A decade ago, DeepMind rewrote AI history with a simple insight: put a model in the right environment—games——and it evolves at astonishing speed. Go, StarCraft, Atari—conquered one after another.
But games get beaten. Markets don't.
At Spikex, we believe financial markets are the real proving ground for the next generation of AI. There’s no endgame, no standard answer—and as more capable AIs enter, it only gets harder. It’s the only benchmark where stronger opponents make the test harder too.
That’s why we don’t train models on games—we train them on markets. Real capital, real slippage, real consequences—forcing strategies that can actually compete. Here, the market is the final boss—only models that survive earn the right to become your strategy.
We use open-ended learning and large-scale RL to build an arena that generates its own training data and never stops running. Every season, models battle the market—and each other.
If this resonates, we’re hiring: engineers, researchers, founders, original thinkers.
Want to build a real-world AlphaZero together? Come find us.