Race to AGI is live

Race to AGI is live. It's a single-player browser strategy game where you run a frontier AI lab racing three rival labs to build artificial general intelligence — and the twist is that getting there first isn't the win condition. Getting there safely is.

I've been heads-down on this for a while, so here's what it actually is.

The loop

You time a volatile compute market — buying on dips, selling into spikes — and pour researchers into a 12-discipline tech tree that branches out from two core tracks, Scaling and Reasoning, into everything from Robotics and Bioscience to AI Security and Green Data-Centers. Underneath that sits a real economy: funding rounds, debt, salaries, team morale, support teams. Random events and rival labs apply constant pressure the whole way.

Progress isn't abstract. Research ships as versioned product releases, and the whole HUD warms from violet toward gold as your lab climbs from Startup to AGI-Titan-ready — you can feel yourself getting closer.

Reaching AGI is graded, not just timed

This is the part I care about most. Rush the finish line recklessly and you get a "Dangerous AGI" ending — you got there, but it doesn't count for much. Capstone Ethics and AI Security, fund your safety teams, and earn public trust along the way, and you can claim the top ending: AGI Titan. Same finish line, very different story depending on how you ran the lab.

Under the hood

The engine is pure and deterministic — the action log is the save. That's what makes seamless resume possible, and it's also what powers a replay-verified weekly leaderboard where everyone races the same seed. No two players' luck diverges; the only variable is how you play. Every week is a fresh, skill-based shot at building AGI faster — and more responsibly — than the world.

Getting to AGI first is easy to simulate. Getting there in a way you'd be proud of is the actual game.

It's free to play, right now, at rtagi.online. If you play a run, I'd love to hear where you landed — Dangerous AGI or AGI Titan.