MENLO PARK, California, May 26, 2026, 05:02 PDT
- Meta now expects to spend $125 billion to $145 billion on AI infrastructure in 2026, raising its range from $115 billion to $135 billion.
- Wolfe Research estimates more than $26 billion in potential new 2027 revenue, while they also list expected 2027 spending at about $200 billion.
- Meta was last at $610.26, gaining $3.05 on its previous close.
Meta Platforms’ AI business could add over $26 billion in revenue by 2027, according to a Wolfe Research note. The report is adding new focus on when investors will see returns on heavy spending, not just from ad sales or product launches.
Meta hiked its 2026 capital spending plan to $125 billion-$145 billion, up from a previous range of $115 billion-$135 billion. That covers big-ticket items like data centers, servers and network equipment. The company blamed the jump on higher component costs and more data-center spending.
Meta’s legacy ad business is still a big profit driver, but its AI costs are getting harder to tie directly to new revenue. The company’s plans on AI spending pushed the shares down, with Crypto Briefing reporting the stock dropped as much as 12% after hours on the update and finished about 8.6% lower the following session.
Chief Executive Mark Zuckerberg responded to the worries. On Meta’s earnings call, he said both Meta’s projects and what he’s seeing in the industry make him “confident in this investment.” Zuckerberg said Meta tends to build out products at scale before it starts to make money from them. Q4 Capital
Meta’s CFO Susan Li told investors the company has kept underestimating compute needs, the power needed to train and use AI models. She said future spending might slow if Meta finds it has more capacity than it needs. Li offered less of a hard stop for investors.
Wolfe is taking a more upbeat view, pointing to potential new revenue streams from the same spending. The firm sees over $6 billion in extra non-ad revenue by 2027 from business AI, subscriptions, and commerce tools. It also forecasts more than $20 billion from AI-driven ad systems, Threads, ads on WhatsApp, and click-to-message ads.
Microsoft, Alphabet and Amazon channel AI demand into their cloud arms, which already handle computing and software sales for businesses. Meta’s focus stays on advertising, messaging, and consumer products, so any AI returns could take longer to show.
Analysts keep coming back to returns. On the call, Brian Nowak at Morgan Stanley pressed Zuckerberg about which markers he’ll track over the next 12 to 24 months to assess return on invested capital—how much profit Meta gets for every dollar it spends.
Meta’s business isn’t broken, but there are worries. Seeking Alpha’s Given Mahlangu wrote that spending on AI, few external monetization options, and softer free-cash-flow margins could limit gains in the medium term.
Meta is now in a tough spot next to its rivals. Reuters quoted Gil Luria, managing director at D.A. Davidson, saying Meta’s results “failed to impress investors” against “stronger numbers from Google parent Alphabet,” despite Meta beating analyst revenue estimates for the first quarter. Reuters
There’s a risk the spending cycle drags on, costs go up, and products need more time to show they can pay off. Meta cautioned that legal and regulatory issues in the U.S. and EU—like youth-safety probes—could hit results hard and possibly bring a material loss.
Right now, investors are being asked to back the usual Zuckerberg approach—spend big first, worry about making money later. What’s new is just how much cash is going out the door.
