Cardinal Alpha is a quantitative trading firm founded in 2020. Everything we do sits on two foundations: rigorous, testable research, and a risk framework that comes before any trade, not after.
We believe markets that are still maturing carry more structural inefficiency than markets that have been arbitraged for a century. We built the research, the infrastructure, and the risk discipline to pursue that inefficiency directly, at scale.
Every idea, at every speed, runs through the same two filters: does the research hold up out-of-sample, and does the risk framework allow it. One research process, one execution and data platform, and one risk framework, sized so that no single trade, and no single regime, should be able to threaten the firm.
That discipline plays out at three different speeds. Arbitrage is fast: it finds two venues, prices, or funding rates that should agree, waits for them to disagree, and closes the gap. Statistical arbitrage is a step slower: it finds correlated assets that have drifted apart and waits for the model's edge to bring them back. The multi-strategy book runs slowest and widest: part of it stays with a market already in motion, and part of it fades a move once it looks stretched. The model weighs both continuously.
See how that plays out day to day on What We Do →
Every signal, on-chain or off, starts as a hypothesis. Most don't survive contact with data they haven't seen yet, and we'd rather kill an idea in research than in production.
Every position is sized to a loss the firm can absorb without changing the plan. Risk limits are set before a trade, never renegotiated during one.
The same exchange connectivity, custody, and data platform serves a spread that closes in minutes, a relationship that reverts in days, and a trend that plays out over months, built for correctness first, speed second.
Researchers ship their own signals. Traders read their own code. Every desk owns its strategy end to end: no black boxes, no handoffs to someone who wasn't in the room.
We work with researchers, engineers, and traders who are comfortable owning a problem end to end, from the idea, to the code, to the position. No trading-floor hierarchy: good signals win the argument regardless of who found them.