Tech Explained: Here’s a simplified explanation of the latest technology update around Tech Explained: On-device AI shockingly threatens data centre dominance in Simple Termsand what it means for users..
On-device AI is at the heart of a dramatic challenge to today’s vast data centre infrastructure, Perplexity CEO Aravind Srinivas has warned.
On-device AI as ‘biggest threat’
Speaking on YouTuber Prakhar Gupta’s podcast, Srinivas described on-device AI as “the biggest threat to a data centre” if intelligence can be “packed locally on a chip that’s running on the device,” removing the need to run inference on central servers. He argued that such a shift could radically change the economics of the billions of dollars currently poured into data-centre buildouts around the world.
On-device AI and a new AI revolution
The Perplexity chief sketched a move away from the prevailing model, where most AI tasks are processed in large, specialised facilities, towards powerful on-device AI that can run models locally. In his view, widespread adoption of on-device AI could usher in a more decentralised ecosystem, undercutting the dominance of cloud-based inference.
Human curiosity versus machine limits
Beyond infrastructure, Srinivas contrasted biological and artificial intelligence, noting that human brains are far more energy-efficient than sprawling data centres on a per-watt basis. He said humans possess intrinsic curiosity and a drive to pose original questions and revisit the familiar in new ways, qualities that current on-device AI and other systems do not naturally exhibit.
On-device AI and the future of work
Srinivas suggested that increasingly capable, personalised on-device AI assistants could narrow the gap between individuals and large institutions, much as smartphones did by putting powerful tools in many hands. He also stressed that age is no barrier to using AI effectively, but that a curious mindset is essential to benefit from on-device AI.
Srinivas’s comments paint on-device AI as both a technological and economic disruptor, with the potential to weaken centralised data centres while empowering users with tailored intelligence on their own hardware. Yet, he cautioned that despite these advances, human curiosity and motivation remain central to asking the questions that on-device AI and other systems are built to answer.