Kelvin Leung Left One of the World's Top Trading Firms for AI — Here's Why

There is a particular kind of clarity that comes from working in markets long enough. You learn to read signals before they become obvious. You develop instincts about where momentum is building, where capital will inevitably flow, and most importantly, when the expected value of staying put no longer justifies the opportunity cost of doing so.

For Kelvin Leung, that moment of clarity came through the slow accumulation of evidence. A quantitative researcher and trader who built and scaled US equity options strategies at Optiver, one of the most respected proprietary trading firms in the world, Kelvin recently made the kind of decision that would give most finance professionals pause: he left a well-defined, upward trajectory to co-found an AI-focused hedge fund.

It sounds bold. But when you hear him explain the logic behind it, it sounds almost inevitable.

Leung's background is not that of someone prone to impulsive moves. He studied mathematics at Cambridge, completing both his bachelor's and master's degrees there, before joining Optiver's Amsterdam office in 2022. Quantitative trading attracted him for reasons that were intellectual as much as financial: the game theory, the fast-paced decision-making, the discipline of forming a view and then actually committing capital to it.

On the single-stock options desk, he developed a US equity options strategy with new signals, drawing on conceptual foundations from the European desk but constructing the American version himself. By the time he departed, the strategy was being scaled up significantly. It was, by any measure, a success story still in motion.

And yet, he left.

Kelvin is not alone in what he observed. Across trading floors and investment banks, a quiet reallocation is underway, not just of capital, but of talent. Some of the sharpest quantitative minds in finance are beginning to ask the same uncomfortable question: is there a better use of my skills given what AI is becoming?

For him, watching the development of AI from inside the markets made the trajectory impossible to ignore.

"AI is moving faster than anything I've seen," he says. "The pace of development is staggering, and it's only accelerating. This is the most transformative technology of our lifetime, and I wanted to be part of it rather than watch from the sidelines."

That sentiment is increasingly reflected in market data, as AI remains a dominant investment theme drawing institutional capital at a pace that has surprised even optimistic forecasters. What is changing now is the profile of those positioning themselves to capture this opportunity. An increasing number of quantitative professionals are being drawn to the space because the core skills used to generate alpha in traditional markets—statistical modeling, signal construction, and rigorous risk management—are the exact tools required to navigate the complexity of AI-driven disruption. This migration represents a recognition that the transition is a natural evolution for those with a shared foundation in quantitative rigor.

What makes Kelvin’s account worth paying attention to is not just the decision itself, but the intellectual honesty with which he describes the process of getting there. He is candid about the psychological friction involved in moving from a stable, high-performing career into genuinely uncertain territory.

"The hardest part has been rewiring my own utility function," he explains. "Naturally, humans punish losses far more than they reward gains, and for a long time that made me default to the lower variance option too often."

It is a remarkably self-aware observation, and one that will resonate with anyone who has wrestled with the tension between a predictable upside and an asymmetric one. In trading, this tension is formalized; you can model it, price it, debate the parameters. In career decisions, it is far messier. The variables are harder to quantify, and the feedback loop is slower.

"Staying on a comfortable path gives you a predictable outcome," he says, "Leaving to build something new gives you a much wider distribution, and I'd rather have that asymmetry."

Kelvin Leung's story is one data point in a much larger trend. The rotation of smart money into AI is not purely a story about technology stocks or venture capital rounds. It is also a story about where the most rigorous and experienced market participants are choosing to place their own careers.

For investors watching from the outside, there is something instructive in the pattern. When the people best equipped to identify mispricing in complex financial instruments start redirecting their attention toward a single theme, it is worth asking whether the market has fully priced in what they are seeing.

Kelvin himself is measured about it. He acknowledges the uncertainty ahead: the challenges of co-founding a fund are real, the competitive landscape is formidable, and the AI investment thesis, while compelling, is not without its own risks. But there is a difference between acknowledging risk and being deterred by it.

"There's certainly going to be a lot more uncertainty and challenges," he says. "But I think it's going to be an incredible journey regardless."

For a quantitative trader, that kind of statement carries a certain weight. These are not people who speak loosely about risk-reward. When someone who built their career on rigorous probabilistic thinking decides the expected value of a new venture outweighs the comfort of a known one, that in itself, is a signal worth noting.