Bill Gates Says a Man-Made Virus Could Be Our Next Big Crisis

In January 2026, Bill Gates wrote in his annual letter. He didn’t describe the next pandemic as a distant possibility or a risk to be modeled. He said a non-government group using open-source AI tools to design a bioterrorism weapon was not just possible – it was, in his word, coming.

Gates has been raising alarms about infectious disease for over a decade. In a 2015 TED Talk, he warned that “if anything kills over 10 million people in the next few decades, it’s likely to be a highly infectious virus rather than a war.” That turned out to be correct. COVID-19 was a painful demonstration of just how exposed a world of increased urbanization and global mobility really is to a rapidly spreading pathogen.

His warnings have increasingly centered on a different kind of threat: bioterrorism. The man-made virus threat, in his view, may now be the more urgent concern – and the combination of advanced gene-editing tools and widely available artificial intelligence has fundamentally changed the risk equation.

What Gates Actually Said About a Man-Made Virus Threat

Gates wrote in his January 2026 annual letter: “In 2015, I gave a TED talk warning that the world was not ready to handle a pandemic. If we had prepared properly for the Covid pandemic, the amount of human suffering would have been dramatically less. Today, an even greater risk than a naturally caused pandemic is that a non-government group will use open source AI tools to design a bioterrorism weapon.“

This wasn’t his first time flagging the bioterrorism risk. In an earlier interview on the Veritasium YouTube channel, Gates said: “Also related to pandemics is something people don’t like to talk about much, which is bioterrorism, that somebody who wants to cause damage could engineer a virus.” He went further in an interview with the BBC, telling journalist Amol Rajan that because the world was starting to think enough about climate change, the big underappreciated threat was “bioterror, which is really awful.”

Gates did not rule out the possibility that the next pandemic would be man-made and released as an act of bioterrorism, noting “that’s a very scary scenario because they could try to spread it in different places at once.”

Gates told The Telegraph that a bioterrorist attack using a contagious virus like smallpox could kill more people than a nuclear weapon, stating plainly that “bioterrorism is a much larger risk than a pandemic” and that “all these advances in biology have made it far easier for a terrorist to recreate smallpox, which is a highly fatal pathogen, where there is essentially no immunity remaining at this point.”

Why AI Makes the Man-Made Virus Threat Different Now

For most of history, bioterrorism required experts and large labs. AI-enabled bioweapons might now need just a computer and bad intent.

Gates wrote that AI’s change to society could be ultimately for the better, but it also poses considerable threats to the global population – particularly if it falls into the hands of bad actors. He drew a direct parallel to his 2015 pandemic preparedness warning, suggesting the world is similarly unprepared for this AI-enabled threat.

This concern is grounded in active research. A 2024 RAND research report determined that either the deliberate release of an engineered pathogen by a state or bioterrorist as a weapon, or the inadvertent release of an engineered pathogen from a research laboratory, could result in a synthetic pandemic. Researchers assessed the risk of a nefarious actor developing such a pathogen using available technology, looking at threat timelines spanning three, five, and ten years out.

A separate 2025 RAND Delphi study surveying experts in biology and AI found that “the risks posed by AIxBio are shifting and could greatly increase beyond just enabling bioterrorism.” Concerns that AI might enable pathogen design are increasing, though risks and timelines remain unclear. Crucially, no strong fundamental limit to AI’s capabilities in this area was identified. The practical implication: the barrier to entry may drop low enough that the threat multiplies across a much wider range of actors, not just well-resourced states.

Among national security professionals and some technologists, there is growing anxiety that AI may dramatically lower the threshold for malicious actors to construct destructive biological weapons – with machine learning technologies potentially enabling terrorists to increase the virulence of existing pathogens or develop blueprints for entirely novel ones.

The Probability Problem

Gates has noted that in a globalized society where intercontinental travel is routine, a pandemic could prove even more deadly than the 1918 influenza outbreak, which killed between 50 and 100 million people.

Against that backdrop, probability estimates deserve attention. Gates authored the book How to Prevent the Next Pandemic in 2022, and by January 2025, he was putting specific numbers on the risk – estimating a 10 to 15 percent chance of a natural pandemic hitting within the next four years, and suggesting that another pandemic in the next 25 years was likely unless a major war intervened. Speaking at the Time 100 Summit in 2022, Gates put the odds of another global pandemic arising in the next two decades at an even 50-50.

