Pandemics and Biosecurity
What pandemic risk looks like, what biosecurity infrastructure exists, and the lessons mostly unlearned from COVID.
(the WHO cites more than $10T for 2020-23; ranges are wide)
(the WHO estimates more than 20 million excess deaths for 2020-23)
(Metabiota modeling cited by the Center for Global Development; contested)
A note on framing. Pandemics largely vanished from political attention after the COVID emergency phase ended in 2022-23, even though most of the conditions that made COVID possible have not changed. An Ebola outbreak in the Democratic Republic of the Congo has been a WHO public health emergency since May 17, 2026, with little attention outside the region. This page walks through the picture as it stands in late September 2026: the threat profile, what biosecurity infrastructure exists and does not, which lessons were learned, and where the debates sit now. The tone is neither alarmist nor reassuring; the record is sobering enough that calm description is more useful than either.
What the threat looks like
Pandemic risk is not one threat. It is a category of related threats with different sources, time profiles, and signatures. Distinguishing them clarifies what each requires.
Natural zoonotic spillover. Most large historical pandemics, including COVID-19 in the natural-origin hypothesis, the 1918 influenza pandemic, HIV/AIDS, MERS and SARS, originated in animals and crossed into humans through close contact, food chains, or environmental disruption. The frequency of spillover has been increasing as human encroachment into wildlife habitats grows, factory-farm scale expands, and climate change shifts disease ecology. The question is not whether more spillovers will happen but which will go pandemic.
Influenza specifically. Avian influenza (H5N1 in particular) is the threat most carefully tracked by public-health agencies. The virus has been circulating in poultry and wild birds for decades; recent years have seen substantial spread in mammals (mink farms, dairy cattle, marine mammals, including cases that suggest mammal-to-mammal transmission). As of late September 2026 it has not adapted to efficient human-to-human transmission, and the US CDC rates the risk to the public as low, but the evolution that would change this is closely watched. A flu pandemic with case-fatality anything like that seen in human H5N1 cases would be far worse than COVID-19.
Lab-related risks. Research on dangerous pathogens has expanded substantially since 2000. Gain-of-function research, where pathogens are deliberately enhanced to study transmissibility or virulence, has been the subject of intense scientific and policy debate. The 2014 US moratorium, the 2017 reversal, and continued ambiguity about what counts as "gain of function" left the framework less robust than it appeared on paper. A US executive order of May 2025 paused federally funded work it called dangerous gain-of-function research; the replacement policy of July 28, 2026 bars federal money for it and sends a wider category to independent review. Lab leaks have happened historically, including in highly secure facilities. Whether COVID-19 had a lab-related origin remains contested; the policy-relevant point is that the risk of lab leaks is real either way.
Deliberate biological threat. Bioweapons programs are formally banned by the 1972 Biological Weapons Convention, which still has no verification protocol and so is much weaker on that count than the nuclear or chemical regimes. Several states accuse others of maintaining covert programs; non-state capability is harder to assess but is rising as biotechnology becomes cheaper. AI-assisted protein design and gene-synthesis services, industry breakthroughs both, are also capability uplifts for would-be biological weaponeers.
Emerging and re-emerging diseases. Antimicrobial resistance (AMR) is producing slow-moving but substantial damage as bacteria and fungi develop resistance to existing drugs. Ebola, Marburg, Nipah, Lassa, mpox and dengue have produced regional outbreaks that did not become pandemics but could in different conditions. The Ebola epidemic in the Democratic Republic of the Congo, running since May 2026, is the largest recorded there: 7,890 confirmed cases and 3,799 deaths by September 23. Diseases vaccines had pushed back are returning too: the Americas lost their measles-elimination status in November 2025 and recorded 47,459 cases by mid-July 2026, the most since 2004. Climate change is shifting vector-borne disease ranges.
What biosecurity infrastructure exists
The international biosecurity architecture has been built across decades and remains incomplete relative to the threat. The components:
The World Health Organization. Coordinates global disease surveillance, declares Public Health Emergencies of International Concern (PHEIC), and supports lower-capacity countries. Its authority is persuasive rather than binding, and the Pandemic Agreement adopted in May 2025 states that it gives the WHO no power to order lockdowns, vaccination or travel bans. The agreement is not in force: it opens for signature only once the annex on pathogen access and benefit-sharing is adopted, and takes effect 30 days after the sixtieth ratification. That annex missed the May 2026 assembly; talks ran on through September, with more rounds set for November and December, and adoption is now aimed at May 2027. The United States left the WHO on January 22, 2026, and the WHO's two-year base budget was cut to $4.27 billion, about 14% below the previous period's $4.97 billion.
