Google merges DeepMind and Google Brain into Google DeepMind
Two research groups that had competed internally for years — DeepMind and Google Brain — were folded into one unit reporting to Demis Hassabis.
- Labs & people
- Major
Alphabet announced that it was merging DeepMind, the London-based research lab it acquired in 2014, with the Brain team from Google Research into a single organisation, Google DeepMind, led by DeepMind co-founder and chief executive Demis Hassabis. Zoubin Ghahramani joined the research leadership and Eli Collins became vice president of product for the combined unit, which brought together two groups that had for years operated with separate research agendas, publication norms and, at times, overlapping projects.
Hassabis said the goal was to “accelerate our progress towards a world in which AI helps solve the biggest challenges facing humanity,” arguing that a single organisation could move “with greater speed, stronger collaboration and execution” than two labs competing for internal resources and attention. The announcement combined DeepMind’s research culture, built around projects such as AlphaGo and AlphaFold, with Google Brain’s closer integration into Google’s product and infrastructure teams and its role in developing large language models including PaLM and LaMDA.
The reorganisation followed roughly two months of visible pressure on Google’s AI position: the rushed, stumbling launch of Bard in February had drawn criticism that Google, despite originating much of the research underlying the current wave of large language models, was struggling to ship products as quickly as OpenAI and Microsoft. Consolidating DeepMind and Brain under one reporting line was widely read as a direct response — an attempt to remove internal duplication and speed the path from Google’s research advantage to shipped products, a strategy the merged organisation followed through on later that year with its first jointly developed model family.