What COVID actually cost the world makes these figures sobering. The COVID-19 pandemic has resulted in an estimated 7.1 million confirmed deaths worldwide. A 2022 WHO analysis found the true toll was far higher: excess deaths directly or indirectly associated with COVID-19 between January 2020 and December 2021 reached approximately 14.9 million. The pandemic also pushed tens of millions into extreme poverty, with economic losses estimated in the trillions.

Gates’ Proposed Solution: A Global GERM Team

Gates has not stopped at diagnosis. His proposed fix is a standing global team dedicated to pandemic detection and response, modeled partly on how militaries prepare for war. He advocates for a Global Epidemic Response and Mobilization (GERM) team, under the mandate of the WHO, that would require $1 billion a year to employ a staff of 3,000 to coordinate pandemic surveillance and preparedness at a global level. Despite his criticisms of WHO, Gates would still make the agency responsible for GERM, with a “special personnel system” to attract the best staff possible, and the authority to declare a pandemic and coordinate the global response.

Gates estimates the cost of GERM would be $1 billion annually, with the U.S. likely contributing around $250 million of that. The same early-detection and rapid-response infrastructure that catches a naturally emerging virus could, in principle, catch an engineered one before it spreads globally.

Gates has identified two key issues arising from AI: the first being artificial intelligence being used by bad actors, and the second being disruption to the jobs market. “Both are real risks that we need to do a better job managing,” he wrote, adding: “We’ll need to be deliberate about how this technology is developed, governed, and deployed.” On the bioterrorism side, that call for deliberate policy reflects a global governance layer that doesn’t yet exist in any meaningful form.

The Gates Foundation has the resources to push this agenda. Since 2000, the Foundation has distributed more than $100 billion in grant funding to thousands of organizations worldwide. It currently provides 45% of the WHO’s NGO funds, equivalent to 12% of the WHO’s total operating expenditure.

The Funding Gap Opening Up Now

The challenge is that pandemic preparedness spending is moving in the wrong direction. The U.S., United Kingdom, France, Germany, and others all reduced their global health and research and development funding in 2025 – at precisely the moment that the infrastructure required to catch an engineered virus early demands consistent, long-term investment rather than the boom-bust cycle the world has run through since COVID.

In May 2025, President Donald Trump signed an executive order restricting gain-of-function research on pathogens with pandemic potential. Gain-of-function research refers to experiments that deliberately enhance a pathogen’s abilities, such as making it more transmissible or more resistant to treatment, in order to study it and develop countermeasures. Gain-of-function experiments, which involve changing genetic material to induce new or enhanced capabilities in microbial organisms, can help researchers better understand human-pathogen interactions and drug resistance, as well as prepare for future pandemics and develop countermeasures under controlled conditions. The policy debate around that executive order reflects the genuine tension at the heart of this issue: the same research that helps scientists prepare vaccines and treatments can, theoretically, provide a roadmap for engineering a dangerous virus.

On the global funding side, some progress is happening. The Pandemic Fund has allocated $1.4 billion through three funding rounds, catalyzing over $10 billion in additional resources across 128 countries – but that is a fraction of the trillion-dollar cost of the next pandemic if prevention fails.

What to Do Now

Policymakers should increase and sustain investments in core pandemic preparedness efforts, such as rapid diagnostics, scalable vaccine platforms, and public health surge capacity. Stakeholders should cultivate a culture of responsible AI use in biological research, particularly among developers and institutions deploying increasingly capable tools.

A country does not suddenly invent epidemiological reporting, oxygen delivery, or vaccine logistics when a crisis begins – it relies on capacities that had to exist beforehand. The same logic applies at the household level. Staying current on recommended vaccinations, maintaining basic emergency supplies, and following credible public health sources rather than social media during an outbreak are concrete actions that don’t require waiting for governments to act. Vaccine hesitancy and misinformation have already been identified as a major barrier to reinvestment in pandemic research and development – meaning public trust in the system directly shapes how well the system can protect everyone.

Researchers at the Center for a New American Security have concluded that “AI-enabled biological catastrophes, though daunting, are far from inevitable” – a conditional framing that acknowledges genuine risk while resisting the assumption that the worst outcome is locked in. Whether global institutions build the surveillance and response infrastructure in time depends almost entirely on decisions made now, while the window is still open.

AI Disclaimer: This article was created with the assistance of AI tools and reviewed by a human editor.