National public-health agencies. US CDC, the UK's Health Security Agency, Germany's Robert Koch Institute, the Chinese Center for Disease Control and Prevention, and equivalents elsewhere. Capacity varies widely; COVID exposed substantial weaknesses in nearly every major agency, and post-COVID reforms have been uneven. The most visible upheaval has been in the United States: four CDC directors or acting directors since mid-2025, and a childhood immunization schedule narrowed in January 2026 from 18 diseases routinely recommended for all children to 11, modeled partly on Denmark's. The administration presents this as matching recommendations to risk and rebuilding trust; the American Academy of Pediatrics broke with the CDC and kept its own broader schedule, arguing US disease risk is not Denmark's. A federal judge stayed the new schedule in March 2026 as likely unlawful; the government has appealed, and an August executive order called for a new one within 90 days.
Disease surveillance systems. Genomic sequencing of pathogens has expanded dramatically since 2020. The GISAID database, the Africa CDC pathogen genomics initiative, and similar efforts allow real-time monitoring of how viruses are evolving. Wastewater surveillance has scaled in many countries. Early detection has substantially improved; the political will and the money to act on it have not. Official development assistance fell 23% in 2025 on the UN's figures, and the WHO says the abrupt aid cuts have disrupted health systems in many countries.
Biosafety laboratory networks. Biosafety Level 3 and Level 4 laboratories, where the most dangerous pathogens are studied, are operated globally, and their safety records have been imperfect across decades. Accidents and near-misses (smallpox in Birmingham 1978, anthrax at Sverdlovsk 1979, SARS releases in Asia in 2003-04) document that even well-resourced facilities fail. The number of high-containment labs has expanded significantly since 2000.
Vaccine and therapeutics infrastructure. mRNA vaccine technology, which enabled the rapid COVID-19 vaccine development, is a real biosecurity advance. The 100-day vaccine target promoted by CEPI (Coalition for Epidemic Preparedness Innovations) is technically plausible for some pathogen categories. US federal support has moved the other way: in August 2025 the Department of Health and Human Services wound down 22 mRNA vaccine projects worth about $500 million, including an H5N1 award, saying the platform protects poorly against upper-respiratory infections. Researchers in the field dispute that reading. Manufacturing capacity remains concentrated in a few facilities, and equitable distribution remains under-resourced.
Stockpiles. The US Strategic National Stockpile and equivalents elsewhere. COVID exposed that many had been depleted and not replenished; rebuilding has been partial.
The COVID lessons mostly unlearned
Few of the lessons identified during COVID have produced durable institutional change. The picture varies by category.
Lessons partially learned. mRNA platform technology has expanded substantially. Wastewater and genomic surveillance has scaled. Several clinical-trial protocols (RECOVERY trial, others) demonstrated that rapid evidence generation is possible.
Lessons identified but not implemented. Improved international coordination through binding agreements (the Pandemic Agreement cannot be signed until its annex is agreed, and needs 60 ratifications after that). Real-time transparent data sharing across borders (still depends on national choices). Adequate stockpiling of medical supplies (partial rebuild only). Sustainable funding for the WHO and disease-surveillance networks (its base budget was cut by about 14% after its largest contributor left). Border-and-quarantine protocols that can be activated quickly (politically difficult). Communication that maintains public trust during sustained emergencies (worse since 2020).
Lessons rejected or reversed. Public-health agencies' ability to issue mandates has been constrained in several jurisdictions through court rulings, legislative changes, and political pressure. Non-pharmaceutical interventions (masking, distancing, school closures) are politically radioactive in ways that make their future use much harder. The UK Covid-19 Inquiry pushed the other way in November 2025, finding that England's first lockdown came a week late at a cost of about 23,000 lives. The investigation of pandemic origins has been compromised by polarization that makes calm scientific assessment difficult.
Lessons made worse. Trust in public-health institutions has declined substantially, and vaccine confidence is lower than pre-pandemic in many countries, including for routine childhood vaccines - the measles figures above are what that looks like in practice. The information environment around health is more polarized, and institutional capacity in several major countries is lower than before COVID.
The pattern is that the lessons that required uncomfortable institutional change or sustained political attention have not been learned, while the lessons that required scientific or technological work have made real if incomplete progress. This is not a stable equilibrium for the next pandemic.
The H5N1 question
H5N1 avian influenza deserves separate attention because it is the threat that pandemic researchers are watching most closely as of late September 2026.
