The Essence of the Announcement
Google DeepMind and Google Cloud announced on October 14, 2026, a new program dedicated to scientific research in Europe. Under the program, up to 200 European researchers will gain direct access to the frontier artificial intelligence models of both companies. Participants will additionally receive a total of $3 million in compute credits. The companies made the announcement in an official statement — according to Reuters, the initiative is designed to help scientists solve complex research tasks faster.
The program is being run jointly by two organizations: Google DeepMind, the division specializing in artificial intelligence research, and Google Cloud, the cloud computing platform within Alphabet. This partnership defines the program's two pillars — on one side, frontier models; on the other, the computing power needed to run them. Further details were published on Google's official blog.
The announcement was distributed simultaneously through two sources: the Reuters news agency, citing the companies' statement, and Google, which published a detailed description of the program on its official blog. Both sources report the same scale of the program — up to 200 participants and a total of $3 million in compute credits.
The announcement text emphasizes that access is granted in direct form: researchers will use the models firsthand, without intermediaries. The compute credits are intended to cover the cost of the capacity required to use the models. In this way, the program gives each researcher two things at once: the frontier models themselves and the computing power needed to run them.
Which Tools Will Be Unlocked
The list of tools to be opened is given in the official announcement. According to it, participants will gain access to Google's leading scientific artificial intelligence tools: AlphaFold 3, WeatherNext, AlphaGenome, AlphaEvolve, and AlphaEarth Foundations. According to the company, these five tools are among its most advanced developments created for scientific research.
Each tool on the list is designed for a distinct field of scientific work — all of them were built to help researchers solve complex scientific problems. Access is granted not in the form of limited testing, but as a right of direct use. Researchers will therefore be able to apply the models' capabilities directly in their day-to-day scientific work.
All five tools are presented as Google's leading developments for scientific research. They are designed to help scientists solve complex scientific problems — the official announcement specifically highlights this goal. Access rights are granted to all five tools at once — scientists can choose and use the ones that fit their research tasks.
Link to the RAISE Initiative
This commitment by the companies is aimed at supporting the European Union's work on artificial intelligence for science. The announcement states that some frontier models will be made available through the RAISE initiative. The initiative is intended to address a problem European scientists sometimes face — the shortage of computing capacity needed to access frontier models.
The program thus meets two needs at once: scientists gain access to frontier models and receive the computing resources needed to use them. Both directions are directly relevant to the everyday research work of the European scientific community. The initiative responds to the most pressing needs of this very community.
The problem is that, in some cases, European scientists cannot find the computing capacity required to access frontier models. The RAISE initiative is designed to fill precisely this gap — to shorten the distance between scientists and the resources they need. The initiative is being implemented as part of the European Union's broader efforts in artificial intelligence for science.
The Program's Goal
According to the companies, the program will let researchers solve complex problems faster. Two areas are cited as examples: forecasting extreme weather conditions and mapping molecular structures. The initiative is also aimed at strengthening the continent's scientific competitiveness.
The official statement illustrates the program's practical significance with two concrete examples. The first is forecasting extreme weather conditions; the second is mapping molecular structures. Both are among the complex scientific tasks that require large amounts of computing.
The program will help researchers tackle complex challenges faster — from predicting extreme weather patterns to mapping molecular structures — and reinforce the continent's scientific competitiveness. — Reuters
The announcement was made on October 14, 2026. The upper limit on participating researchers is 200, and the total amount of compute credits allocated is $3 million.
