Weather is one of the few things a prediction market can settle without any argument. The temperature either crossed 95 degrees on July 15 or it didn't, and there's a measurement station to prove it. That clean settlement, plus the fact that weather touches almost every corner of the economy, is why these markets have been growing while traditional weather derivatives stayed too expensive and fiddly for most people to bother with.
What weather contracts actually look like
Most of them are binary questions. Will the high in New York City top 95 degrees on a given day? Will a hurricane make landfall in a specific region this season? Will seasonal rainfall come in above or below average?
How you write the contract matters more than people expect. "Will NYC's high exceed 95 on July 15" is precise and verifiable, and there's nothing to fight about at settlement. "Will this summer be hotter than average" drags in three separate arguments: which average, which station, and which dates even count as summer. The vaguer the wording, the more the resolution turns into a dispute instead of a payout.
Kalshi has leaned into weather harder than most, which tracks with the CFTC being comfortable with the category. These contracts also run shorter than political markets, usually weeks to months, so your capital turns over faster. Even a small per-trade edge can annualize into something decent when you're recycling it every few weeks.
Who's actually trading them
The natural crowd here is anyone with real weather exposure: farmers, energy companies, event planners, insurers. When those participants hedge, they're not guessing, they're pricing in actual economic information, and that information ends up in the market.
Meteorologists have gotten active too. If you run your own numerical weather model, you can line your output up against the market price and trade the gap when they disagree. Do that across enough forecasters and the price ends up being a blend of many models, which historically beats any single one of them.
Retail adds liquidity but tends to trade off vibes and recency. It's been hot all week, so they buy the heat contract. That bias is exactly the opening for anyone who bothers to know the base rates and seasonal patterns.
Where the efficiency breaks down
For 1 to 7 day forecasts, these markets are honestly pretty efficient. Short-range weather forecasting has gotten good, and the market price hugs the best numerical models closely. To beat it in that window you need a better model or faster data, and that's a hard game.
Push out to 2 to 8 weeks and it changes. Forecast accuracy falls off, subseasonal forecasting is genuinely hard, and far fewer people bother trading the longer-dated contracts. Thinner pool plus a fuzzier signal is where edge tends to live, if you've got real meteorological chops or access to specialized tools. It's the same pattern I see across most prediction categories on Blockcircle: the efficient window is short, and the money is in the part everyone else finds annoying to model.
The climate versions, and the practical use
Climate contracts are the newer and messier cousin: annual temperature records, multi-year trends, frequency of extreme events. They're harder to price because the horizons are long and you've got barely any relevant precedent to anchor on. Still, they could become a real way to quantify climate risk, and that category should grow as climate keeps showing up in financial decisions.
On the practical side, this is where prediction markets beat traditional weather derivatives for a lot of businesses. A farmer who wants to hedge a drought doesn't have to negotiate a custom OTC contract with a bank. They buy contracts on a platform with low minimums and simple mechanics. You give up customization and some liquidity on very localized risks, but for most people the easier access is worth that trade. If you're a business with weather exposure and you've never priced a derivative before, this is a reasonable place to start small and see how the hedge behaves before you scale it.