Tradeweb Markets Inc. TW

Key data

The business

Tradeweb runs the electronic marketplaces where institutions buy and sell fixed income: U.S. and European government bonds, interest-rate swaps, corporate credit, mortgages, and short-term money-market instruments. It makes money three ways, a fee each time something changes hands (about two-thirds of revenue, and the part that flexes with volume), fixed subscription fees for access, and selling the market data its own venues generate. The moat is liquidity: the venue with the most buyers and sellers gets the next order, which deepens the pool again, and dealers and funds wire their systems straight into it, so leaving is costly. The 2024 purchase of ICD pulled corporate treasurers and their money-market cash onto the platform, adding steadier subscription-like flow. Refinitiv, now part of LSEG, still controls the vote through a second class of shares, so public holders own the economics but not the wheel.

The numbers

Revenue has roughly doubled in four years and the profit line has grown faster, though the last two prints tell a subtler story.

Q1 2026 was the blowout, $618M of revenue on a rates market whipsawed by the Fed. Q2 gave part of it back, down to $559M, up just 9% against a strong year-ago quarter. Note the GAAP spike in Q4 2025, a $325M net-income quarter carrying a large deferred-tax gain, which flatters the trailing profit and the 25 times earnings you pay. On the company's adjusted measure the multiple is nearer 28.

Across the five full years, revenue compounded about 17.5% a year and operating income about 23.5%. Reported EPS grew faster, from $1.09 to $3.78, but tax items did some of that work, so operating income is the honest gauge. Cash conversion is the real prize: the business turns close to half of every revenue dollar into free cash, holds $2.06B against almost no debt, and yields about 4.7% on that free cash at today's price. Against consensus it has cleared the bar four quarters running, each by a couple of cents.

The one number that decides the next year is fee capture. Volume is booming, the take on it is not.

The blended fee per million is sliding as the mix tilts toward larger, lower-margin government-bond tickets and cheaper execution protocols. That is either the price of winning share, which pays back as those clients do more, or the start of a structural squeeze. What this memo believes the market does not: the decade-low cash-flow multiple treats the Q2 deceleration as the new trend, and one clean rates-volatile quarter with steadier fees would break that. The print that settles it is the October Q3 revenue growth and the per-million fee by asset class in that filing.

Management

Insiders are one-way sellers: zero buys and $44.6M sold across 23 filings in a year, led by CEO Billy Hult at $17.5M over February and March and CFO Sara Furber at $7.3M. Plan status is not disclosed in the filings pulled, so whether these are pre-set 10b5-1 sales or discretionary cannot be called. At a controlled company with heavy stock pay, steady selling is closer to routine than signal. Buybacks are modest, $104M last year against about $81M of stock compensation, so repurchases roughly offset dilution rather than shrink the count. Pay rewards revenue and adjusted earnings growth, both delivered.

How it fails or surprises you

Fee compression bites (downside). Rates volume grew 23% year-on-year while total revenue grew 9%, so the fee per million is falling. If that gap holds through the next two prints, high-teens volume growth converts to single-digit revenue growth and the multiple stays low for a reason. The first tell is the variable-transaction-fee-per-million disclosure in the October 10-Q.

The rates market goes quiet (downside). About two-thirds of revenue flexes with volume and volatility. A Fed that stops moving and a calm bond market, the opposite of the last two years, would drain volume fast. Q2's sequential drop from $618M to $559M is a preview. Monthly ADV releases through year-end are the early warning.

Credit and money-market share compounds (right tail). Electronic penetration of U.S. corporate credit is still rising, and ICD is early in cross-selling money-market cash. If credit revenue keeps compounding above 20% and subscription flow scales, EBITDA reaccelerates and a 13x multiple looks like a mistake. The market is not paying for it. The stock sits at its cheapest cash-flow multiple in a decade. Watch the credit and market-data lines in the next two prints.

Closing thoughts

The October quarter will show whether volumes can keep growing while fees stabilize, or whether fee compression is permanent and the de-rating was earned. If revenue reaccelerates above about 12% with fees holding, the decade-low multiple was the opportunity. If volumes rise and fees keep sliding, the high-quality compounder stays fairly priced with a real question mark over whether the moat converts flow into margin. An ambiguous print leaves you holding at a defensible valuation while waiting for clarity. The left tail is not ruin: net cash, free cash near half of revenue, and a genuine moat mean the downside is a lower multiple, not a broken business. The fatter tail, on this evidence, is the upside, because fee compression on rising volume is the ordinary cost of taking share.

The bet is still that the world's bond volume keeps moving onto Tradeweb's screen and that owning the deepest pool lets it earn on that flow. What breaks it is fee capture falling faster than volume rises. Watch rates ADV against transaction revenue each quarter, and if volume keeps climbing while revenue growth stalls in the single digits for two more prints, the moat is not converting and the thesis is wrong.