The basic facts: H5N1 has been circulating in poultry and wild birds for decades. Since 2020 an unusually transmissible variant (clade 2.3.4.4b) has spread globally, killing tens of millions of birds. It has increasingly infected mammals - mink farms in Spain (2022), seal die-offs across multiple coastlines, dairy cattle in the United States (substantial outbreak from 2024), and individual cases in dogs and cats. Human cases have been few and severe: 1,003 confirmed cases and 479 deaths reported to the WHO since 2003, a case-fatality ratio of about 48%, though those are severe presentations with limited testing of milder ones. Ten cases and two deaths were reported in 2026 to early September, all in Asia.
What would change everything is efficient human-to-human transmission. The genetic mutations needed for this are known; whether and when they will occur in nature is uncertain. The 2024 dairy-cattle outbreak in the United States raised the surveillance level because dairy work involves close contact and opportunities for adaptation. It produced most of the 71 US human cases recorded since 2024, one of them fatal; none became clusters, and none has been reported in 2026. Credible researchers disagree about the probability that this virus produces a pandemic in the next several years.
What is well-established is that a flu pandemic with case-fatality substantially above seasonal flu would be much worse than COVID-19. Flu spreads faster; respiratory infectivity in pre-symptomatic carriers is high; existing flu infrastructure (vaccine production, antivirals) is more developed than for novel viruses but far from adequate for a major pandemic. The stockpile of pre-pandemic H5N1 vaccines is partial and uneven across countries, and the United States canceled its mRNA H5N1 vaccine work in 2025.
The paths from here
Continued elevated baseline risk without major event
Surveillance improves slowly. Periodic regional outbreaks test response capability without becoming global pandemics. The Ebola emergency in the Democratic Republic of the Congo is the current test, fought with a vaccine unproven against this species and, the WHO says, too few responders. The H5N1 watch continues. No major novel pandemic for several years, and the institutional drift continues without reform pressure.
A second major pandemic before 2035
Published estimates cluster in the 20-40% range for an event of comparable scale to COVID before 2035. Metabiota's modeling, cited by the Center for Global Development, implies roughly a quarter over a decade; the forecaster Airfinity put it at 27.5% by 2033. Against the current degraded baseline, the response would likely be worse than 2020-22.
H5N1 specifically goes pandemic
The most-watched scenario. If the virus adapts to efficient human transmission while retaining substantial virulence, the resulting pandemic would be qualitatively worse than COVID. Pre-positioned vaccines and antivirals would help; the time pressure on scaling production would be severe.
Antimicrobial resistance produces slow crisis
Rather than acute pandemic, the slow-moving AMR threat materially erodes the effectiveness of existing antibiotics, antivirals, and antifungals. Surgical procedures, cancer chemotherapy, and standard infections become more dangerous. This is already happening; the trajectory may accelerate.
A lab-related event reshapes biosecurity politics
A confirmed lab leak, biological-research accident, or deliberate event produces sustained political attention to research-and-laboratory biosecurity. Specific reforms (oversight of dual-use research, gene-synthesis screening, export controls on dangerous pathogens) become globally coordinated. The United States tightened its funding rules in July 2026 without waiting for an incident; no binding international standard exists.
Where serious analysts disagree
Pandemic risk is the most under-prepared major threat
Comparison with the resources spent on much smaller threats (terrorism in particular) shows that pandemic preparedness has been chronically underfunded relative to expected harm. The COVID experience should have closed this gap; the political economy of post-emergency funding has not let it. Without major reform, the next pandemic finds the world inadequately prepared.
Held by: Bill Gates ("How to Prevent the Next Pandemic"), parts of the Center for Global Development, the Nuclear Threat Initiative's biosecurity program, and the broader pandemic-preparedness community. The case has empirical support and policy traction in some jurisdictions.
Biosecurity restrictions on legitimate research are themselves a risk
Restrictions on pathogen research, gain-of-function studies, and related work risk reducing the basic-science capability that future responses depend on. The right balance involves more transparency and oversight, not blanket restriction. On this reading the US funding ban of July 2026, and the ending of federal mRNA vaccine work, are net negative for long-run preparedness.
Held by: parts of the virology and infectious-disease research community, several major science journals' editorials, and a strand of biosecurity-policy thinking. The case is contested; the empirical question of how much basic-science capability matters for response is debated.
