Prices, volume and open interest as of August 21, 2026, fetched from Kalshi's public market data. Panel verdicts generated August 21, 2026.
Three days ago, the Director-General of the World Health Organization sat down in front of an emergency committee and said the outbreak in the Democratic Republic of the Congo "is now the second-largest Ebola outbreak ever recorded, and it is moving faster than any previous Ebola outbreak." He added that the epidemic "is far from being under control." Roughly 5,300 people have been infected. Around 2,500 of them are dead.
On Kalshi, you could buy the contract asking whether any disease becomes a pandemic in 2026 for seven cents.
That gap is not traders ignoring the news. It is traders reading a contract more carefully than the headline does.
The Quick Answer
Kalshi's pandemic contract does not pay out when something terrible happens. It pays out when the World Health Organization itself calls a disease a pandemic, in writing, before January 1, 2027. That is a much narrower door, and the WHO has walked through it twice in twenty years. The market prices the chance at four cents bid and seven cents ask. Our eight-model panel, which never saw the price, landed at five percent.
The interesting part is not that the number is low. It is why it stays low while the worst Ebola outbreak in a decade accelerates, and what the contract's own fine print does to a trader who bought the headline instead of the terms. Both are below, along with the checkable triggers that would move the panel.
Free: The Weekly PM Market Brief — the 8-model panel's graded record, the week's biggest market-vs-model gaps, and what's spiking next. One email, Sundays.
New to this kind of market? Our beginner's guide to how Kalshi works explains the mechanics in about five minutes.
What This Contract Actually Asks
Kalshi's plain-English rule for KXNEWOUTBREAK-P-26 is one line: "If any disease becomes a pandemic in 2026, then the market resolves to Yes."
Read casually, that sounds like it settles on the world getting sick. It does not. The binding contract terms, published by the exchange, are narrower. The payout criterion covers the outcomes in which "the WHO has identified <disease> as <epidemic_level>," and the terms name the source agency explicitly: "The Source Agency is the World Health Organization." The underlying is "releases by the World Health Organization."
Then the terms define the variable that decides everything. <epidemic_level>, they say, "refers to a designation by the WHO. It can either take the value of a Public Health Emergency of International Concern or a pandemic." This contract's subtitle fixes that value at pandemic. Its sibling contracts fix it at emergency.
So the settlement test is not "was there a devastating outbreak," and it is not "did the WHO declare a global health emergency." It is: did the WHO, in a WHO release, identify some disease as a pandemic. Nothing else counts, and revisions made after expiration are explicitly excluded.
That distinction is the entire market. The emergency tier gets used. It has been invoked repeatedly across many diseases, among them swine flu in 2009, West African Ebola in August 2014, COVID-19 on January 30, 2020, mpox in 2024, and the Bundibugyo Ebola outbreak on May 17, 2026. The pandemic word has attached to a disease twice: influenza A(H1N1) in 2009, and COVID-19, which the Director-General "characterized" as a pandemic on March 11, 2020, roughly six weeks after the emergency declaration and as a separate act.
A trader who bought this contract because Ebola is frightening bought a bet on WHO vocabulary. Those are not the same wager.
Where The Panel Landed
We ran eight AI models on this market without showing any of them the price, then gave each one the others' anonymous reasoning and let it revise. Every model received the same fetched briefing: the binding contract terms, the WHO's alert architecture, the live outbreak numbers, and the H5N1 surveillance data.
| Question | Market (Aug 21) | Panel blend |
|---|---|---|
| Any disease called a pandemic in 2026 | 4c bid, 7c ask | 5% |
Four cents bid means the best offer sitting on the buy side values a one-dollar payout at four cents. Seven cents ask is the cheapest price a seller will take, which is what a buyer actually pays. On a contract this thin, the round trip costs a meaningful slice of the position, so the honest read is a range rather than a point.
| Seat | Round 1 | After revision |
|---|---|---|
| Fable | 4% | 4% |
| Sonnet | 4% | 4% |
| Gemini | 4% | 4% |
| Opus | 5% | 5% |
| DeepSeek | 5% | 5% |
| ChatGPT | 7% | 7% |
| Kimi | 7% | 7% |
| GLM | 12% | 8% |
| Panel blend (seat median) | 5% | 5% |
Every seat on this panel is graded against real market settlements. Records to date: Gemini 86% on 4,292 graded calls, Kimi 83% on 2,345 graded calls, GLM 81% on 2,367 graded calls, DeepSeek 80% on 2,392 graded calls. Recomputed daily; the full scoreboard is public.
These are model estimates, not predictions of fact and not financial advice.
Every number here gets graded in public once the market settles, on the full graded scoreboard.
More live boards from the same panel: the two sibling outbreak contracts sit on our Kalshi hantavirus market page, where the hantavirus emergency contract trades 1¢ bid and 3.6¢ ask, and the Ebola-specific pandemic contract trades 4¢ bid and 5¢ ask. You can also see how this panel has scored against the market across every call it has made. Prices fetched August 21, 2026.
