Meta Platforms reported its second-quarter earnings with one issue dominating investor attention: the soaring cost of its artificial-intelligence expansion.
Meta is not alone. Amazon raised its expected 2026 capital spending to approximately $220 billion, while Microsoft expects to invest around $190 billion in capital expenditures in 2026.
But Meta faces a more complicated question than its cloud-computing rivals.
Meta is still working to prove that its enormous AI investment can generate comparable new revenue streams.
Those concerns deepened after Meta raised the lower end of its 2026 capital expenditure forecast to $130 billion from $125 billion and said heavy infrastructure investment would continue beyond this year.
Morgan Stanley sees four major earnings opportunities
Morgan Stanley analyst Brian Nowak said Meta’s existing social media business continues to strengthen as the company prepares to introduce new revenue sources.
Meta now reaches 3.6 billion people across its apps each day, while Instagram has surpassed 2 billion daily active users.
Morgan Stanley views that audience as both a competitive moat and a platform from which Meta can launch new consumer and business products.
More Meta:
- Meta layoffs take disturbing turn in new lawsuit
- Meta business model in trouble from $1.4 trillion lawsuit
- Meta doubles down on smart glasses amid always-on feature concerns
The firm estimates that four opportunities could collectively add more than $9, or roughly 25%, to its base 2028 earnings forecast of $34.59 per share.
The largest potential contribution could come from Meta using its infrastructure as a “neocloud,” allowing outside companies to purchase computing capacity.
Morgan Stanley estimates that the opportunity could add about $2.97 to projected 2028 earnings per share.
That possibility is not entirely theoretical.
“We’re getting a lot of offers for compute at a significant premium over what we paid for it,” Zuckerberg said during the company’s earnings call.
He indicated, however, that Meta expects to generate higher margins from selling AI-powered products and services than from simply renting out raw computing capacity.
Morgan Stanley estimates that Meta AI search could add another $2.89 per share, subscriptions approximately $1.88, and API revenue roughly $1.22.
Some of those revenue streams are already developing.
Family of Apps revenue outside advertising reached $1 billion for the first time during the quarter, increasing 73% as paid messaging and subscriptions expanded.
Meta also plans to begin charging businesses for its Business Agent during the second half of 2026 through subscriptions and usage-based pricing.
More than 1 million businesses are already using the agents each week, according to the company.
“META seems to be pricing in all of the spend/capex/opex and isn’t pricing any of the call optionality of products to come,” Morgan Stanley analysts wrote, referring to the potential value of businesses that are not yet included in the firm’s base forecast.
Morgan Stanley reiterated its Overweight rating, kept the company as its top internet stock pick, and maintained a $775 price target.
The target represents about 32% upside from Meta’s July 29 closing price of $585.61.
Meta’s advertising business remains strong
Morgan Stanley’s bullish view is not solely based on future products.
Meta reported second-quarter revenue of $60.8 billion, up 28% from a year earlier, as its core advertising business continued to expand.
Advertising revenue rose 27% to $59.36 billion.
Ad impressions across Meta’s Family of Apps increased 14%, while the average price per advertisement climbed 12%.
Meta projected third-quarter revenue between $61 billion and $64 billion.
Morgan Stanley said the upper end would imply approximately 26% year-over-year growth, excluding foreign-exchange effects, and was stronger than expected.
The firm said Meta’s ability to recommend more relevant organic content and advertisements is improving as the company incorporates large language models into its systems.
That can increase engagement by helping Facebook and Instagram show users more relevant posts while improving advertising performance.
Instagram reached 2 billion daily active users, Threads surpassed 500 million monthly active users, and Facebook remained above 2 billion daily users.
Morgan Stanley said this reach and improving monetization give Meta the cash-generating foundation needed to fund its AI expansion.
Meta’s AI spending remains the main risk
Yet, this expansion carries a massive cost.
Meta spent $19.84 billion on capital expenditures in the first quarter and raised its full-year forecast to $125 billion to $145 billion.
Combined with the second quarter spending of $31.08 billion, Meta’s capital spending has exceeded $50 billion during the first half of 2026.
Now in Q2, the company has narrowed its full-year capital expenditure forecast to $130 billion to $145 billion.
As infrastructure spending consumed most of the cash produced by operations, Meta generated only $784 million in free cash flow during the quarter, down from $12.39 billion a year earlier.
Costs and expenses climbed 55% to $42.03 billion, including $2.4 billion in legal charges and $1.18 billion in severance expenses related to Meta’s May workforce reduction.
Operating income declined 8% to $18.78 billion, while the operating margin fell to 31% from 43%.
Net income dropped 14% to $15.85 billion, and diluted earnings declined 13% to $6.18 per share.
However, Morgan Stanley estimated adjusted earnings of $8.65 per share, nearly 7% above its $8.10 forecast.
The firm also raised its projected Meta earnings per share for 2027 and 2028 by approximately 2% and 4%, respectively, following the stronger advertising results.
Risks to Morgan Stanley’s bullish Meta view
Morgan Stanley’s outlook depends on Meta turning its AI spending into stronger engagement, better advertising performance, and new revenue streams, but the risks remain significant.
Weaker engagement, slower Reels monetization, softer advertising demand, or tighter regulation could all weigh on growth.
The biggest uncertainty is whether Meta can generate sufficient returns from its rapidly expanding infrastructure.
Morgan Stanley warned that missteps in the company’s data-center buildout could lead to higher long-term capital requirements and weaker returns on invested capital.
Further increases in operating expenses and capital expenditures could also pressure profits and free cash flow.
Meta says it remains constrained by available computing capacity and expects to find productive uses for infrastructure coming online in 2026 and 2027.
The company acknowledged, however, that it has less visibility into its computing needs beginning in 2028.
That leaves investors with the central question behind Morgan Stanley’s thesis: whether Meta’s new AI products can eventually generate enough revenue to justify the scale of its spending.
Related: Cathie Wood buys $14.3 million of tumbling semiconductor stock
