By Amara Osei

Near-term prediction-market prices (traders’ estimates of event probabilities) favor Jensen Huang’s continued tenure at Nvidia: Kalshi assigns a 54% probability that he departs before January 2029. Polymarket puts a resignation-or-removal announcement at 16% by year-end and 22% by March 2027.

Nvidia shares have fallen nearly 10% since Monday morning, underscoring the financial stakes attached to the leadership question. Traders have turned the corporate scandal into a running forecast of possible departures, testimony and personal outcomes.

What do the CEO contracts show?

Kalshi’s contract on Huang’s CEO departure has drawn more than $266,000 in trading volume. Traders place the probability of his departure before January 2029 at 54%, compared with 52% before January 2028 and 33% before July 2027.

A separate Kalshi contract has drawn roughly $2,500 in trading volume. The probability that Huang will testify before Congress reaches 46% by January 2027 and rises to 79% by July 2027.

Polymarket’s CEO contract has drawn more than $305,000 in volume, making it the largest of the three CEO-related markets cited. Traders put the probability of a resignation-or-removal announcement at 16% by year-end and 22% by March 2027.

Polymarket has also established a personal-life market carrying around $159,000 in bets. The current probability of Huang’s divorce by March 2027 stands at 23%.

What triggered the trading?

Kalshi and Polymarket opened their markets after Huang acknowledged on Sunday to Nvidia’s board that he had an intimate, extramarital affair with a business executive. U.S. intelligence officials reportedly identified that executive as a suspected Chinese spy.

One day after Nvidia announced an investigation into Huang, the platforms had established contracts covering his tenure, possible testimony before Congress and his private life. That sequence links the Sunday disclosure, Nvidia’s investigation and dated contracts, giving traders a framework for revising estimates as facts emerge.

How do prediction markets work?

Kalshi and Polymarket allow users to trade on outcomes across sports, politics and weather. Prices reflect traders’ expectations of how events will unfold, rather than providing a guaranteed forecast.

Each event contract (a wager tied to a specified dated outcome) converts a question into a tradable probability. The CEO contracts range from Huang’s departure date to a resignation-or-removal announcement and congressional testimony.

The platforms drew substantial attention during last week’s 2026 midterm elections. Kalshi recorded $1.4 billion in trading, while traders across the platforms favored Democrats to win control of both chambers of Congress.

How reliable are their forecasts?

The midterm results bolstered prediction markets’ reputation as forecasting tools. The platforms rose to prominence during the 2024 presidential election, when traders favored Donald Trump while polls showed a close race between him and Kamala Harris.

The record remains uneven: markets broadly anticipated Democratic gains in the 2026 midterms, but their prices favored Republicans in Texas’s governor’s race and South Carolina’s Senate race. Democrats won both contests. An earlier surprise in Brazil’s presidential election also raised questions about forecasting accuracy.

What does the market signal mean?

With Polymarket’s exit-announcement prices at 16% by year-end and 22% by March 2027, the CEO contracts place greater near-term weight on continuity than on an immediate leadership change. Kalshi’s 46% January and 79% July probabilities for congressional testimony show a sharper rise in the prospect of formal scrutiny.

For Nvidia, the strategic distinction is material because a prolonged inquiry and testimony could sustain pressure on governance and reputation even without a change in command. The nearly 10% share decline since Monday morning indicates that the market already assigns the episode a substantial cost.

None of these probabilities resolves the underlying questions. Prediction markets provide a live, probabilistic assessment of unfolding events, but their record includes wrong directional calls in Texas, South Carolina and Brazil.