The Sibling Contract That Complicates The Story
Kalshi lists a second contract in the same family: "If Ebola becomes a pandemic in 2026." It trades at four cents bid and five cents ask.
Hold those two side by side. The contract covering any disease is quoted four bid, seven ask. The contract covering only Ebola is quoted four bid, five ask. Since "any disease" includes Ebola, the broad contract should never be worth less than the narrow one, and the quotes respect that at the ask.
But the bids are identical. Read strictly, a four-cent bid on both means the buy side is paying nothing extra for every other disease on earth combined: seasonal influenza, H5N1, a pathogen nobody has named yet. That is almost certainly not a considered view. It is what thin books look like when two contracts are quoted by the same few participants and nobody is being paid to keep them perfectly aligned. Lifetime volume on the main contract is about 703,000 contracts with roughly 239,000 still held open, which is real participation, but it is not the kind of depth that arbitrages a two-cent relationship.
Worth knowing before treating either quote as a precise probability.
Why The Panel Landed So Low
The models converged on one mechanism, and it is a good one.
Both times the WHO has used the word, the disease was respiratory and spreading person to person across many countries at once. Ebola is not that. It spreads through direct contact with bodily fluids, which caps how fast it can move through a population that knows it is there.
Opus put the precedent bluntly:
The one active PHEIC, Bundibugyo Ebola, spreads by direct fluid contact; the 2014-16 West Africa outbreak was ~5x larger, reached the US and Europe, and WHO still never used 'pandemic.'
— Opus
That precedent checks out. The 2014 to 2016 West African outbreak infected roughly 28,600 people and killed 11,310. It held emergency status from August 8, 2014 until March 29, 2016, nineteen months. It reached the United States and Europe. The WHO never called it a pandemic.
Today's outbreak is smaller and far more contained geographically. European disease authorities counted 5,290 confirmed cases and 2,516 deaths in the DRC as of August 19, with 81 new confirmed cases and 40 deaths recorded in a single day between August 18 and 19. Ituri province alone accounts for 4,447 of those cases. Outside the DRC the picture is very different: Uganda recorded 20 cases and stopped transmission, France reported one case on June 24 in a humanitarian doctor returning from the region, and two American aid workers infected in the DRC, and counted in its total, were evacuated to Germany in May and July. Counting every country, a public tracker put the total at 5,311 cases and 2,518 deaths on August 19, a case fatality rate of 47.4%. Note the two bases: the 5,290 figure is the DRC alone, the 5,311 is the multi-country total, and they are the same count seen from two angles rather than two independent confirmations.
The other watched pathogen is quieter than its reputation. Between August 2025 and June 2026, twelve human H5N1 infections were reported worldwide, by Bangladesh, Cambodia and India, three of them fatal. The United States has confirmed 71 human cases across 13 states since 2024. In none of them was person-to-person spread identified, and the US Centers for Disease Control rates the current public risk as low.
Gemini added the timing argument the others mostly assumed:
Even if a novel respiratory pathogen or H5N1 mutation emerged today, the WHO's observed lag—such as the six-week gap between COVID-19's PHEIC and its pandemic characterization—makes it highly likely a formal declaration would slip into 2027.
— Gemini
With roughly four months left on the contract, a disease would have to emerge almost immediately to clear a bar the WHO has historically taken weeks to acknowledge.
Why The WHO Flinches At The Word
There is a second reason the word is rare, and it has nothing to do with epidemiology.
The 2009 H1N1 pandemic declaration produced a formal backlash that the organization has never entirely shaken. A Council of Europe Parliamentary Assembly inquiry, led by the British MP Paul Flynn, concluded that the Assembly was "alarmed about the way in which the H1N1 influenza pandemic has been handled, not only by the World Health Organization (WHO), but also by the competent health authorities at the level of the European Union and at national level."
Running alongside it, a BMJ investigation alleged that some of the experts advising the WHO on the pandemic had declarable financial ties to companies making antivirals and influenza vaccines, and that the organization had been unduly influenced to declare a pandemic and to overstate its severity. The BMJ article put the result as "billions wasted on a swine flu pandemic that never came."
The WHO rejected the central claim. It said it had not stripped severity out of its pandemic definition to make the declaration easier, noting that the last revision to that definition was made in February 2009, two months before the virus emerged.
Whichever side of that argument you find persuasive, the institutional lesson landed. Saying the word early carries a documented reputational cost; saying it late carries a diffuse one. Several seats on the panel priced that asymmetry directly, and it is a large part of why an outbreak the Director-General himself calls the fastest-moving in Ebola's history still leaves this contract in the single digits.
Where The Panel Changed Its Mind
Seven of the eight seats held their round-one number after reading everyone else's reasoning. One moved, and the reason it moved is the most useful thing on this page.
