On May 17, 2026, the World Health Organization declared a public health emergency of international concern for an Ebola outbreak in the Democratic Republic of the Congo and Uganda. The causative virus was Bundibugyo \u2014 first identified in a Ugandan district in 2007, characterized in the scientific literature, discussed at international conferences, listed on priority pathogen watchlists, and granted nearly two decades of institutional concern. Doctors treating patients in 2026 had no approved vaccine. They had no approved treatment. They had a case fatality rate somewhere between 25 and 40 percent.<\/p>\n\n\n\n
Not because no one knew Bundibugyo existed. Not because the biology was intractable. The rVSV vector platform that produced Ervebo \u2014 the licensed Zaire ebolavirus vaccine, approved by the FDA in 2019 \u2014 is adaptable technology, and the relevant biological knowledge for applying it to Bundibugyo had been accumulating for years. After WHO declared the emergency, CEPI moved quickly: three investigational candidates fast-tracked, including a $3.2 million award to IAVI for an rVSV-based Bundibugyo candidate. Reactive, unproven, weeks from any possible clinical use at best.<\/p>\n\n\n\n
Three point two million dollars. Weeks into a declared emergency. For a virus known since 2007.<\/p>\n\n\n\n
The eighteen-year gap between discovery and that emergency award has nothing to do with biology. It has everything to do with a calculation \u2014 specific, documented, rational \u2014 that pharmaceutical companies are legally and financially required to make before beginning any development program. Understanding that calculation is understanding the failure. And the failure, when you understand it, is worse than the easy version.<\/p>\n\n\n\n
If you were a pharmaceutical company in 2009 \u2014 two years after Bundibugyo was first characterized, looking at the outbreak data across your conference table \u2014 the relevant question was not “can we make a vaccine for this?” The relevant question was “will the expected return on developing this vaccine be positive?” Those are completely different questions, and confusing them is how this problem gets misread as a scientific one.<\/p>\n\n\n\n
Vaccine development is expensive in ways that compress the number of actors willing to attempt it. The 2024 brief from the US Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation \u2014 the most comprehensive recent accounting of vaccine development economics \u2014 found that bringing a vaccine to regulatory approval costs, on average, $886.8 million. The development timeline averages 15.3 years from research program to regulatory approval. Out-of-pocket costs per trial phase average $6.5 million for Phase 1, $14.6 million for Phase 2, and $104.4 million for Phase 3 \u2014 but the $886.8 million figure is an expected capitalized cost \u2014 it incorporates the cost of capital across the fifteen-year development timeline and adjusts for the high probability of failure at each stage, since the majority of vaccine candidates never reach approval. It is, in short, the figure a pharmaceutical company uses when deciding whether to begin.<\/p>\n\n\n\n
So a company making a development decision is looking at roughly a billion dollars of capital deployment, committed over fifteen years, against the probability of technical success, against projected market size and per-dose price, against competitors, reimbursement dynamics, and the time value of money. The exercise has a name in pharmaceutical finance: expected net present value analysis. It is not a moral judgment. It is an accounting exercise, and it produces a binary output \u2014 positive or negative. A positive NPV means development is commercially rational. A negative NPV means it is not.<\/p>\n\n\n\n
Expected NPV applies a discount rate to project cash flows, which means a dollar earned fifteen years from now is worth less today than a dollar earned in five years. For a vaccine requiring a decade and a half of investment before any revenue appears, that discount effect alone substantially inflates the real cost of development in present-value terms. Add in the probability weighting \u2014 the expected NPV must account for the likelihood that this specific program will fail, multiplied by the full sunk cost of the failed investment \u2014 and the required market return to justify an NPV-positive decision becomes large. Large enough that most pharmaceutical executives, presented with Bundibugyo’s profile in 2009, would not have needed a spreadsheet to reach a conclusion.<\/p>\n\n\n\n
For a statin or a cancer therapy or a diabetes medication \u2014 diseases affecting hundreds of millions of people in countries with deep insurance systems and high willingness to pay \u2014 the expected NPV is reliably positive. Markets are large. Prices are high. The probability of regulatory success is well-characterized by historical data. Returns, while uncertain, are predictable within ranges that justify the risk.<\/p>\n\n\n\n
