Meta q2 earnings put AI spending and ad growth in focus

Meta q2 earnings are due after the close of regular trading Wednesday, with Wall Street expecting another quarter of fast revenue growth and a sharper test of investor tolerance for AI spending. Analysts polled by LSEG expect adjusted earnings of $7.22 a share on revenue of $60.17 billion.

That revenue estimate would represent 26% growth from $47.52 billion a year earlier, according to LSEG data cited by CNBC. Meta’s core advertising business remains the main profit engine, while the company is trying to prove that its artificial intelligence investments can expand beyond ad targeting and engagement tools into models, services and paid products.

What are analysts expecting from Meta q2 earnings?

StreetAccount expects Meta to report 3.61 billion daily active people across its family of apps in the second quarter. Analysts also expect average revenue per person of $16.65, a closely watched measure of how effectively Meta turns user activity across Facebook, Instagram, WhatsApp and its other apps into sales.

Investors are also expected to focus on capital expenditure, after Meta lifted its full-year capex outlook in April to as much as $145 billion from a previous high of $135 billion. Analysts expect second-quarter capex of $33.9 billion and $136.7 billion for the full year, according to CNBC.

Capital expenditure is money a company spends on long-lived assets such as data centers, chips and network equipment. For AI companies and large internet platforms, higher capex can support more computing capacity, but it can also weigh on free cash flow if revenue from those assets develops more slowly than spending.

Meta’s AI infrastructure buildout has accelerated alongside similar spending by Alphabet, Amazon and Microsoft. CNBC reported that Alphabet shares fell last week after the company raised its 2026 capex guidance to as much as $205 billion, up from a prior forecast of up to $190 billion.

Why is Meta’s AI spending under scrutiny?

Meta does not have a cloud infrastructure business comparable to the other major hyperscalers, which means investors are watching how the company plans to earn returns from large-scale AI capacity. CNBC reported that Chief Executive Mark Zuckerberg has recently indicated Meta may sell some AI capacity to outside customers, and CNBC confirmed earlier this month that Anthropic is in preliminary talks to lease AI-related compute from Meta.

Meta has also announced several large data center commitments in July. The company said Tuesday it was working with BlackRock on a $14 billion data center project in El Paso, Texas, according to CNBC. That followed disclosures of a Hyperion data center project in rural Louisiana costing more than $50 billion and plans for a $9 billion data center in Alberta, Canada.

The company’s AI product push has also broadened. Earlier this month, Meta introduced Muse Spark 1.1, which AI chief Alexandr Wang described as the company’s “strongest model for agentic and coding work yet,” according to CNBC. Meta has also released Muse Image, including features available to power users and creators through monthly subscription plans disclosed in May.

Wedbush analysts wrote last week that “the gap between capex intensity and diversified monetization remains the central debate for the stock.” The firm has the equivalent of a hold rating on Meta shares and said that uncertainty explains its cautious stance despite Meta’s valuation discount to peers.

Meta shares are down 10% this year, according to CNBC, trailing the Nasdaq as investors weigh the durability of advertising growth against the scale and timing of AI returns. The company’s Reality Labs unit, which develops virtual reality, augmented reality and AI-powered wearable devices such as Ray-Ban Meta glasses, is expected by StreetAccount to post a quarterly loss of $5.07 billion on revenue of $423.4 million.

This story draws on original reporting from CNBC.