Perplexity CEO Aravind Srinivas has warned that the rush to build massive, multi-billion-dollar data centres could prove misplaced if artificial intelligence shifts decisively to on-device processing.
In a recent podcast interview with YouTube creator Prakhar Gupta, Srinivas argued that packing sophisticated AI models directly onto user devices represents an existential risk to centralised infrastructure.
“Well, the biggest threat to a data centre is if the intelligence can be packed locally on a chip that's running on the device and then there's no need to inference all of it on like one centralised data centre,” he said.
Such a shift would decentralise AI computation, eliminating latency from remote servers while enhancing privacy, as personal data would remain on the device rather than traversing networks.
Personalised 'digital brain' could be the future
Srinivas described the potential for devices to host adaptive models that learn from individual user behaviour, creating what he termed a personal 'digital brain'.
"That way you don't have to repeat it. That's your intelligence. You own it. It's your brain," he noted, adding that this could render vast investments in centralised facilities uneconomical.
He questioned the rationale behind spending hundreds of billions - or even trillions - on global data centre networks if local silicon proves capable of handling complex tasks reliably.
Breakthrough yet to arrive
The Perplexity chief acknowledged that no company has yet delivered a model efficient and intelligent enough to run sophisticated large language models entirely on consumer hardware for demanding applications.
“That's not yet happened. No one's actually shipped a model that can be packaged on your computer, a very efficient chip locally and then it's very intelligent enough that it can complete tasks reliably,” Srinivas said.
When such a breakthrough occurs, however, he predicted it would fundamentally reshape the industry.
Srinivas's comments come amid intense investment in hyperscale data centres by technology giants, some of whom are exploring unconventional solutions including orbital facilities to address energy constraints.