For a hemorrhagic fever virus concentrated in Central Africa with outbreak sizes under a hundred cases, the calculation looks different in every variable simultaneously.<\/p>\n\n\n\n
In 2009, the Bundibugyo data showed a disease that had produced 93 putative cases in its first confirmed outbreak, in Uganda in 2007\u20132008. The 2012 outbreak in DRC produced 52 confirmed cases and resolved. The effective market for any resulting vaccine \u2014 the buyers who would actually purchase doses \u2014 consisted of government health ministries in DRC and Uganda, operating within procurement frameworks that cap prices at a few dollars per dose, in countries where annual per capita GDP is below $600 and government health budgets cannot support large advance purchase commitments. International emergency stockpiles and donor-funded outbreak response exist, but they are structured to buy at humanitarian prices, not at prices that recover hundreds of millions in development costs. The time-to-revenue problem, meanwhile, compounds the discount-rate problem: a vaccine for an epidemic that erupts every decade cannot be monetized steadily over time the way a chronic-disease drug can.<\/p>\n\n\n\n
A rational investor, looking at all of this in 2009, concludes the expected return is negative.<\/p>\n\n\n\n
They are right. The calculation is correct. This is not a story about callousness or pharmaceutical greed \u2014 those framings are both too comfortable and too easy to dismiss. This is a story about rational actors using the correct inputs and producing the correct output, and that correct output being catastrophic for the populations it excludes. That’s what a structural market failure actually looks like: not bad decisions, but decisions that are individually correct and collectively devastating.<\/p>\n\n\n\n
A Bundibugyo vaccine program, optimistically modeled, might project a market of 10\u201315 million high-risk individuals in DRC and Uganda, reachable over a decade at $3 per dose \u2014 peak annual revenue of maybe $30\u2013$45 million in a very generous scenario, earned only in outbreak years, heavily discounted over the fifteen-year development horizon. Against a billion dollars in upfront costs. No pharma CFO requires a detailed model to read that picture. The company spends its capital elsewhere. The pattern repeats across Nipah, across Crimean-Congo haemorrhagic fever, across Rift Valley fever. The market doesn’t fail Bundibugyo specifically. It fails any pathogen whose outbreak population cannot afford to pay development-sustaining prices and whose government cannot commit to pay them on its behalf.<\/p>\n\n\n\n
Bundibugyo doesn’t fail the pharmaceutical investment calculation once. It fails it three times simultaneously, and each failure compounds the others.<\/p>\n\n\n\n
The first failure is purchasing power. DRC’s GDP per capita sits below $600. The global procurement architecture for vaccines in low-income settings \u2014 run through GAVI and UNICEF \u2014 has driven prices for routine vaccines to roughly $2\u2013$3 per dose, sometimes lower. That pricing structure exists precisely to extend vaccine access into the countries that most need it. But it means a manufacturer trying to recover $886.8 million in development costs from this market faces arithmetic that closes badly. At $2 per dose, recovering the development cost alone \u2014 before production costs, before distribution, before any return on invested capital, before accounting for the failed programs that didn’t make it to approval \u2014 requires selling roughly 450 million doses. The entire population of DRC is approximately 100 million people. There is no version of this arithmetic in which the numbers approach each other at any plausible dose price within this market.<\/p>\n\n\n\n
The counterargument is sometimes floated: wealthy-country outbreak response stockpiles, emergency procurement at higher prices, humanitarian donor purchases. These exist. They are not large enough to change the calculation. WHO outbreak response funds and donor emergency stockpiles operate at humanitarian prices, not development-recovery prices. And they are structured for reactive purchasing \u2014 buying doses of something that already exists, not funding the decade-long process of creating something that doesn’t. Even optimistically sized, their aggregate purchasing capacity for any single neglected pathogen represents a fraction of what would be required to flip a negative expected NPV to positive for an investor deciding in Year 1 whether to open a Phase 1 file.<\/p>\n\n\n\n
