Dear Readers,

It’s striking how quickly the ground under our feet is shifting. One moment, AI feels like a stack of narrow tools; the next, a single lab is weaving games, biology, chips, weather, and virtual worlds into something that looks uncannily like a unified path toward general intelligence. Today, we follow that thread back to its source: Google DeepMind. What began as a small London startup with the ambitious goal of “solving intelligence” has quietly become the engine shaping Google’s—and possibly this century’s—technological direction. If you’ve ever wondered how a Go-playing agent, a protein-folding system, a weather model, and a 3D world generator fit into the same story, this issue will give you the map.

In the pages ahead, you’ll step into DeepMind’s evolution from startup to AGI laboratory: the Atari experiments that lit the spark, the Nobel-level breakthroughs of AlphaFold, the chips and algorithms designed by AI itself, and the new agents learning inside open-world games. You’ll see how these seemingly distant breakthroughs connect, and why they matter for the next decade across science, industry, and daily life. If you’re ready to understand how one lab became the quiet architect of the AI era, let’s dive in.

All the best,

If you had to choose a handful of moments that changed the story of AI, Google DeepMind would appear again and again. A Go program quietly defeats world champion Lee Sedol on live television. An algorithm “solves” a 50-year-old biology puzzle and later wins a Nobel Prize. Weather models outperform the best physics-based simulations. And just days ago, an agent was released that can step into open-world games, understand spoken goals, and learn how to act in real time.

All of these breakthroughs lead back to the same place: a London research lab founded in 2010 with an almost absurdly ambitious mission, to “solve intelligence” and use it to tackle humanity’s hardest problems. Google DeepMind, as the lab is known today, has become both a symbol and a driving force of the modern AI revolution. It operates in a unique space, balancing games and science, near-term profit and long-term research, and Google’s advertising-driven business and the speculative future of artificial general intelligence.

This article traces that story on two levels. First, the institutional: how DeepMind emerged from the London startup scene, why Google acquired it, and how it evolved into the core of Google’s AI strategy. Second, the technical: a chronological tour of the lab’s most significant systems — from AlphaGo to AlphaFold, from weather models to AI-designed chips, from world-building engines like Genie 3 to the newly released SIMA 2 agent.

Beneath these historical details lies a more fundamental question: how do these seemingly disparate projects — board games, proteins, weather, 3D worlds, and competition math — fit together into a coherent path toward general intelligence? And what does that path reveal about the future Google is attempting to build, with DeepMind at its center?