Google DeepMind to train AI in war and diplomacy in EVE Online

Google DeepMind has entered into a research partnership with Fenris Creations (the developer of EVE Online) and will begin training AI models in one of the most complex MMORPGs. This week, the studio separated from its owner Pearl Abyss for $120 million and simultaneously received an investment from Google.

Why EVE Online Is an Ideal Laboratory for AI

EVE Online is not just a game about space. It is a fully functional simulation with its own economy, currency, complete player corporations (with hierarchy, diplomacy, and wars), a dynamic market, and political alliances. Thousands of people inhabit this world simultaneously, trading, building coalitions, dividing power, and planning operations.

For AI, this is an ideal sandbox because the environment is unpredictable — there is no single correct solution; constant adaptation is required. An agent interacts with other AI and players simultaneously; a mistake can be costly, and success requires understanding motives and negotiation. In typical simulations, AI learns either to win or to optimize a single metric.

In EVE Online, it must do everything at once: compete, cooperate, trade, fight, and survive.

What DeepMind's AI Agents Will Learn

The research program will cover several areas simultaneously:

• Cooperation — how to act in coordination with other agents to achieve a goal
• Conflict and warfare — strategic planning under opposition
• Economics — resource management, trading, finding the balance between supply and demand
• Diplomacy — negotiation, building alliances, reaching cooperation agreements
• Adaptation — quickly changing tactics when market conditions change

Experiments will take place on a local (offline) server — an isolated copy of EVE Online with its own economy and rules. This allows DeepMind to iterate quickly: launch the model, observe results, roll back failed versions, without risking impact on the live economy of the main game, where real people invest time and money.

How Google Strategically Positioned Itself

Google's investment in Fenris Creations is not simply sponsorship of academic research. The company obtained a minority stake directly in the developer, which guarantees it long-term access to the environment and influence over its development. Fenris Creations, meanwhile, has fully retained its independence: management, teams, and development plans for EVE Online, EVE Vanguard, and the new game EVE Frontier remain in the studio's hands. Google does not control the studio, but has established itself as a strategic partner in AI research — a typical Big Tech move in which a company invests in the future while simultaneously securing rights to research.

What This Means for AI and Games

Games are increasingly becoming laboratories for AI research. Complex economic systems, social interactions, multi-agent scenarios, unpredictability — all of this is necessary to prepare AI agents for real-world tasks. If an AI can conduct warfare, trade, and negotiate in EVE Online, it can handle more abstract human scenarios as well. EVE Online is a fully hyper-complex system that reflects reality better than any artificial task. For players themselves, this has an interesting consequence: in the future, they will encounter AI agents directly in the game, competing with them on equal terms.