The post-PHEIC response itself demonstrates the failure precisely. CEPI’s $3.2 million emergency award to IAVI after the 2026 outbreak was declared is the largest rapid mobilization the system managed for Bundibugyo \u2014 and it represents reactive funding for an investigational candidate that wasn’t in clinical development before the emergency was declared. That $3.2 million, reactive and insufficient as it is, is not part of the expected-return calculation a pharmaceutical company makes in 2009. It is a donation that materializes only after the catastrophe \u2014 after the deaths that would have been prevented by a pre-existing stockpile. Emergency humanitarian funding, however welcome, does not retroactively rewrite the investment logic that left the stockpile empty.<\/p>\n\n\n\n
The second failure is government procurement leverage. The mechanism that made COVID-19 vaccines commercially viable was not just market size \u2014 it was the US government’s willingness to act as a large, creditworthy buyer committed to paying development-sustaining prices before any regulatory approval existed. The initial US government contract for Pfizer\/BioNTech’s COVID-19 vaccine was $19.50 per dose \u2014 a signal, early and substantial, that a buyer with an enormous balance sheet was committed to purchase at scale. That signal changed the expected-NPV calculation for manufacturers. It meant that even before Phase 3 results existed, the revenue side of the ledger had a floor. DRC has no comparable lever. Its government cannot commit to buy a future Bundibugyo vaccine at development-sustaining prices, because it doesn’t have the budget or the institutional infrastructure to make such a commitment credible. What it has, when an outbreak arrives, is a request for international emergency assistance.<\/p>\n\n\n\n
These two failures are about money. The third failure is structural, and it may be the most perverse of the three.<\/p>\n\n\n\n
Regulatory approval requires demonstrating efficacy in a large randomized controlled trial with sufficient incident cases to measure the protective effect statistically. The Merck VSV-ZEBOV trial that produced Ervebo enrolled 11,841 participants across Guinea in 2015. That trial was possible because the 2014\u20132016 West Africa Ebola epidemic generated approximately 28,000 cases across Guinea, Liberia, and Sierra Leone over two years \u2014 a sustained outbreak with the geographic concentration and duration to make a large, well-powered trial logistically achievable.<\/p>\n\n\n\n
Bundibugyo has never produced 28,000 cases. Its outbreak pattern is precisely the opposite: sporadic, brief, and geographically scattered. The 2007\u20132008 Uganda outbreak produced 93 putative cases and was over in weeks. The 2012 DRC outbreak produced 52 confirmed cases and resolved before research infrastructure could be mobilized at scale. There is no epidemiological reason to expect this pattern to change. Bundibugyo’s disease characteristics \u2014 the same ones that make it lethal but rare \u2014 also make the evidence-generation process required for regulatory approval essentially impossible to execute under current frameworks.<\/p>\n\n\n\n
The Phase 3 problem<\/strong>\n\nA protective efficacy trial for an outbreak pathogen like Bundibugyo requires thousands of participants at active risk of exposure, with enough incident cases in the trial population to distinguish vaccine-induced protection from baseline low transmission. Ring vaccination designs \u2014 enrolling contacts and contacts-of-contacts of confirmed cases \u2014 require sustained outbreak chains to function at all. The Merck VSV-ZEBOV trial enrolled 11,841 participants and achieved zero cases in vaccinated individuals after ten or more days of follow-up, a result made possible by two years of active, sustained transmission across three countries.\n\nRunning such a trial in Bundibugyo-affected territory requires: vaccine cold-chain capacity in remote DRC; certified sample collection and transport to reference laboratories; ethical oversight from both national and international review bodies; active regulatory engagement with DRC authorities; and crucially, enough cases in a tight enough geographic area for long enough to reach statistical significance. Bundibugyo's last two outbreaks were over before most of these prerequisites could be assembled. Adaptive trial designs and platform approaches can reduce some of these barriers \u2014 but none of them are funded or tested for Bundibugyo, because funding them requires an expected positive return that doesn't exist.<\/code><\/pre>\n\n\n\nThis is the catch-22 that no amount of scientific ingenuity has resolved: the disease’s outbreak pattern prevents the evidence-generation process required for regulatory approval. You cannot run the trial that would give you the approval that would let you stockpile the vaccine that would let you respond when the next outbreak arrives. Not because the science is hard. Because the outbreaks are too small and too fast and too remote to serve as the substrate for large clinical trials.<\/p>\n\n\n\n
