Alex Gerko, the "super geek" that is reshaping FX

What interests me about Alex Gerko and XTX Markets is how much of the business sits inside its software. Research, computing infrastructure and price forecasts determine what the firm can do. Headcount tells only part of the story.

For someone building an AI company, that is worth studying. Engineering choices can change how much work a team handles without requiring the organisation to grow at the same rate.

Compared with a large financial institution, a specialised firm can organise around a narrower problem. Data, computing infrastructure and the ability to test models become central operating choices.

Researchers and engineers still choose objectives, assess data and monitor systems. Automation changes where the effort goes and how far a team’s decisions can reach. A mistake can travel just as far as a good decision.

For other businesses, I would ask which repeated decisions can be supported by software, what evidence would make that support trustworthy, and where human judgement remains necessary.

I would look closely at the feedback loop: how quickly does the team discover that an assumption is wrong, and how quickly can it change the system? Computing power is useful only if those decisions improve with it.