Executive Summary
On 8 July 2026, the Atlantic Council’s Bipartisan Commission on Biodefense – a group of former senior US officials that advises on pandemic preparedness – named artificial intelligence the 16th priority in its Apollo Program for Biodefense. This marked the commission’s first addition to the American strategic initiative to develop technologies against all biological threats, since its founding in 2021. Speaking to HSToday, Commission Co-chair and former US Homeland Security Secretary Tom Ridge warned that “the federal government is simply not equipped to address these rapid advances in technology.” Yet the FY2026 budget cuts funding to the Biomedical Advanced Research and Development Authority (BARDA) – which funds the development of treatments for public health emergencies – by $361 million, the Strategic National Stockpile – the emergency reserve of medical supplies – by $230 million, and proposes eliminating the Administration for Strategic Preparedness and Response (ASPR), the agency that runs both. AI is compressing drug discovery’s slowest stage to make it cost-effective, while Washington is defunding the stage that comes next.
Key Points
- AI named 16th Apollo Program priority on 8 July, first new priority since 2021.
- Co-chair Tom Ridge: the government is “not equipped” for AI-era biological risk.
- Co-chair Donna Shalala: without investment, AI “could benefit the attacker before it benefits the defender.”
- Novartis (WEF): AI-driven simulation narrowed target identification for a rare kidney disease (ADPKD) to under a year; generative modeling cut a Huntington’s compound-design program from thousands of lab syntheses to roughly 60.
- BARDA cut $361M, to $654M (from $1.015B).
- Strategic National Stockpile cut $230M, to $750M (from $980M).
- FY2026 proposes eliminating ASPR; total ASPR funding falls $800M, to $2.83B.
- Biosurveillance/early-warning funding rises 17%, to $927M, in the same request.
- Office of the Director of National Intelligence (ODNI) dissolved the National Counterproliferation and Biosecurity Center (NCBC), its biothreat intelligence arm, in August 2025.
Analysis
COVID’s failure was the lag between identifying a threat and manufacturing scale-up, distribution, and stockpiling of the vaccine. This budget is worth reading in that context; it funds the side of the pipeline that spots a pathogen, not the side that turns a spotted pathogen into a countermeasure people can actually receive. Private AI-bio investment follows a different logic: Novartis and comparable efforts are pouring resources into drug discovery for diseases with a commercial market (chronic and rare disease, neurodegeneration), while biodefense-specific countermeasures have no market until a crisis hits. That is the standard public-goods rationale that already justifies state subsidies for semiconductors and critical minerals, but biodefense manufacturing has not gotten the same treatment. NCBC’s dissolution and ASPR’s proposed elimination make this worse, removing the institutions that would translate AI output into fielded capability.
Policy Implications
These are primarily relevant to Congress and Health & Human Services/BARDA leadership:
- Authorize BARDA to issue Advance Market Commitments. Binding pre-purchase agreements for countermeasure candidates that clear AI-validated development milestones give private AI-bio investment a positive demand signal.
- Convert Strategic National Stockpile funding into a capacity-linked mandate. Tie replenishment to a fixed share of validated domestic production capacity via multi-year (no-year) authorization, rather than a line item that resets, and can be cut, every budget cycle.
- Create a CHIPS Act-style investment tax credit for biomanufacturing. A federal credit for domestic production capacity such as fill-finish, cell and gene therapy lines, and rapid-response facilities pulls private capital into manufacturing the same way the CHIPS Act pulled it into fabs.
- Give the AI-to-market translation function statutory protection. Whatever replaces ASPR must retain explicit responsibility for moving AI-discovered candidates from lab to Stockpile; without a named institutional owner, that function disappears along with the agency.