Traders just pushed a grim Manifold contract to the top of the board. A market asking whether there will be an assassination attempt on the CEO of a major AI lab before January 1, 2035 sits at a 91% implied probability of YES as of July 24, 2026, and roughly 54% of its all-time volume changed hands in a single day. The reason is not hard to find: real attacks and threats against AI executives have turned a once-hypothetical question into an active bet.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
What the Manifold market is asking
The contract lives on Manifold, a play-money prediction platform where users trade on the probability of future events. The question, created by user IhorKendiukhov, reads: “Will there be an assassination attempt on a CEO of a major AI lab by January 1, 2035?”
The resolution criteria are specific. An assassination attempt is defined as any deliberate action intended to kill the CEO of a major AI lab, and it must be publicly reported and confirmed by credible sources such as news agencies, law enforcement, or official company statements. The market limits “major AI labs” to a short list, currently OpenAI, Anthropic, and Google DeepMind, with no more than five labs counted at any one time. The market does not close until January 1, 2035, which gives the question a nearly nine-year runway.
Because Manifold runs on its own virtual currency rather than real dollars, the odds reflect participant sentiment and forecasting skill rather than regulated financial exposure. If you are new to how these platforms convert trades into a percentage, our explainer on how Manifold markets price an event is a useful primer.
The odds, and what 91% implies
As of July 24, 2026, the market shows a 91% implied probability of YES on Manifold. In prediction-market terms, that reads as traders assigning roughly nine-in-ten confidence that at least one publicly confirmed attempt to kill a qualifying AI-lab CEO will occur, or will be judged to have already occurred, before the 2035 deadline. Manifold’s page lists 149 holders and 203 trades on the contract at the time of writing.
An implied probability is not a guarantee and it is not a forecast of timing. A 91% YES tells you what the current pool of bettors collectively believes, weighted by how much virtual capital they are willing to commit. For a refresher on translating a headline percentage into what it actually claims, see our note on reading prediction-market odds. The number is high, but the more telling signal here is the speed of the move, not the level.
What is driving the sudden move
The 24-hour surge, with more than half of the market’s lifetime volume trading in one session and 159 unique bettors reported active, lines up with a stretch of news in mid-July 2026 about escalating violence and threats aimed at AI leadership. It is the accumulation of several verified incidents over the spring and summer, including a March 2026 investigation that found large language model chatbots could be prompted into generating detailed plans for violent attacks, up to and including political assassinations, that have made the underlying question feel less abstract to forecasters.
Two threads are doing most of the work. First, a documented April attack on OpenAI CEO Sam Altman’s San Francisco home. Second, a broader, well-sourced reporting cycle in July about AI executives across the industry hiring armed security as threats climb. Together they moved a contract that had traded quietly for months into the most active question on the source’s board.
The April attack on Sam Altman’s home
The anchor event is concrete and on the record. In April 2026, Daniel Moreno-Gama, a 20-year-old from Spring, Texas, allegedly threw an incendiary device, described in some accounts as a Molotov cocktail, at Sam Altman’s San Francisco residence at roughly 3:40 a.m. on April 10, having traveled from the Houston area with the stated intent of killing Altman over concerns about AI. According to CNN, he was charged with two counts of attempted murder, one naming Altman and one naming a security guard, along with attempted arson and weapons counts. Al Jazeera reported the same charges at the state level, with additional federal charges tied to an unregistered firearm and attempted destruction of property by means of explosives.
Prosecutors alleged premeditation. Authorities said Moreno-Gama traveled from the Houston area with multiple incendiary devices, a jug of kerosene, and a lighter, and that he also struck the glass doors of OpenAI’s headquarters with a chair while saying he wanted to burn it down. The Department of Justice said he carried a three-part document, the first section titled “Your Last Warning,” that advocated killing the CEOs of AI companies and their investors and listed names and addresses of apparent executives and board members. The FBI later raided a home in Spring, Texas, connected to the suspect, as Fox News reported.
