Business
Bottom Line
Marex is a diversified non-bank multi-asset wholesale broker — a hybrid of a futures commission merchant, an inter-dealer broker, a market maker and a structured-product issuer, all on one regulated balance sheet. The economic engine compounds because each line feeds the others: clearing brings the client, the client leaves cash, the cash funds market making and the market-making book hedges the structured notes. The market mostly debates the wrong thing — interest-rate sensitivity — and underweights the right thing — that contracts cleared and client balances both compounded through the 2025 rate-cut cycle, so the franchise is structurally widening even as the NII tailwind reverses.
How This Business Actually Works
Marex sells access, plumbing and balance sheet to anyone who needs to clear, hedge, finance or price-discover in the global commodity and derivatives markets. A producer hedges Brazilian sugar, a hedge fund clears LME copper, an asset manager finances an equity portfolio, a private bank distributes a structured note — Marex is the regulated counterparty on the other side, and earns a different fee on each.
The unit economics layer four revenue streams onto a single client. Commissions, trading spread, net interest income and structured-product margin are not the same business even though they share infrastructure.
The four streams interact: clearing wins the client and parks the float; the float earns NII; the same client buys hedges at a spread; the structured-notes desk uses the broader trading book to hedge bespoke payoffs and recycles the proceeds as cheap funding. Marex says clients now sit on $16bn of average clearing balances (Q1 2026), up from $13bn in FY2025 and $11bn in FY2024 — the float keeps growing whether or not the next dollar earns 5% or 3%.
The bottleneck is people, not capital equipment. Compensation is 61% of revenue; capex is well under 1% (FY2025 capex $13M against $2.0B of revenue). This is a high-touch human-capital business with a software wrapper, not a software business with humans attached. Operating leverage therefore works through the comp-to-revenue ratio, not through D&A unwinding.
Where the bargaining power sits. Exchanges and clearing houses set the toll — Marex is a price taker on venue fees and passes them through. Banks have stepped back from sub-scale prime-brokerage clients under Basel III, so Marex sets the price for the mid-AUM hedge funds, commodity producers and private banks who can no longer get bulge-bracket coverage. Pricing is set client-by-client, not from a rate card.
The Playing Field
There is no like-for-like comp for Marex. The five listed names below are each a partial overlap: StoneX is the closest in shape, TP ICAP and BGC are the inter-dealer brokers Marex competes against in voice products, Interactive Brokers is the asset-light electronic extreme, Flow Traders is the principal market-making analogue. None of them does all four on one balance sheet.
Three reads from the peer set. IBKR is the only platform earning >70% pre-tax margin because it is electronic-only and retail-anchored — no human-touch broker replicates that and Marex doesn't try. The inter-dealer brokers (TCAP, BGC) have been stuck at 7–10% pre-tax margins for a decade as voice spreads erode against a heavy cost base — the trap Marex has so far avoided by leaning into clearing and prime services. MRX earns the highest ROE in the cohort (27.5%) on the lowest tangible book base of the five non-IBKR peers because the float-funded clearing engine works the equity capital hardest. The market pays 13x earnings for that today versus 27x for SNEX and 34x for BGC.
The shape "good" looks like here is StoneX with a faster growth rate and a heavier float — what Marex has delivered since the 2022 ED&F Man Capital Markets and 2023 Cowen Prime acquisitions filled out the platform.
Is This Business Cyclical?
Yes — but on three different clocks running at three different speeds. The cycle hits volume, spread and float in turn, and the diversification across all four segments is what mutes the swings at the consolidated level.
This is the single most important chart in the deck. Net interest income fell $74.5m year-on-year as Fed Funds dropped roughly 100bps, but trading income grew $360m and commissions grew $121m on the back of growing volumes and Q4 volatility. Diversification did exactly what management has claimed it would. The bear case for Marex has historically been "this is a rate-cycle stock dressed up as a growth stock"; FY2025 is the first full year of evidence against that.
The cycle reading for valuation: the 2024 NII number was a peak, not a baseline, and the 2025 number is closer to a normal-rate level. Anyone underwriting Marex by extrapolating FY2024's $227M NII run-rate is mispricing the business; anyone extrapolating FY2025's $153M is closer to fair.
The Metrics That Actually Matter
Marex's filings publish dozens of numbers. The five below are the only ones that explain whether the business is creating or destroying value. Industry quartiles are taken from the FY2025 peer set above.
Two of these metrics deserve more weight than they get in sell-side notes. Cleared contracts and clearing balances are leading indicators: both grew through the FY2025 rate-cut year, which means the franchise expanded even as the NII tailwind reversed. Compensation/revenue is the operating-leverage lever: it has been pinned at ~61% through three years of growth, which is unusual for a people-business that has been hiring aggressively. If this ratio bends down — say to 58% — the multiple expands materially without revenue having to do anything more.
What Is This Business Worth?
This is a single integrated economic engine, not a sum-of-the-parts. All four segments share the same clients, the same regulated balance sheet and the same Neon technology stack. Clearing brings the client and the float; Agency and Execution sells more services to that same client; Market Making absorbs the inventory that Solutions creates; Solutions issues structured notes that fund the broader business. Pulling them apart would lose the inter-segment economics, which is why we resist the SOTP framing.
The right lens is therefore earnings power × persistence × capital productivity. The relevant question is not "what multiple should each segment earn" but "is the FY2025 ROE of ~28% a peak or a baseline?"
Marex trades at the lowest P/E of the cohort despite having the highest ROE and the fastest earnings CAGR. The discount is partially explained: a recent IPO (April 2024), foreign-private-issuer status with lighter US disclosure, real NII sensitivity, and a cohort-leading comp ratio. But the gap to SNEX (the closest peer in business mix) is wide — 13x vs. 27x — and is hard to defend if FY2026 numbers continue to come in at Q1 2026's run rate.
The right framing. Marex is not a SOTP story. The four segments share the same clients, same balance sheet and same float — pulling them apart loses the cross-segment economics. Underwrite it as one engine and ask whether a 25–30% ROE persists. If yes, the multiple has room to compress to peer; if no (rates fall further or competition compresses spreads), the discount is fair.
What I'd Tell a Young Analyst
Watch four things, ignore the rest.
Three things the market is most likely getting wrong. Marex is mis-categorised in capital-markets screens against TW, VIRT and electronic venues — the right comp set is SNEX plus an IDB blend. The FY2024 NII number was a peak; anyone basing on it is wrong, FY2025 is closer. The comp ratio has held flat through aggressive hiring and acquisition integration — a 200bps reduction over two years is not a stretch and would add roughly $40m to pre-tax at current revenue.
Three things that would change the thesis. A clearing-margin or credit event that breaks the float-NII model (the 2022 LME nickel episode is the closest recent precedent); a comp ratio drifting above 63% as M&A integration falters; or a regulatory or rate event compressing the NII spread retained by clearing brokers below 50% of policy rates. None is showing up yet — but each is worth a quarterly check.
The Bermuda redomicile vote on May 21, 2026 is worth watching for tax-rate and capital implications, but it is not thesis-changing in either direction.