Meta announced a new cloud business to sell its excess AI computing power on July 2, 2026 [1].
This shift challenges the prevailing industry narrative that owning the AI model itself is the primary source of value. By moving toward a commoditized compute-as-a-service model, Meta is repositioning the infrastructure underlying artificial intelligence as a critical revenue stream.
Dan Niles, an investor, said this development undermines the market's belief in model ownership as the sole driver of AI profitability [1]. The strategy treats computing power as a utility, allowing the company to monetize the hardware and energy required to run large-scale systems.
Parallel to this business pivot, new data from Ramp and Revelio have highlighted the ongoing impact of AI on the labor market [1]. The findings reinforce concerns that AI is actively displacing jobs across various sectors.
These combined developments suggest a transition in the AI economy. While the first phase of the boom focused on the creation of proprietary models, the current phase emphasizes the physical infrastructure, and the resulting shifts in employment [1].
“Meta announced a new cloud business to sell its excess AI computing power”
Meta's pivot indicates a strategic shift from software-centric value to infrastructure-centric value. By commoditizing compute, the company is betting that the 'picks and shovels' of the AI era—the hardware and power—are more sustainable profit centers than the models themselves. Simultaneously, the employment data suggests that the economic gains from this infrastructure are coinciding with a tangible contraction in the human workforce.
