Meta is taking a cue from SpaceX’s playbook by entering the cloud business arena, repurposing its unused AI computing resources for external clients while also offering access to its proprietary AI models. As one of Nvidia’s largest GPU customers, Meta has reallocated substantial funds into AI infrastructure development following a series of layoffs, with an anticipated investment of $145 billion this year. Despite the introduction of its advanced large-scale model, Muse Spark, internal demand has yet to fully utilize the extensive hardware investments. This has made the resale of surplus computing power a financially prudent decision, creating a fresh revenue stream through B2B monetization.

Earlier, SpaceX had already set a precedent by leasing GPU capacity originally acquired for xAI, securing computing power leasing agreements with Anthropic and Google valued at $1.25 billion and $920 million per month, respectively. This news has had a positive impact on Meta’s market performance, with its stock price climbing roughly 10% during intraday trading.

As tech giants encounter a temporary surplus in AI hardware investments, the resale of computing capacity has emerged as a prevalent strategy among industry leaders. This approach not only optimizes asset utilization but also eases financial pressures related to capital expenditures, potentially reshaping the competitive landscape of the traditional cloud computing market.