stock priced at 12 times earnings, the firm compounds and the owner watches from the platform. The proof: another acquisition funded with equity before the per-share line catches the net-income line.

Float and credit turn at once. A chunk of what the firm keeps is interest on client cash, which thins as rates fall, and the enlarged physical-commodity book adds counterparty exposure in a price spike. Interest coverage sits near 1.4 times. A rate-cut cycle that shrinks float income while a commodity dislocation strains a counterparty is the pincer a thin coverage ratio does not forgive.

Scale is finally real (right tail). If the 89% jump in money kept comes with a cost base that grows slower, the incremental margin built quietly over four years steps up, and a market pricing SNEX as a cyclical broker at 12 times re-rates it as the compounder its net-revenue curve already looks like. The print: operating margin expanding two quarters running after O'Brien.

Closing thoughts

The shape of the payoff is a firm whose enterprise value compounds faster than its share price, unless the September report shows the incremental margin stepping up. This is an uncertainty a named print resolves: the full-year margin on the combined firm. If it expands, 12 times is a gift on a 20%-return business and the fatter tail is up. If it is flat, the dilution read is right and you own a fine business whose owners keep getting diluted, priced correctly at 12 times. The left tail, the one to actually avoid, is a rate turn and a commodity counterparty loss arriving together against 1.4 times coverage.

The bet stays what it was: StoneX keeps being the middleman the real economy needs to lock in prices and move money, and keeps buying rivals to get bigger. What breaks it is the pair to watch, net revenue and diluted share count. If the first keeps doubling while the second keeps climbing to pay for it, the compounding never reaches you. The falsifier is simple: a September full year where money kept rises again and diluted earnings per share does not.