Meta Delays Muse Spark API as 2026 Capex Rises
Meta postponed the Muse Spark developer API with no firm launch date while raising 2026 capital spending guidance to $125–$145 billion.
Meta Platforms has delayed the public release of its Muse Spark developer API and has not given a firm launch date. The Muse Spark model was unveiled in April as the first major release from the company’s Superintelligence Labs.
Meta AI chief Alexandr Wang posted on X in April that the API was “coming soon.” Nearly two months later the tool is not broadly available. A Meta spokesperson confirmed the company is testing the API with early partners and expects to release it “within the month,” without providing a fixed date.
The delay coincides with a larger capital plan. In April Meta raised its full-year 2026 capital expenditure guidance to $125-145 billion, up from a prior range of $115-135 billion. Company executives attribute the higher range to investments in data centers, chips and additional compute capacity to support AI work.
Investors and analysts have tracked earlier schedule slips for internal models this year. Muse Spark had been presented as an opportunity to show faster delivery from the AI group; the API delay creates a near-term point of focus for product timelines.
In the first quarter Meta reported advertising revenue of $55 billion. Ad impressions rose 19% and the average price per ad increased 12% year over year. Company executives say AI tools are already being used to improve ranking, targeting and ad delivery across Facebook, Instagram and WhatsApp.
On the company earnings call, when asked how Meta will measure returns on the increased capital outlay, Mark Zuckerberg called the question “a very technical question.” The stock fell in after-hours trading after the results were released.
Analysts remain largely positive: 64 analysts rate Meta a “Strong Buy” with an average 12-month price target of $826.75, compared with a recent share price near $623. Until the Muse Spark API is widely available, investors will have fewer direct metrics to measure developer adoption and will rely on operating results to evaluate the company’s compute and data-center investments.








