Perplexity AI CEO Aravind Srinivas said that spending on artificial intelligence will become more efficient very soon [1].
This shift in spending efficiency is critical as corporations and investors weigh the massive capital expenditures required to maintain AI infrastructure against the actual productivity gains realized by businesses.
Srinivas said the trajectory of the industry during an appearance on CNBC’s ‘Power Lunch’ program [1]. During the segment, he addressed whether AI spending is slowing and how companies can improve their operational efficiency as they integrate these tools into their workflows [1].
Beyond the financial aspects of the technology, Srinivas said the societal shifts accompanying the AI transition. He said that AI could eliminate up to 100 million jobs in the next decade [4].
While the prospect of large-scale job loss is a central point of debate for regulators and labor advocates, Srinivas suggested a change in how individuals spend their time. He said, "Spend less time doomscrolling on Instagram; spend more time using the AIs" [3].
The conversation at CNBC focused on the next phase of AI development, specifically how the industry will move from the current era of high-cost experimentation toward a more sustainable model of efficiency [1].
“AI spending will become more efficient very soon.”
The assertion that AI spending will soon become more efficient suggests a transition from the 'build-out' phase of generative AI—characterized by massive GPU procurement and energy consumption—to an optimization phase. If the cost of deploying these models drops while utility increases, it may accelerate the adoption of AI in sectors that previously found the technology too expensive to implement, even as it heightens the risk of workforce displacement.
