UBS: Meta's Sale of AI Computing Power May Not Be Bad News, Could Actually Alleviate Profit Pressure

BlockBeats News, July 2nd - UBS believes that if Meta were to sell AI computing power or provide model access services to external clients, it should not necessarily be interpreted by the market as a negative signal of "AI infrastructure oversupply." Instead, this could potentially become a path for Meta to more quickly monetize its massive AI investment.

In its First Read report released on July 1st, UBS mentioned that Meta is considering two commercialization approaches: selling "raw" computing power to external companies and providing access to AI models hosted on Meta's infrastructure. The report stated that Mark Zuckerberg had previously mentioned similar options in public, so this is not entirely new information.

However, this direction may still make some investors uneasy. Meta's previously outlined long-term growth opportunities to the market mainly included advertising, more immersive content experiences, business messaging, Meta AI, AI devices, among others, rather than directly transforming into a cloud computing or computing power provider. Therefore, if the company indeed starts selling computing power externally, the market may question whether this is an active monetization move or a passive digestion after excessive AI capital spending.

UBS takes a more pragmatic view. The bank believes that one of the core issues Meta currently faces is the excessively long AI investment cycle with unclear revenue realization timing. Selling cloud computing power or model access could potentially bring in revenue more promptly compared to waiting for Meta AI chatbots or enterprise intelligent agents to ramp up, thereby alleviating investor concerns about flat or declining EPS in 2027.

The report maintains Meta's Buy rating with a 12-month target price of $865, while the stock price listed in the report is $601.85. UBS expects Meta's diluted GAAP EPS to be approximately $32.6 and $33.0 in 2026 and 2027, respectively, and stated that profit forecasts will not be adjusted until the company confirms relevant news. The target price is still based on 26 times the full-year expected diluted GAAP EPS as of the first quarter of 2028.

The implication of this report is that Meta's AI strategy is entering a new phase: the market is no longer just asking how much it will spend on buying GPUs and building data centers, but is starting to demand to see how these investments translate into returns. For UBS, selling computing power is not a strategic retreat but an additional cash flow outlet for the AI investment cycle.