And because the market never justified funding the adaptive trial approaches that might have worked around it.<\/p>\n\n\n\n
The rVSV platform that could produce a Bundibugyo vaccine \u2014 the same one CEPI’s $3.2 million emergency award is now funding \u2014 was validated and sitting in laboratories before the 2026 outbreak. What didn’t exist was anyone willing to pay for the development process, because the development process costs far more than any plausible market could return. Compare this to influenza or hepatitis B, both scientifically comparable to outbreak ebolaviruses in terms of platform complexity: both have licensed vaccines, because both affect enormous global populations with sufficient purchasing power to generate positive expected NPV. Influenza mutates constantly and requires annual reformulation \u2014 a development challenge of substantially higher complexity than adapting an existing rVSV vector. The variable that differentiates Bundibugyo from pathogens with vaccines is not biological complexity. It is market geometry.<\/p>\n\n\n\n
The fixes<\/h3>\n\n\n\n
The market failure is understood. It has been understood, with increasing analytical precision, since at least the mid-1990s. Every mechanism designed to address it is a response to a specific theory of what’s wrong. Most of them are technically sophisticated. The question is whether they solve the actual problem at the scale the actual problem requires.<\/p>\n\n\n\n
They do not. And the reason is not simply that they’re underfunded \u2014 it’s structural.<\/p>\n\n\n\n
Advance market commitments were designed to solve a specific version of the problem: the situation where a vaccine exists but manufacturers have no viable market signal to produce it at scale for low-income populations. The pneumococcal AMC is the canonical success. A $1.5 billion donor fund \u2014 assembled from the Bill & Melinda Gates Foundation and five governments \u2014 guaranteed manufacturers a viable purchase price for pneumococcal vaccines delivered to low-income countries. Two manufacturers already had approved products: GSK with Synflorix, Pfizer with Prevenar. They needed a credible demand signal to invest in low-income-market supply capacity and bring prices down. The AMC provided it. By the end of 2020, more than 225 million children had been vaccinated and over 700,000 deaths had been prevented. The mechanism worked.<\/p>\n\n\n\n
The conditions under which it worked explain why it hasn’t generalized.<\/p>\n\n\n\n
The pneumococcal AMC solved a distribution problem. There were existing products. There were willing manufacturers. What was missing was a viable buyer. The AMC created one. The Bundibugyo problem is a different problem: there is no product. A hypothetical Bundibugyo AMC would need to function as a “pull” mechanism \u2014 committing today to purchase a vaccine that doesn’t yet exist, at a price high enough to make development commercially rational, contingent on successful development. This theoretical construct exists in policy literature under various names: advance market commitments for development, pull financing, innovation inducement prizes. What it has not done is materialize as a funded mechanism for any neglected outbreak pathogen at the scale that would change the expected-NPV calculation from negative to positive. The guarantee would need to promise purchase at prices substantially above GAVI humanitarian levels, for a product that may never clear regulatory approval, from manufacturers that haven’t started the development process. What that guarantee looks like, in practice, is a government contract with AMC language layered on top of it.<\/p>\n\n\n\n
CEPI was designed for exactly this gap. Founded at Davos in January 2017, directly after the 2014\u20132016 West Africa Ebola epidemic made vivid what the absence of preemptive development infrastructure actually costs, CEPI began with approximately $460 million in commitments and a mandate to accelerate vaccine development for high-risk outbreak pathogens with no viable commercial market. It has done real work: CEPI-funded programs have advanced Lassa candidates, Nipah candidates, a Chikungunya vaccine now approved in multiple jurisdictions. Sudan ebolavirus candidates moved rapidly through CEPI’s pipeline during Uganda’s 2022 Sudan ebolavirus outbreak and were available for emergency deployment. The accelerated timelines CEPI has achieved for several pathogens are genuinely significant compared to the uncoordinated baseline.<\/p>\n\n\n\n
CEPI 2.0, covering 2022\u20132026, had a fundraising target of $3.5 billion. It received approximately $2.5 billion \u2014 $1 billion short of its own stated goal. This shortfall arrived during a period in which the COVID-19 pandemic had made the case for epidemic preparedness as forcefully as anything in modern public health history. Every government that faced the COVID emergency had, by 2021, experienced in real time what happens when novel pathogens spread through a world without pre-positioned vaccine development infrastructure. They were then asked to fund CEPI at $3.5 billion. They sent $2.5 billion. CEPI 3.0 is now seeking $2.5 billion additional investment for 2027\u20132031.<\/p>\n\n\n\n
