Google DeepMind announced that it will use the popular 23-year-old MMORPG game Eve Online to train its models. Here are the details.
Google DeepMind took action to train its new models on the space-themed online role-playing game Eve Online, which has been running for more than twenty years. This huge universe, prepared by Iceland-based developer CCP Games, attracts attention with its complex player interactions and economic structure.
DeepMind has invested millions of dollars in a minority stake in the game’s developer, the studio recently renamed Fenris Creations. This investment took place simultaneously with the management of the game being taken back from Pearl Abyss.
Strategic Cooperation and Education Process
Eve Online offers a very complex structure where players can establish their own companies and engage in a wide range of activities, from mining to space wars. Google DeepMind CEO Demis Hassabis stated that games have always been at the center of DeepMind’s technological breakthroughs.
Hassabis stated that games are an ideal environment to develop algorithms, as in previous projects such as Atari DQN, AlphaGo and AlphaStar. Research will be conducted on isolated servers of the game to avoid affecting live servers.
This study aims to create more advanced models by analyzing players’ behavior. The data obtained is also planned to be used to improve the gaming experience in the future.
A Complex Player-Focused Universe
Throughout its 23-year history, Eve Online has been known for player-driven stories and hours of large-scale battles. These wars can lead to the loss or gain of virtual assets worth hundreds of thousands of dollars in the real world.

The main reason why Google chose this game is that players make decisions that resemble real-world risks. This provides a unique data source for models to understand complex decision-making processes.
Fenris Creations moved to an independent structure and formed its own board of directors, thus gaining a structure that can make faster decisions. The company predicts that this collaboration with DeepMind will contribute to both technological research and the future of the game.
The research process will be expanded over time and the findings will be used to improve the mechanics of the game. How do you think such complex data in the gaming world will change the future of the technology world?