Meta’s Q2 Results Reveal Mixed Performance
Meta released its second-quarter results Wednesday after the market closed. While the earnings per share (EPS) were below expectations, the sales figures surpassed forecasts. The company has also scaled back its capital spending plans for 2026, yet its revenue outlook for the third quarter is somewhat disappointing, coming in below the midpoint.
Immediately following the earnings report, Meta’s stock plummeted nearly 8%.
During the quarter, Meta reported an EPS of $6.18, alongside revenues totaling $60.8 billion. Analysts had anticipated an EPS of $7.14 and revenues around $60.2 billion, according to Bloomberg consensus estimates.
This time last year, the company’s EPS was $7.14, with a revenue of $47.51 billion.
Looking ahead to the third quarter, Meta forecasts sales between $61 billion and $64 billion, while Wall Street had hoped for a midpoint closer to $63.1 billion.
Meta’s second-quarter advertising revenue hit $59.3 billion, slightly higher than the expected $59.07 billion.
Interestingly, the company raised its capital expenditure estimates, now projecting between $125 billion and $145 billion for the upcoming years, with a significant portion allocated for data center construction.
In a recent announcement, Meta and BlackRock revealed plans for a $14 billion, 1 gigawatt data center in Texas, with BlackRock owning 80% and Meta controlling the remaining 20%.
Earlier this month, Meta CEO Mark Zuckerberg mentioned that leasing out data center capacity to customers might be a direction worth exploring.
It remains unclear how exactly Meta plans to approach this, but they may look at SpaceX as a model. SpaceX recently signed a multibillion-dollar agreement to provide AI capabilities to various partners.
Meta could also follow a path similar to neo-cloud companies like CoreWeave, potentially renting out its hardware.
This month also marked the introduction of new AI models, including the Musk Spark 1.1, which came with a pricing strategy that notably reduces the costs compared to offerings from competitors like OpenAI.
Interestingly, Meta is charging developers $1.25 per million input tokens and $4.25 per million output tokens, which is considerably less than Anthropic’s prices.
With this aggressive pricing, it’s likely Meta could capture some market share from larger AI labs, especially as cost-conscious clients seek better deals.
