Meta faces $125B capex test as BofA flags AI returns path

Meta Platforms Inc. is pouring as much as $145 billion into artificial intelligence infrastructure in 2026, and Bank of America argues the payoff is starting to come into view after months of investor skepticism.

"AI is already delivering benefits within Meta's core advertising business by improving content recommendations and ad targeting," Justin Post, an analyst at Bank of America, said. Post reiterated a Buy rating and $835 price target on the stock.

Meta raised its 2026 capital expenditure guidance in April to between $125 billion and $145 billion, nearly double last year's spending, while first-quarter revenue rose 33% year-on-year to $56.3 billion. The stock still fell more than 6% in after-hours trading when the spending hike was disclosed. Chief financial officer Susan Li told analysts the company had "continued to underestimate our compute needs even as we have been ramping capacity significantly."

The spending cycle has put Meta at the center of a broader debate about whether hyperscaler AI investment can generate returns commensurate with the scale of outlay. Morgan Stanley estimates AI-related global debt issuance will top $570 billion in 2026, with Alphabet, Amazon, Microsoft and Meta expected to spend $700 billion in combined outlays this year.

Meta's path to monetization rests on two factors, according to BofA: frontier-level model development over the next nine months and scaled adoption of new AI products. The firm pointed to potential revenue streams from AI subscriptions, enterprise offerings and business agents as catalysts that could shift investor sentiment.

The company has already moved about 7,000 employees into AI-focused roles and cut roughly 10% of its workforce in May — the fourth wave of layoffs this year — to redirect resources toward compute infrastructure. Meta also deepened its partnership with Mukesh Ambani's Reliance Industries, announcing a 168-megawatt AI data centre in Jamnagar, India, that will support its core products and growing AI compute requirements.

The $1 Trillion Question for Hyperscalers

Meta's spending trajectory mirrors a broader industry trend. Morgan Stanley forecasts hyperscaler capital expenditure will surpass $1 trillion in 2027, driving a shift toward debt financing as companies that long relied on cash flows turn to bond markets. AI-related global debt issuance stood at nearly $236 billion as of May 31, fourfold higher than the same period last year.

For Meta specifically, the question is whether its AI investments can translate into measurable revenue growth before depreciation costs pressure margins. BofA's Post said that a combination of stronger AI capabilities, successful product adoption and continued advertising outperformance could improve investor confidence and support a higher valuation. The next two earnings reports, he said, should provide greater clarity on AI monetization.

Meta shares, which have underperformed since the capex guidance hike, trade at a discount to the narrative that the company's AI buildout will eventually pay off. The stock's re-rating depends on whether Meta can demonstrate that its $125 billion-plus bet is producing returns — not just compute capacity.

This article is for informational purposes only and does not constitute investment advice.