What Actually Made Congo's Ebola Outbreak the Fastest in History

The Democratic Republic of Congo announced this week that confirmed Ebola deaths have passed 2,000, in what health officials call the fastest-spreading Ebola epidemic ever recorded. The outbreak, declared on May 15, has reached 4,449 confirmed cases and 2,061 deaths¹, making it the second-largest in history behind West Africa's 2014-16 catastrophe. The pace is the alarming part: in Congo's 2018-2020 outbreak, it took a little over twelve months² to reach 2,000 deaths. This time, the toll jumped from 1,000 to 2,000 in under a month. The question that matters, both for the people of Ituri province and for the Washington strategists watching their critical-minerals partnership with Kinshasa stall, is why.

Start with a comparison that discipline requires. Uganda declared an outbreak of the same virus on the same day Congo did. It recorded 20 cases and two deaths, followed all 836 identified contacts³ through their monitoring periods, and declared itself Ebola-free on July 28⁴. Same pathogen, no vaccine available to either country, radically different outcomes. Whatever explains Congo's disaster, it is not the virus alone.

The explanation most people reach for is the war. Eastern Congo is contested by Rwanda-backed M23 rebels, the Islamic State-linked Allied Democratic Forces (ADF), and a constellation of smaller militias, all fighting partly over the region's gold, tin, and tantalum. That story is real, and I will come back to it. But the evidence I find most persuasive points somewhere less cinematic: this outbreak spread fastest because the response arrived late, broke, and unarmed.

Unarmed in a specific sense. This is Congo's seventeenth Ebola outbreak, but the first large one caused by Bundibugyo virus rather than the Zaire species. That distinction is load-bearing. The world's only licensed Ebola vaccine, Ervebo, targets Zaire ebolavirus, and it was the workhorse of the 2018-2020 response, where teams used ring vaccination, immunizing the contacts of each new case and the contacts of those contacts, to wall off transmission chains. That campaign vaccinated more than 303,000 people⁶ in an active conflict zone. For Bundibugyo, no licensed vaccine or specific treatment exists⁵, largely because prior Bundibugyo outbreaks produced only 149 and 57 cases⁷, too few to justify commercial vaccine development. The responders of 2026 walked into the same geography as 2018 with the single most effective tool removed from the kit.

Then there is the head start. A study published in Science concluded the outbreak began in January or earlier⁸, near the mining town of Mongbwalu, with more than 500 suspected cases accumulating before the May 15 declaration. Early patients were treated for malaria and typhoid. When testing did begin, health workers initially used assays built for the wrong Ebola species, and samples sent to Kinshasa were mishandled. Four months of silent spread is how you get the numbers we now see: the outbreak blew past 1,000 confirmed cases within 40 days⁹ of the response starting, a threshold the 2018 outbreak took roughly 235 days to reach. By July, WHO officials estimated that around 80 percent of new infections were emerging outside known transmission chains, meaning surveillance was not tracking the epidemic so much as trailing it.

And the head start was not bad luck. USAID was in the process of awarding a five-year disease-surveillance contract for Congo when the agency was abruptly shut down in 2025¹⁰, gutting exactly the early-warning capacity that would have flagged a cluster of hemorrhagic deaths in January. Once the outbreak was declared, Africa CDC (the African Union's disease-control agency) and WHO launched a $518 million six-month response plan¹¹; weeks later, Africa CDC's director general was complaining that less than $2 million in pledges had become real money¹². As of mid-July, WHO had received less than half the funding it needed. By early August, health workers in Ituri were striking over unpaid wages¹³ while the WHO chief admitted the outbreak was outpacing the response, and only then did the State Department announce an additional $242 million, bringing US assistance past $512 million. Money arrived, in other words, on roughly the schedule the death toll did.

The security-first counterargument deserves its full weight before I discount it. In June, a burial team in Katana, a South Kivu town under M23 control, was attacked and forced to abandon a coffin¹⁴, leaving the highly infectious body to be handled by the community. Suspected ADF fighters killed at least 16 villagers in Ituri¹⁵ even as tracing teams tried to work the same territory. Conflict and displacement genuinely complicate surveillance, and Uganda, with no insurgency on its soil and a contained importation cluster rather than months of undetected rural spread, is not a perfectly clean control. But two facts cut against making the war the primary villain. First, the 2018-2020 outbreak unfolded amid the same militias, the same attacks on treatment centers, the same distrust, and the response still registered a quarter-million contacts and completed tens of thousands of safe burials, because it was resourced to push through. Second, response officials say most of the more than a dozen attacks¹⁶ on health facilities this time have come from angry mobs, not militias, and the fix is unglamorous: at one displacement camp near Bunia, teams who were initially stoned gained full access after sitting down with community leaders¹⁷. Community engagement is a budget line. Unpaid, understaffed responses do it badly. The war is the constant in this equation; the collapse in tools and financing is the variable that changed.

Which brings us to the minerals. Congo is the world's top cobalt producer, and Washington has spent the past year building a partnership with Kinshasa designed to pry supply chains away from China, a shortlist that includes Rubaya, the M23-held coltan field supplying an estimated 15 percent of global tantalum ore¹⁸, from which the UN estimates the rebels tax at least $800,000 a month. There is an argument, and it is not crazy, that this deal is ultimately good for outbreak control: a bargain that pacifies mineral corridors would restore the access responders lack. But that theory is speculative, and the observable causality currently runs the other way. Reuters reports the epidemic is disrupting the minerals partnership itself¹⁹: 21-day quarantines, postponed investor visits, a scrapped Washington meeting on Congolese projects. The epidemic Washington under-financed is now taxing the strategic asset Washington prizes.

That is the judgment I keep landing on. The United States treated eastern Congo's minerals and eastern Congo's health system as separable problems, built diplomatic architecture for the first, and let the surveillance infrastructure guarding the second lapse with USAID's demise. A five-year monitoring contract, the kind that costs a rounding error against $512 million in emergency spending, was the cheapest insurance policy the minerals partnership could have bought. It was cancelled in 2025, the virus began circulating unseen that winter, and every delayed meeting in Washington since is part of the interest on that decision.