Meta stock has over 50% upside, but investors may be missing this catalyst

Meta Platforms could turn its aggressive investment in artificial intelligence infrastructure into a significant new source of revenue if it chooses to sell some of its excess computing capacity to external customers, according to Evercore ISI.

The social media company is targeting 14 gigawatts of total compute capacity by 2027 as it rapidly expands its data-center footprint and develops its own AI infrastructure.

Meta shares have faced pressure from the scale of those investments.

The stock is down about 4% over the past month despite briefly gaining momentum following the launch of a new AI model in early July.

The stock is down 12% this year, even as the S&P 500 remains up by 12%.

However, Evercore analyst Mark Mahaney believes the infrastructure spending could ultimately give Meta an opportunity to benefit from the current shortage of AI computing capacity.

Mahaney raised his price target for Meta to $860 from $820 in a Monday note.

The target implies more than 50% upside from the roughly $570 level where the stock closed Tuesday after gaining 2%.

Compute shortage creates opportunity

Meta is currently the only major hyperscaler without a public cloud business, but recent reports have suggested the company could establish a Meta Compute division to monetize some of its data-center capacity beyond its own operations.

"In a market where almost every scaled operator is sold out, Meta is a very rare potential merchant seller of surplus capacity," Mahaney wrote.

The analyst does not expect Meta to transform into a full-fledged neocloud provider like Nebius or CoreWeave.

Instead, he sees the potential business as a limited effort to monetize a portion of capacity that would otherwise remain unused.

Mahaney estimates that putting just 0.5 gigawatts to 1 gigawatt of capacity on the market could generate between $11 billion and $22 billion in annual external revenue.

Leasing 1 gigawatt would represent only about 7% of Meta's targeted 14 gigawatts of capacity in 2027, while potentially adding as much as $4.32 to earnings per share, according to his estimates.

Zuckerberg remains focused on AI returns

Meta has signaled that it is not looking to sacrifice its longer-term AI ambitions for short-term infrastructure revenue.

CEO Mark Zuckerberg said during the company's latest earnings call that Meta has received offers to purchase computing capacity at a premium.

However, he cautioned that "it would be foolish to basically just sell all of the compute and take a short-term profit," because "there will continue to be a significantly higher margin on selling intelligence rather than selling compute directly."

That suggests Meta's priority remains using its infrastructure to power its own AI products and services, where it expects to capture greater margins.

Mahaney described the potential compute business as a "call option" for investors rather than something that should be treated as an established revenue stream.

"We would use compute optionality as a reason to own the AI infrastructure play in Meta, but we would not underwrite a 'full-blown' hyperscaler compute annuity stream," Mahaney wrote.

Meta is also planning to manufacture its own Iris chips as part of its broader infrastructure strategy.

Investors weigh AI spending

Meta trades at about 17 times analysts' earnings estimates for the next 12 months, according to Mahaney, within 10% of its trailing three-year trough valuation multiple.

The analyst believes Meta can demonstrate returns from its substantial AI investments.

"Over the last two years, Meta has decisively proven that it can effectively deploy AI to materially improve its customer experience," Mahaney said.

For investors, the possibility of Meta Compute therefore provides another potential payoff from the company's infrastructure spending, even as Meta continues to emphasize that selling AI-powered intelligence remains more valuable than selling computing capacity itself.