When WHO declared a PHEIC for Bundibugyo in 2026, no Bundibugyo vaccine candidate was in clinical development. CEPI, which exists specifically to prevent this situation, didn’t have a Bundibugyo program. This is not a failure of design or organizational intent. It is the direct, mechanical consequence of chronic underfunding: with a budget that funds roughly two to three complete development programs at average cost, and eight WHO emergency-priority disease groups to cover, CEPI has to prioritize. Prioritization means making bets \u2014 Lassa over Bundibugyo, Sudan ebolavirus over Crimean-Congo haemorrhagic fever \u2014 on which pathogens are most likely to drive the next emergency. Bundibugyo, with its sporadic outbreak history and small confirmed case counts, didn’t make the cut before 2026. The $3.2 million emergency award to IAVI after the PHEIC was declared is reactive vaccine development \u2014 exactly the pattern CEPI was built to replace.<\/p>\n\n\n\n
Starting from zero clinical development, “as fast as it can” is measured in years, not weeks. That is what chronic underfunding means in practice when an outbreak arrives.<\/p>\n\n\n\n
Priority review vouchers are the third mechanism, and they are genuinely clever \u2014 which is worth acknowledging before explaining why that cleverness doesn’t close the structural gap. The PRV program, created by Congress in 2007, awards FDA vouchers to pharmaceutical companies that gain regulatory approval for drugs or vaccines targeting qualifying neglected tropical diseases. The voucher converts a standard FDA review \u2014 approximately ten months \u2014 to a priority review of approximately six months, saving roughly four months of time-to-market for whatever product the holder applies it to. Vouchers are transferable and have sold at auction for between $67 million and $350 million, according to GAO analysis.<\/p>\n\n\n\n
The theory is that this creates a tradable asset from FDA review capacity. A small company develops a neglected-disease drug, earns a voucher, sells it to a large company racing to launch a new product, uses the cash to fund its next neglected-disease program. Private incentives, no direct public expenditure, market mechanism.<\/p>\n\n\n\n
The practice is more constrained. PRVs are awarded at regulatory approval, meaning the entire development investment \u2014 Phase 1, Phase 2, Phase 3, regulatory submission \u2014 must be committed upfront before any voucher benefit materializes. The voucher doesn’t change the expected-NPV calculation at the moment that actually matters, which is before Phase 1, when companies decide whether to begin development. A voucher worth $67 million to $350 million is not an adequate return on a program that costs $886.8 million to complete and whose end market cannot support commercial price recovery. Most PRVs have been earned for drugs, not vaccines \u2014 vaccines cost more to develop, require more extensive clinical trials for a narrower population of eligible diseases, and have a lower probability of earning a voucher per development dollar spent. And the secondary market for vouchers only functions because the voucher value is modest relative to what large companies earn from faster launches: the mechanism creates value by selling something cheap. The program has produced some additional neglected-disease approvals at the margin. It has not changed the structural dynamics of vaccine development for outbreak pathogens with negative expected NPV.<\/p>\n\n\n\n
Any mechanism that still routes through private expected return as its organizing principle will produce the same result the market produces \u2014 a negative expected return \u2014 because the disease characteristics that make the market fail don’t change when you design a cleverer instrument. The poverty of the affected population doesn’t change. The government procurement constraints don’t change. The epidemiology that makes Phase 3 trials logistically near-impossible doesn’t change. These are not problems that better financial engineering can redesign around.<\/p>\n\n\n\n
When you try to scale any of these mechanisms to the point where they actually close the gap, something happens: they become their opposite. A pull AMC large enough to make Bundibugyo vaccine development commercially rational is structurally indistinguishable from a government procurement contract \u2014 it is a government commitment to buy a future product at a guaranteed price, which is what government contracts are. CEPI funded to cover all eight WHO emergency-priority disease groups simultaneously is a public vaccine-development agency, accountable to state funders, making allocation decisions on public health grounds rather than expected commercial return, producing vaccines as public goods. A PRV sized to offset an $886 million negative-market development program would need to sell at roughly $900 million or more, at which point the secondary-market arbitrage that is the mechanism’s entire functional logic has collapsed \u2014 no company pays $900 million for four months of regulatory time savings.<\/p>\n\n\n\n
