Quant trading firms make money by collecting a large number of small, statistically favorable payoffs: the bid-ask spread on trades they facilitate, fleeting price discrepancies they correct faster than anyone else, and predictive signals that are right slightly more often than they are wrong. No single trade matters much. The business is a positive-expectancy coin flipped millions of times, with risk management making sure no single flip can end the game.
That one paragraph covers Jane Street, Citadel Securities, Optiver, Jump, and Renaissance alike — but the mix of the three engines differs enormously by firm, which is why interview processes differ by firm too. Here is each engine in detail.
Market making: getting paid the spread
A market maker quotes two prices simultaneously: a bid (where it will buy) and an ask (where it will sell). Anyone who wants to trade right now pays for that immediacy by crossing the spread. If a market maker quotes a stock at \$9.99 bid / \$10.01 ask and trades once on each side, it earns \$0.02 per share while holding roughly zero net position.
The catch is adverse selection: some counterparties know something. If a firm buys at \$9.99 from a seller who knows bad news is coming, the price drops before the position can be offloaded. So the honest per-trade economics look like this:
$$\text{E[PnL per trade]} = \underbrace{\frac{\text{spread}}{2}}_{\text{immediacy fee}} - \underbrace{\text{adverse selection cost}}_{\text{losses to informed flow}} - \underbrace{\text{fees} + \text{hedging cost}}_{\text{frictions}}$$
A worked example. Suppose a firm quotes a \$0.02 spread and does 5 million shares a day in one name. It earns \$0.01 per share of half-spread (\$50,000), gives back \$0.006 per share to informed traders and hedging (\$30,000), and clears roughly \$20,000 — on one instrument. Multiply across thousands of instruments and you see why options market makers like Optiver and SIG hire aggressively and why their interviews lean so hard on options pricing and market microstructure. The skill they are screening for is exactly this: setting a spread wide enough to survive informed flow but tight enough to win volume.
High-frequency trading: getting paid for speed
HFT profits come from being first. Two representative mechanisms:
- Cross-venue arbitrage. The same asset (or a tightly linked one, like an ETF and its basket, or an index future in Chicago and stocks in New York) briefly trades at different prices on different venues. The fastest firm buys the cheap leg and sells the rich leg. The discrepancy might be worth a fraction of a cent and last microseconds — hence microwave towers and FPGA engineering.
- Stale-quote sniping. When news moves the future, a resting quote on a correlated product is momentarily wrong. Whoever hits it first captures the difference; whoever is quoting eats it. This is the arms race in one sentence: fast firms earn what slow quoters lose, so everyone pays for speed in self-defense.
Note that HFT is mostly market making done fast — speed is how you avoid being the stale quote, not a separate business. Per-trade profits are tiny; the moat is infrastructure and engineering, which is why firms in this space run the hardest technical loops (see our firms ranked by interview difficulty).
Statistical arbitrage: getting paid for prediction
Stat arb firms and systematic funds (Renaissance, Two Sigma, DE Shaw, PDT) build models that forecast returns over minutes to weeks: mean reversion between related stocks, momentum, signals extracted from alternative data. A signal that is right 51–53% of the time is a great signal — the money comes from deploying it across thousands of positions so the law of large numbers converts a small edge into a steady PnL stream. This is the researcher's side of the business: less about reacting in real time, more about not fooling yourself with overfit backtests.
The revenue mix by firm type
| Firm type | Primary engine | Holding period | Representative firms |
|---|---|---|---|
| Options market maker | Spread capture + volatility trading | Seconds to days (hedged) | Optiver, SIG, Akuna, IMC |
| HFT / electronic MM | Spread capture + latency arbitrage | Microseconds to minutes | Jump, HRT, Tower, Virtu |
| Stat arb / systematic fund | Predictive signals on client or partner capital | Minutes to weeks | Two Sigma, DE Shaw, PDT, Renaissance |
| Multi-strategy pod shop | Many independent teams, centralized risk | Varies by pod | Citadel, Millennium, Point72 |
One structural distinction worth knowing: prop firms (Jane Street, Optiver, Jump) trade their own capital and keep all the PnL, while funds charge fees on outside capital. That difference flows straight through to how each type pays.
Why this is exactly what interviews test
Once you see the business model, quant interviews stop looking arbitrary. Mental math and probability questions test whether you can price a small edge under time pressure. The classic “make me a market” game tests whether you understand spread, inventory, and adverse selection viscerally. Bet-sizing questions are the Kelly criterion in disguise: firms survive on repeated small edges only if no single bet is oversized. The interview is a compressed simulation of the job.
The fastest way to internalize all of this is to do it: play our market making game and feel adverse selection eat your quotes, work through the market microstructure question bank, or browse the full set of trading games that simulate the decisions these firms make all day.
Frequently asked questions
How do quant trading firms make money?
Quant firms make money three main ways: capturing the bid-ask spread as market makers, exploiting fleeting price discrepancies faster than competitors (high-frequency trading), and trading predictive statistical signals across thousands of positions (statistical arbitrage). Each individual trade earns very little; profits come from repeating a small positive edge at enormous scale with tight risk management.
How do market makers make money if they hold no position?
Market makers earn the bid-ask spread: they buy at the bid, sell at the ask, and pocket the difference while keeping their net inventory near zero. Their main cost is adverse selection, meaning losses to counterparties who trade on information the market maker does not have yet. The business is profitable when half the spread exceeds adverse selection plus fees and hedging costs, multiplied over millions of trades.
How does high-frequency trading make money?
HFT firms profit by reacting to information faster than anyone else, typically by arbitraging tiny price differences between related instruments or venues, and by updating their own quotes before informed traders can hit stale prices. Individual profits are often fractions of a cent per share and last microseconds, so the edge comes from speed infrastructure like FPGAs and microwave networks rather than long-horizon prediction.
Do quant firms trade their own money or client money?
It depends on the firm type. Proprietary trading firms like Jane Street, Optiver, and Jump trade the partners' own capital and keep all trading profits, while systematic hedge funds like Two Sigma or DE Shaw manage outside capital and earn management and performance fees. Multi-strategy platforms such as Citadel and Millennium run outside capital allocated across many independent pods with centralized risk control.
Practice the real thing
QuantVault has 2,800+ quant interview problems with full solutions, intuition, and hints, firm-by-firm interview funnels, and an auto-graded coding judge. Start free.