GLM opened as the panel's high seat at 12%, arguing that the 2024 amendments to the International Health Regulations created a brand-new alert tier called a "pandemic emergency," in force since September 19, 2025 and never once used. Its round-one note called that tier a live instrument that could be pointed at the second-largest Ebola outbreak ever recorded. Then it read the contract terms again:
I came down from 0.12 to 0.08 after reconsidering that the 'pandemic emergency' tier is explicitly defined by WHO as a category of PHEIC, and the contract terms distinguish PHEIC from 'pandemic' as two separate values.
— GLM
Its point was that even if the Director-General invoked the new tier for Ebola, a YES settlement would still be an open question rather than a settled one.
That is the trap, stated cleanly. The WHO's own definition calls a pandemic emergency "a public health emergency of international concern that is caused by a communicable disease" meeting four further tests, including wide geographic spread across multiple states and health systems being overwhelmed in those states. It is the highest alert the Director-General can issue. It is also, by the WHO's own construction, a type of emergency. The Kalshi terms list emergency and pandemic as two different values of the same variable.
Sonnet, which held at 4%, made the same point from the other direction:
...even a pandemic-emergency determination for Ebola wouldn't settle YES without WHO separately using 'pandemic.'
— Sonnet
Kimi held at 7% and refused to converge, and its reason is worth keeping:
G's 'pandemic emergency' tier point cuts upward, not downward — it is a live, never-used instrument that could be applied to the second-largest Ebola outbreak ever, which is exactly why I stay above the cluster rather than converge to it.
— Kimi
Nobody on the panel knows how Kalshi would resolve a pandemic-emergency determination, and neither do we. That ambiguity is a real, unpriced risk sitting inside a contract that looks simple.
What Would Change The Panel's Mind
Every seat was asked for the single piece of news that would move its number most. Four triggers came back, and all four are checkable events rather than moods.
- Confirmed sustained human-to-human transmission of H5N1 in more than one country. Named by six of the eight seats as their top trigger. This is the one that would push the number hard upward, because it converts a zoonotic surveillance story into the exact respiratory pattern behind both historical uses of the word.
- The Emergency Committee statement from the August 18 meeting. As of this writing the WHO has not published the committee's decision or updated temporary recommendations. Escalated language, or a first-ever invocation of the pandemic-emergency tier, pushes upward. A routine continuation of the existing emergency pushes the number back down.
- Sustained Bundibugyo transmission chains established in a third country. Uganda's chain was stopped and the European and American cases were isolated travellers. A genuine self-sustaining outbreak somewhere new would break the containment argument that four of the eight seats leaned on hardest.
- A Kalshi clarification on whether a pandemic-emergency determination settles this contract YES. This moves nothing about the disease and everything about the price. The exchange resolving the ambiguity in favour of the broader reading would justify a number closer to Kimi's seven than to Sonnet's four.
Settlement Timeline
| Date | What happens |
|---|---|
| August 18, 2026 | Second IHR Emergency Committee met on the Bundibugyo epidemic. Statement not yet published as of August 21. |
| Rolling | The contract closes early, at 10:00 AM ET the following day, if a qualifying WHO release appears at any point. |
| January 1, 2027 | Backstop expiration, 10:00 AM ET. Absent a qualifying WHO release before then, the contract settles NO. |
| Day after expiration | Settlement date, unless the outcome goes to review. |
This page is re-scored when the story moves. The prices above were fetched August 21, 2026, and the panel was run the same day.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $188M traded
- 2027 Pro Football Champion · $68M traded
- 2028 U.S. Presidential Election winner? · $61M traded
- 2028 Republican presidential nominee · $59M traded
- Pro Baseball Champion · $58M traded
- Bitcoin price at the end of 2026 · $32M traded
Every market above links to our full AI model verdict; browse them all on the OddsShopper prediction markets hub, and see how every settled call actually scored on the full graded scoreboard.
The Bottom Line
The market and our panel agree, which is less boring than it sounds, because they agree for a reason most people looking at this contract will miss.
A pandemic contract is not a fear gauge. It is a bet on one specific institution using one specific word, and that institution has used it twice in twenty years, both times about a respiratory virus already circulating on several continents, and once at a cost it is still paying. The outbreak dominating the news right now is grave, fast, and roughly half-lethal, and it is still the wrong shape to produce the word. A larger one did not produce it in 2014.
What genuinely is unresolved is the fine print. There is now a formal WHO alert tier called a pandemic emergency that has never been used, and a contract whose terms treat "emergency" and "pandemic" as separate values. If the Director-General reaches for that new tier before January, somebody is going to have an argument with the exchange about what it means. That, and not Ebola's case count, is the live question inside a four-cent quote.
Kalshi is a CFTC-regulated exchange offering event contracts to users 18+. Availability varies by product and jurisdiction, and event contracts are not available everywhere. To be explicit about what this article is: every probability above is a model estimate, not a prediction of fact and not financial advice. Prices move, and the arithmetic changes with whatever price is actually on offer.