These mechanisms don’t scale into more efficient versions of themselves. They scale into public finance dressed in market language.<\/p>\n\n\n\n
There is no market solution. Not because markets are malicious, but because the structural characteristics of these diseases \u2014 the outbreak populations, the purchasing power, the trial logistics \u2014 cannot be altered by instrument design. A mechanism sufficient to close the gap is, by definition, not a market mechanism anymore.<\/p>\n\n\n\n
The arithmetic of a real solution<\/h3>\n\n\n\n
The WHO’s “priority diseases for R&D in emergency contexts” \u2014 its canonical emergency preparation list \u2014 identifies eight named disease groups requiring urgent vaccine and treatment development: COVID-19; Crimean-Congo haemorrhagic fever; Ebola and Marburg viruses; Lassa fever; MERS-CoV and SARS; Nipah and henipaviral diseases; Rift Valley fever; and Zika. Plus Disease X \u2014 the placeholder for the next novel pathogen, unknown identity, confirmed threat. Each disease group needs at least one complete development program: preclinical research, clinical trials through Phase 3, regulatory approval, manufacturing scale-up, stockpile.<\/p>\n\n\n\n
CEPI 2.0 received approximately $2.5 billion over five years. At the ASPE average development cost of $886.8 million per program, CEPI’s entire five-year budget funds roughly two to three complete programs from preclinical research to regulatory approval \u2014 before administrative overhead, before program failures, before the organizational infrastructure required to run an international public health institution. There are eight disease groups on the WHO emergency list. CEPI, at its actual funding level, covers at best three of them.<\/p>\n\n\n\n
Global pharmaceutical R&D spending in 2024 was approximately $288 billion. CEPI’s annual budget \u2014 roughly $500 million in a good year \u2014 is less than two-tenths of one percent of what the pharmaceutical industry spends annually, almost entirely on diseases affecting wealthy or middle-income populations.<\/p>\n\n\n\n
Two mechanisms can close this gap. Both are explicitly non-market.<\/p>\n\n\n\n
The first is regulatory mandate: governments requiring pharmaceutical companies, in exchange for the patent protections and market exclusivities that underpin the entire pharmaceutical business model, to invest some defined share of R&D in neglected disease development. The structural logic is coherent \u2014 companies receive an enormous public subsidy in the form of 20-year monopoly rights enforced by state power, and attaching public interest conditions to that subsidy is reasonable. The politics are not sound, at least not at present. The TRIPS Agreement \u2014 the WTO treaty framework governing pharmaceutical intellectual property, negotiated under sustained pharmaceutical industry lobbying pressure \u2014 protects patent rights aggressively and has been designed with public health exceptions narrow enough to require years of litigation to invoke. The COVID-19 pandemic produced a TRIPS waiver for vaccine production, adopted at the WTO’s 12th Ministerial Conference in June 2022 \u2014 a partial, hard-won concession achieved over two years of opposition. India and South Africa first proposed a broad COVID-related IP waiver in October 2020. The United States, the European Union, the United Kingdom, and Switzerland opposed it. The US reversed its position in May 2021; the EU held out for another year. The final text covered vaccine production only, excluded diagnostics and therapeutics, and was substantially narrower than originally proposed \u2014 the pharmaceutical industry characterized it as acceptable precisely because it didn’t require technology transfer or shared manufacturing know-how. A mandatory neglected-disease R&D requirement faces the same opposition \u2014 more organized, better resourced, and without even the COVID-scale emergency as political cover.<\/p>\n\n\n\n
The second mechanism is public subsidy at scale: governments funding vaccine development directly through public agencies, owning the resulting intellectual property, and producing vaccines as public goods priced for access rather than return on private investment.<\/p>\n\n\n\n
This is not a theoretical construct. It has been demonstrated once, under conditions that generated the political will to deploy it.<\/p>\n\n\n\n
