A flu forecasting model built with Google AI ranked first at predicting flu-related hospital admissions for the 2025-26 season, according to a CDC end-of-season analysis announced this week. Of 39 eligible models, Google’s best matched the season’s observed hospital admissions.
The result comes from FluSight, the CDC’s collaborative forecasting program that pools weekly submissions from government, industry, and academic teams. A top ranking does not mean perfect week-to-week accuracy, but it does show how Google’s approach compared with other participants in that season.
What FluSight is, and why it matters
From October through May, FluSight collects weekly forecasts of U.S. flu hospital admissions for the current week and three weeks ahead. The CDC uses the combined forecast each week to communicate anticipated state-level demand for medical services.
In plain terms: a stronger hospital-admission forecast can help planners prepare staffing, beds, and supplies before a wave peaks. That is why a ranking like this matters beyond the research community.
How Google built the forecasts
Per Google Research’s official post, the forecasts were developed with Empirical Research Assistance (ERA), an AI tool that generates optimization algorithms across scientific fields. Research on ERA was published in Nature, and ERA’s underlying technology is now available to trusted testers as part of Google’s experimental science tools.
What to take away, and what not to
The ranking covers the 2025-26 season and models that met FluSight’s eligibility rules. It does not guarantee the same result in every future season or every region, and it does not replace clinical judgment. It is a research validation of hospital-admission forecasting inside a defined competitive evaluation.
More Google research on this site
For related Google coverage here, see Gemini 4 Argon, our piece on DeepMind’s SynthID Bio, Project Suncatcher in orbit, and updates in AI News.
Google says the result validates its confidence that AI combined with human ingenuity will improve the ability to forecast diseases worldwide. The post was published by Zahra Shamsi of Google Research on September 30, 2026.
What it means for you
- In the US: FluSight forecasts feed the CDC’s weekly communication about expected state-level demand for medical services, so better forecasts matter most to US hospitals and health planners.
- In the EU and UK: the announcement does not mention European or UK health agencies or any use of the model outside the US.
- No product to try: ERA’s underlying technology is available only to trusted testers through Google’s experimental science tools. There is no price or open sign-up.
- Google tools you can use today: for Gemini plans and prices in USD, GBP and EUR, see our Gemini guide.
Sources
- Google Research (Zahra Shamsi), Google’s AI ranks #1 for predicting flu hospitalizations: blog.google/innovation-and-ai/models-and-research/google-research/google-science-ai-flu-forecasts/