Reporting also noted that days later an intruder entered Anthropic’s lobby by following an employee inside, claiming he needed to warn that a top Anthropic executive was going to be killed. CNN Business framed the Altman case as exposing a violent fringe of the anti-AI movement.
The wider surge in threats against AI leaders
The July news cycle broadened the story from one suspect to an industry pattern. Per Entrepreneur, citing Wall Street Journal reporting and figures from security firm Liferaft, threats targeting AI leaders and data centers jumped roughly sevenfold between late February and May 2026. Companies responded by expanding round-the-clock protection, hiring armed guards, and quietly extending security to executives’ families.
The spending trail is visible in corporate filings and in the market data behind them. Palantir’s executive-protection budget rose about 150% to nearly $3 million in 2025, and Oracle’s climbed about 85% to $5.6 million, largely to protect chairman Larry Ellison, according to the same reporting. Across S&P 500 technology companies, the share disclosing executive-protection spending grew from 26.8% in 2021 to 38.1% in 2024. Zooming out, an April 2026 industry analysis valued the global executive-protection market at $427.8 million in 2024 and projected it to reach $853.7 million by 2032, a 10.4% compound annual growth rate, with personal-protection services already accounting for 45% of that market. Palantir CEO Alex Karp tied the anger to job-loss anxiety, saying that when people are told their job will disappear, “people go for the pitchfork.” Anthropic has maintained round-the-clock security since 2024.
A timeline traders were pricing
The table below sets out the verified events behind the market’s move.
| Date | Event | Why it matters to the market |
|---|---|---|
| Dec 4, 2024 | UnitedHealthcare CEO Brian Thompson killed in New York | Base-rate reference point for targeted violence against a US corporate chief |
| Since 2024 | Anthropic maintains round-the-clock security | Signals the sector treated the threat as real early |
| Late Feb to May 2026 | Threats to AI leaders and data centers rise about 7x (Liferaft, via WSJ) | Quantifies the escalation traders are reacting to |
| Apr 10, 2026 | Incendiary attack on Sam Altman’s San Francisco home; suspect later charged with attempted murder | The most direct real-world event matching the market’s wording |
| Apr 2026 | Intruder enters Anthropic lobby claiming an executive would be killed | Second named lab drawn into the security story |
| Mid-Jul 2026 | Reporting that AI execs travel with armed guards as threats climb | Catalyst for the 24-hour volume spike on the contract |
Who and what the market covers
The contract is narrow by design. It counts only the sitting CEOs of the labs on its list, which today means Sam Altman at OpenAI, Dario Amodei at Anthropic, and Demis Hassabis at Google DeepMind. A threat against a researcher, an investor, or a board member does not by itself qualify, even though the alleged Moreno-Gama document reportedly named investors and board members as well. The event also has to be publicly reported and confirmed, which rules out rumor and unverified claims.
That framing is worth holding in mind when reading the 91%. The market is not asking whether AI executives face danger in general, a proposition the July reporting already supports. It is asking the narrower question of whether a confirmed attempt to kill one specific category of person, a listed lab’s chief executive, is on the record before 2035.
Why traders disagree over whether it has already resolved
Here is the subtlety that keeps the market from trading at 99%. The April attack on Altman produced attempted-murder charges, which sounds like it should settle the question YES. But resolution depends on the market creator’s reading of whether a given incident meets the bar of a confirmed, deliberate attempt to kill a listed CEO. In the contract’s comments, the creator indicated that a victim’s characterization alone does not decide resolution and pointed toward official law-enforcement confirmation as the standard.
So two live debates sit under the price. One is whether the Altman case already satisfies the criteria, given the filed charges. The other is whether, if it does not, the broader threat environment makes a future qualifying event nearly inevitable over an eight-and-a-half-year window. A trader can arrive at a high YES from either direction, which helps explain why the number is elevated even before any formal resolution. For readers tracking how AI-adjacent contracts behave when news breaks, our coverage of a separate Manifold market on an AI security event shows a similar dynamic.
