Meta has confirmed to TechCrunch that Santosh Janardhan, the company's longtime head of #infrastructure and co-head of engineering, is stepping down. A key figure in Meta's evolution into a social media and tech powerhouse, Janardhan will be replaced by Dave Brown, who joined the company earlier this year as vice president of infrastructure. According to his LinkedIn profile, Brown spent 19 years at Amazon before joining Meta, focusing on #compute, cloud, and AI.
"Santosh will stay on for transition and he is moving into an advisory role on Meta Compute," a Meta representative said. He is expected to leave next year. Janardhan originally joined the company in 2009 and helped it grow into the massive internet presence it is today. Bloomberg first reported his departure, citing a recently circulated internal memo. The memo did not specify reasons for his exit, and Meta did not immediately respond to a request for clarity. However, Janardhan's memo apparently cited the need to "pass the baton."
In recent years, Meta's investment in infrastructure has exploded as it seeks to keep pace with competitors in the AI race. Earlier this year, Janardhan was tasked with leading Meta Compute, the company initiative to scale its data center investments and strategy. The leadership change comes at a time of unprecedented AI spending, and the continuity of Meta's infrastructure strategy is under scrutiny. Brown's #Amazon background, particularly in compute and AI, could bring fresh perspectives, but he will need time to adapt to Meta's unique hyperscale infrastructure demands.
[AgentUpdate Depth Analysis] The change at the top of Meta's infrastructure team is more than an executive shuffle; it reflects the intense compute demands of the AI Agent ecosystem. Janardhan's Meta Compute initiative was designed to support the training and inference of large-scale AI agents. Compared with rivals like Google and Microsoft, Meta has been slightly behind in custom silicon and hyperscale data center buildouts, and this leadership change may signal an acceleration. For AI Agent developers, infrastructure stability and cost directly impact deployment efficiency and scalability. Dave Brown's cloud background could bring mature cloud-native and AI compute scheduling expertise, but whether it translates into a competitive advantage for Meta's agent ecosystem remains to be seen. In the long run, infrastructure competition will determine who can host the next billion-agent real-time interactions. Is Meta being proactive or reactive? Only time will tell.
