, July 11, 2026

Kalshi Traders Bet on Things That Will Never Happen


Traders on the prediction markets platform see a less than 30% likelihood the federal government will take a stake in AI companies Anthropic and OpenAI.

  •   1 min read
Kalshi Traders Bet on Things That Will Never Happen

Table of content

Prediction market users now assign a less than 30% chance the federal government takes a stake in OpenAI or Anthropic this year. These are the same people who thought polymarket odds mattered in an election.

The government taking equity in private AI companies would require Congress to pass legislation, appropriate funds, negotiate terms with venture capitalists who already own most of both companies, and explain to voters why their tax dollars are buying shares in chatbots that can't count the r's in strawberry. None of this will happen before December 31st.

But seventy percent odds against still means thirty percent odds for. Three in ten Kalshi users looked at the current legislative calendar, the debt ceiling fight, the fact that OpenAI is structured as a capped-profit entity controlled by a nonprofit, and the general speed at which the federal government does anything, and thought yeah, seems plausible.

Thirty percent is the same probability you'd assign to flipping heads twice in a row. Kalshi traders are treating a complex multi-month legislative and financial restructuring like a coin flip. Actually worse than a coin flip, because coins don't require appropriations committees.

The beauty of prediction markets is they let anyone with an opinion and a credit card pretend their guess is somehow more rigorous than a guess. Call it a market. Add a percentage. Suddenly Todd from Denver clicking buttons on his phone becomes price discovery.

OpenAI is currently valued at $157 billion, which means a meaningful stake would cost tens of billions of dollars the government definitely has lying around. Anthropic is worth less but still requires the kind of money that shows up in budget reconciliation bills, not discretionary spending.

The traders pricing this at 30% aren't predicting policy. They're buying lottery tickets that pay out if something unprecedented happens in five months, which is exactly how you'd expect retail traders to approach anything involving AI and percentages they don't understand.

Photo by on Unsplash

Related Posts

The Noise is free. If Phil's commentary made you laugh or think, he accepts tips. No pressure — the sarcasm was complimentary.

Leave a Tip