Context for a tech and finance reader
Prediction markets have drawn scrutiny for listing contracts on death- and violence-adjacent outcomes, and this market sits squarely in that uncomfortable category. It is worth being precise about what does and does not exist here: there is no evidence of a broader, formal commercial “assassination market” built around AI CEOs, and the closest historical analogues are the Bitcoin- and Tor-based sites that surfaced between 2013 and 2015, not an organized 2025-2026 betting venue. This is also not the only CEO-mortality contract Manifold has hosted; a separate market created in August 2025 asked whether another CEO would be murdered in the US between August 11 and December 31, 2025, defining “CEO” broadly as the leader of any company valued at $5 million or more. A 2025 Politico piece on prediction markets raised “assassination market” scenarios as a dark hypothetical within political betting, noting that platforms enforce rules against them and that experts see the scenario as very unlikely in practice. The mechanism that makes prediction markets useful, aggregating dispersed information into a single price, is the same mechanism that makes a market on human harm feel exploitative to critics. That tension is not unique to Manifold. It runs through the broader debate about where prediction markets belong, a debate we track in our work on regulated event exchanges like Kalshi and on how offshore platforms such as Polymarket operate.
For a finance audience, the more durable takeaway is what the money is signaling about corporate risk. Executive-protection budgets are now a line item large enough to appear in filings, and security firms are selling threat-monitoring to an industry that a few years ago treated physical risk as a Silicon Valley afterthought. The Manifold contract is a play-money proxy for a real balance-sheet trend.
What to watch next
Several developments would move this contract. The clearest is the outcome of the Moreno-Gama prosecution and whether the market creator judges it to meet the resolution bar, which could resolve the question YES outright. Short of that, any new confirmed threat or attack naming a listed CEO would push the price further up, while a sustained quiet period and falling threat statistics could soften it. Changes to the lab list, since the market caps it at five, could also shift which people count. Watch the volume as much as the level: another one-day spike would signal fresh news rather than drift.
Frequently asked questions
What are the current odds and where do they come from? As of July 24, 2026, the market shows a 91% implied probability of YES on Manifold, a play-money prediction platform. The figure reflects participant trading, not a regulated financial price.
Has an AI-lab CEO actually been the target of an attack? Yes. In April 2026, a Texas man was charged with attempted murder after allegedly throwing an incendiary device at OpenAI CEO Sam Altman’s San Francisco home, according to CNN and Al Jazeera. Whether that specific case resolves the market YES is a judgment call for the contract creator.
Which CEOs does the market cover? Only the sitting chief executives of the listed major labs, currently OpenAI, Anthropic, and Google DeepMind, capped at five labs at any time. Threats against researchers, investors, or board members do not by themselves qualify.
Why did volume spike now? Mid-July 2026 reporting that AI executives are hiring armed security as threats rise roughly sevenfold, per Wall Street Journal figures cited by Entrepreneur, coincided with the surge in trading.
Can I bet real money on this? Not on Manifold, which uses virtual currency. Real-money availability elsewhere is restricted by jurisdiction, and this article is not a recommendation to trade.
The Bottom Line
Sources
- Manifold — Will there be an assassination attempt on a CEO of a major AI lab by January 1, 2035?
- CNN — Suspect in attack at Sam Altman’s house charged with attempted murder and attempted arson
- CNN Business — The attack on Sam Altman exposed a dark underbelly of the anti-AI movement
- Al Jazeera — Man charged with attempted murder after attack on OpenAI CEO Altman’s home
- Entrepreneur — As threats against AI executives surge, tech giants are beefing up security
- Fox News — FBI raids Texas home of suspect accused of attacking Sam Altman’s house
Prediction markets carry risk and are not investment or betting advice. Odds cited are play-money implied probabilities on Manifold as of July 24, 2026 and can change quickly. Real-money availability is restricted by jurisdiction: Polymarket is not available to US persons, while Kalshi is a CFTC-regulated US exchange. Participation age limits of 18+ or 21+ apply as relevant. If gambling is a problem for you or someone you know, call 1-800-GAMBLER for confidential help.
