Meta has selected Robert Fergus, former research director at Google DeepMind, to head its long-standing Fundamental AI Research (FAIR) lab. The appointment, first reported by Bloomberg, comes as Meta restructures its AI research operations following internal leadership changes and talent shifts.
Fergus brings deep experience in artificial intelligence, having spent the last five years at DeepMind, where he led various research initiatives. Before his time at Google, he worked at Meta as a research scientist, making this a return to familiar ground.
His leadership is expected to help stabilize and refocus FAIR, which has played a foundational role in Meta’s AI development but has seen significant changes in recent years.
FAIR, established in 2013, was behind Meta’s early AI efforts, including the creation of the first and second versions of its large language model series, Llama 1 and Llama 2. However, the lab has recently faced internal challenges, including a wave of researcher departures and growing overlap with Meta’s newer GenAI division. GenAI is the team responsible for Llama 4 and is now seen as Meta’s more active AI product development group.
The shift in leadership follows the recent exit of Joelle Pineau, Meta’s former vice president of AI research, who announced in April that she would be leaving the company for a new opportunity. Pineau played a key role in driving Meta’s AI strategy and helped elevate its research visibility.
With Fergus stepping in to lead FAIR, Meta appears to be signaling a renewed commitment to its foundational AI research arm, even as it ramps up more product-focused development in GenAI. The move suggests a desire to maintain a balance between cutting-edge innovation and long-term scientific exploration in AI.
Meta has not yet announced additional structural changes at FAIR, but Fergus is expected to bring fresh energy and vision to the team. His dual experience at both DeepMind and Meta positions him uniquely to align research priorities with the company’s growing ambitions in generative AI and machine learning.