Lab-leak hypotheses deserve serious investigation
The political polarization of the COVID-origin question has obscured that lab-related events are a real category of pandemic risk whatever COVID-19's origin was. In the declassified US assessments the FBI (moderate confidence), the Department of Energy and, from January 2025, the CIA (both low confidence) lean toward a lab-related origin; four other agencies and the National Intelligence Council lean toward natural zoonotic origin, also with low confidence. No agency has reached high confidence, and the June 2026 declassification added documents but no new judgment. Germany's foreign intelligence service put a lab accident at 80-95% in a 2020 report revealed by German newspapers in 2025. The WHO's expert group concluded in June 2025 that the weight of available evidence supports zoonotic spillover, while saying all hypotheses must remain on the table.
Held by: Alina Chan and Matt Ridley ("Viral"), parts of the US intelligence community, the French Academy of Medicine in a 2025 report, several biosecurity analysts. Critics including virologists Kristian Andersen and Robert Garry argue natural spillover remains better supported by the genomic and epidemiological evidence. The question is not settled either way, and the early sequences and laboratory records the WHO has asked China for have not been released.
Antimicrobial resistance is the bigger long-term threat
Acute pandemic response gets political attention. The slow accumulation of antimicrobial resistance is producing larger aggregate harm and is much harder to reverse. The agricultural use of antibiotics, the absence of new antibiotic development, and the difficulty of paying for antibiotics that are only used as backstop are all worsening the trajectory.
Held by: the antimicrobial-resistance research community (Jim O'Neill's UK review, the WHO's antimicrobial-resistance work, parts of the pharmaceutical-economics literature). The case has empirical support and political marginality in roughly equal measure.
Public health overreached and lost trust
Specific decisions during COVID (school closures of contested necessity, vaccine mandates beyond the highest-risk groups, communication strategies that overstated certainty) damaged public-health credibility in ways that will compromise the next response. The right correction is more humility about what is known, more honest acknowledgment of trade-offs, and less reliance on coercive tools.
Held by: Vinay Prasad (ran the FDA's biologics center in 2025-26, now back at UCSF), Jay Bhattacharya (NIH director; acting CDC director for part of 2026), parts of the heterodox-pandemic-policy community. The case is contested. Some claims have been validated by post-COVID review; the UK inquiry found England's first lockdown came too late rather than too readily.
What the available evidence actually supports: pandemic risk is real and substantial, and the post-COVID institutional response has been disappointing relative to the lessons identified. Technology has improved but has not been matched by political and institutional progress. H5N1 deserves close attention; lab-related risks are real regardless of COVID-19's origin; AMR is a slower but possibly larger threat. The trust environment is worse than in 2019, which will make any future response harder.
What this means for you
Personal preparedness
Modest practical preparation that does not require alarm. Stay current on routine vaccines, including seasonal flu (which affects pandemic-flu severity meaningfully). Keep a small supply of N95 or equivalent respirators on hand. Know what local public-health resources look like before you need them. Maintain enough non-perishable food and water for two weeks. None of it requires belief in a specific scenario; all of it is the kind of preparation that matters when something does happen.
If you are at heightened personal risk
Older adults, people with immune compromise, pregnant women and healthcare workers benefit from more attention to seasonal respiratory-virus dynamics, prompt antiviral access, and conservative behavior during local surges. Specific vaccines (RSV, updated COVID, pneumococcal) are widely available and underused, though which ones are routinely recommended changed in the United States in January 2026.
If you read pandemic-related coverage
Specialized sources (Bulletin of the Atomic Scientists biosecurity coverage, Center for Global Development, The Atlantic's pandemic coverage, the Global Health Security Index, several substack-based experts) have been consistently more accurate than mainstream coverage. Mainstream media oscillation between alarm and reassurance has been particularly poor on this topic. Reading scientific source material when major claims are made is more useful than usual, and the sources are accessible: the WHO publishes outbreak reports and monthly H5N1 case tables.
If you work in healthcare or public health
Maintaining institutional and personal capacity for the next major event, under degraded public trust, shrinking aid budgets and political pressure, is harder than it was in 2019. The work matters; the credit for it is currently low. Keep records, maintain professional networks, and protect your own resilience.
If you vote on biosecurity policy
The unglamorous questions matter most. Public-health agency funding and structure. Pandemic Agreement negotiating positions, and whether your country ratifies once the text opens for signature. Gain-of-function research oversight. Stockpile rebuilding. Lab biosafety. AMR drug-development incentives. These are not headline issues; they are where the actual preparedness picture is being shaped. Politicians who engage them substantively are doing the work; those who treat pandemic policy as ordinary partisan material usually have not.
