Why We're Building MNX

The AI economy needs a real exchange.

As founders of @ManifoldMarkets, we’ve seen the power of prediction markets firsthand. Our markets called the Russian invasion of Ukraine months before conventional wisdom caught up. They priced in AI winning gold on the International Math Olympiad while experts claimed AI is incapable of human-level reasoning. Over 30 million predictions later, the verdict is clear: prediction markets work for forecasting.

But prediction markets today have a problem. They are not nearly ambitious enough.

The real opportunity for prediction markets isn’t giving retail traders another way to bet on who wins the presidency or the Super Bowl. It’s giving a fund manager with a billion-dollar portfolio a way to hedge an idiosyncratic risk that no existing instrument covers. The moment prediction markets become useful for institutions managing real economic exposure, the volume potential rises by orders of magnitude.

But no one has built the infrastructure to get there.

Prediction markets aren’t the only underleveraged innovation from crypto (pardon the pun). If you want leveraged exposure to Nvidia today, you trade options, which means navigating Greeks, expiry dates, and time decay. If you want to short it, you borrow shares through a prime broker and pay daily fees. If you want to trade TSMC or Samsung, you need access to foreign exchanges.

A perpetual future or perp solves all three problems with a single instrument. It’s a synthetic contract that tracks any underlying asset, lets you go long or short with leverage, never expires, and doesn’t require access to the underlying market. In crypto, perps already dominate spot volume on every major token. They haven’t transitioned to traditional assets yet, only because the necessary infrastructure didn’t exist.

That’s changing.

We believe that within five years, all major equities and commodities will see more volume on perps than on their spot markets.

We have two powerful financial innovations, prediction markets and perps, both currently stuck in the retail world, waiting for someone to build a bridge to institutional use. The obvious place to start is AI.

AI will be the most important sector in the global economy, and it has almost no purpose-built financial infrastructure. Consider what doesn’t exist today: there are no compute futures, so hyperscalers and startups have no way to hedge the cost of training runs that can exceed $100 million. There are no AI benchmark markets, so investors have no instrument to express a precise view on the rate of capability progress. There are no valuation markets for private AI labs, so employees and LPs holding stakes in companies like OpenAI or Anthropic, worth hundreds of billions, have no path to liquidity or price discovery. Each of these gaps represents a question that will determine how trillions of dollars in capital get allocated, and traditional finance can’t offer most of them since the assets are too novel and the regulatory frameworks are too slow.

This is exactly where crypto infrastructure shines. A few years ago, building an institutional-grade exchange onchain would have been a fantasy. That’s no longer true. Stablecoin adoption has crossed a tipping point: USDC and USDT now settle more transaction volume than Visa. High-performance chains like @MegaETH can match the throughput and latency that serious traders expect. And crucially, post-FTX, non-custodial settlement means that traders don’t have to trust an exchange with their funds. The infrastructure to build a real, institutional-grade trading platform onchain finally exists.

And that’s why we’re building MNX: the AI economy needs a real exchange.

Try testnet here:
https://testnet.mnx.fi

Follow @MNX_fi for updates.

Trading goes live end of Q1.

— @StephenGrugett & @realpalebone