Operation Warp Speed committed approximately $18 billion of US government funding to COVID-19 vaccine and treatment development. The government absorbed clinical trial risk by funding Phase 3 trials directly. It contracted with Moderna, Pfizer\/BioNTech, Johnson & Johnson, AstraZeneca, Novavax, and others for hundreds of millions of doses before any of those vaccines had demonstrated efficacy in humans \u2014 essentially pre-paying for products that might never work, so that manufacturing capacity could be built in parallel with clinical development. It paid for failures: AstraZeneca’s US rollout stumbled, Novavax’s timeline slipped, others were never deployed. Multiple vaccines were licensed in under a year \u2014 the fastest vaccine development program in history by a substantial margin. The model works. The science knows it. The policy apparatus knows it.<\/p>\n\n\n\n
It has been deployed for exactly one disease. That disease killed people in the United States, the United Kingdom, Germany, France, and Japan at a rate that put politicians facing re-election in those countries under direct electoral pressure. COVID-19 threatened populations whose governments had both the budget and the political motivation to respond \u2014 governments that faced accountability at the ballot box for the bodies accumulating in hospital car parks. The same regulatory infrastructure, the same emergency procurement authorities, the same fast-track pathways, the same public health agencies exist now and could be pointed at Bundibugyo, Lassa, Nipah, Crimean-Congo haemorrhagic fever, Rift Valley fever. The WHO’s priority list is not a secret document. It has been published, updated, and cited in global health policy discussions for years. The risk information was available. The development model was available.<\/p>\n\n\n\n
What has been absent \u2014 consistently, across multiple outbreaks, across multiple administrations in multiple countries \u2014 is political motivation. And political motivation tracks with a precision that the polite version of this story declines to name: it tracks whether the threatened population has governments with both the budget and the electoral incentive to treat the outbreak as an emergency.<\/p>\n\n\n\n
DRC does not have that lever. Uganda has it only partially. The governments of the countries where Bundibugyo kills at a 25\u201340% case fatality rate do not have the fiscal capacity to mount an Operation Warp Speed for a disease that threatens primarily their own populations. And the governments that do have that capacity \u2014 the G7, the major bilateral donors, the multilateral institutions \u2014 have no electoral constituency pressing them to act with anything like the urgency they displayed for COVID-19.<\/p>\n\n\n\n
The gap between what was spent on COVID-19 and what has been spent on the WHO’s other seven named emergency-priority disease groups is not a budgetary accident. It reflects a consistent pattern: the same institutions that moved at unprecedented speed to fund COVID vaccines have been serially asked to fund CEPI at its stated goal and have serially come up short. The failure is not that political systems cannot respond to pandemic risk. They can, demonstrably. The failure is that “pandemic risk” in practice means “risk to populations whose governments have the resources and the political incentive to treat them as worth an emergency response.” Bundibugyo does not produce those governments. Lassa does not produce those governments. Nipah does not produce those governments.<\/p>\n\n\n\n
This is not an inference drawn from charitable assumptions about how governments work. It is a structural observation, verifiable across decades of outbreak response data. The same governments, the same institutions, the same mechanisms \u2014 applied differently to different populations, producing predictably different outcomes.<\/p>\n\n\n\n
The word for that is not complexity. It is priority.<\/p>\n\n\n\n
In June 2026, CEPI awarded $3.2 million to IAVI for an rVSV-based Bundibugyo vaccine candidate. The rVSV platform had been available for years. The award came after a PHEIC was declared, for a virus known since 2007, characterized in detail in scientific literature that anyone could read. The $3.2 million is roughly 0.36 percent of the cost of a full development program. The same governments that assembled $18 billion for Operation Warp Speed \u2014 the same budget processes, the same procurement authorities, the same public health agencies \u2014 have not assembled anything comparable for Bundibugyo, or Lassa, or Nipah, or the other pathogens on the WHO’s own priority list.<\/p>\n\n\n\n
The distinction between “can’t” and “won’t” is the precise nature of this failure. The death toll from Bundibugyo in 2026 is verifiable. So is the decision that produced it. Both of those facts will be true, in the same relationship to each other, when the next pathogen on the WHO list arrives.<\/p>\n\n\n\n