Earlier this year, Meta cut thousands of employees and told investors the reductions were necessary to fund its AI ambitions. The company redirected that payroll into infrastructure, researchers, and models designed to compete with OpenAI, Anthropic, and Google. Mark Zuckerberg called it the most important investment Meta would ever make.
The first real test of whether that trade-off is working just arrived , and the answer is not straightforward.
What the WSJ reported about Muse Spark and what Meta has said
Meta (META) has delayed plans to release the developer API for its Muse Spark AI model multiple times, and as of June 3 had no planned launch date, according to Reuters, which cited people familiar with the matter. The holdup stems from software bugs and infrastructure problems discovered during testing.
Meta Chief AI Officer Alexandr Wang had told developers on X two days after Muse Spark’s April debut: “The Muse Spark API will be coming soon. We have been thrilled with the amount of excitement amongst developers who want to try Muse Spark inside their agentic harnesses. Stay tuned!”
The API never arrived. The timeline slipped from April to May, then to June, according to TipRanks.
More Tech Stocks:
- Morgan Stanley sets jaw-dropping Micron price target after event
- Nvidia’s China chip problem isn’t what most investors think
- Quantum Computing makes $110 million move nobody saw coming
A Meta spokesperson told Reuters the company is testing the API with early partners and expects to release it this month.
That statement is the company’s public position. It does not answer why a model unveiled in April still has no confirmed developer access date in June, according to The Next Web.
Why Muse Spark matters and what makes this delay more than routine
Muse Spark is not a minor model release. It is Meta’s first closed-source AI model, developed by the TBD Lab within Meta Superintelligence Labs, and it represents a deliberate strategic departure from the open-source Llama approach the company had built its AI reputation on. It was also the first major output from Meta Superintelligence Labs since Zuckerberg invested $14.3 billion in Scale AI and hired its founder Alexandr Wang to lead the division, according to TipRanks.
The closed-source decision makes the API delay especially consequential. For open-source models like Llama, developers can download the weights and run the model themselves.
For Muse Spark, the API is the only channel through which outside developers can access the technology. There is no workaround. Until Meta releases it, Muse Spark exists for the developer community in name only.
That gap matters commercially. OpenAI and Anthropic generate meaningful revenue by selling API access to businesses building AI-powered products.
Meta’s $125-145 billion capex guidance for 2026 requires a monetization path, and the Muse Spark API was supposed to be a primary part of it, The Next Web notes.
Economou/Getty Images
How the Muse Spark delay connects to Meta’s layoffs and AI spending
The context that gives this delay weight is what Meta told investors when it cut thousands of jobs earlier this year. The layoffs were not framed as cost-cutting.
They were framed as a resource reallocation: headcount redirected from lower-priority functions into AI infrastructure, research, and development. The implicit promise was that the AI output would justify the human cost.
A flagship model that cannot be accessed by developers two months after its debut does not immediately break that promise. But it does invite a question investors are now asking more openly: Is Meta’s AI spending translating into products that can generate revenue on a timeline that justifies the scale of the investment?
The WSJ report lands at a moment when that question is more sensitive than it has ever been, according to TipRanks.
Key context on Muse Spark, Meta Superintelligence Labs, and the API delay:
- Alexandr Wang, Meta’s Chief AI Officer, founded Scale AI and was hired by Zuckerberg in early 2025 with a $14.3 billion investment in Scale AI as part of the deal; Muse Spark is his first major model delivery at Meta, making the repeated API delays a direct test of his tenure and the thesis behind the acquisition, TipRanks said.
- Neither Meta nor the WSJ’s sources cited a performance problem or a safety hold as the reason for the delay; the only stated causes are bugs and infrastructure issues, which means the model itself is not known to be technically deficient, according to Reuters
- Meta abandoned a previous model codenamed Behemoth last year before committing to Muse Spark; that prior setback means this is the second time Meta Superintelligence Labs has faced a public stumble on a high-profile model, reports TipRanks.
- OpenAI, Anthropic, and Google have all made developer API access a launch-day feature for their frontier models; Meta is the only major AI company that has unveiled a flagship closed-source model and then withheld API access for nearly two months without a confirmed release date, a gap that is directly ceding developer mindshare to rivals, The Next Web notes.
- Meta stock carries a consensus Strong Buy rating on TipRanks based on 32 Buy and six Hold ratings assigned in the last three months; the delay has not yet materially altered analyst sentiment, but continued slippage would likely change that calculus, according to TipRanks.
What the Muse Spark delay means for Meta stock investors
The delay is not a crisis. Meta’s core advertising business continues generating substantial cash flow, and a two-month slip in a developer API is not the kind of event that typically moves the stock significantly.
Meta’s spokesperson says the API ships this month. If that holds, the story ends as a footnote.
The reason it deserves attention is the sequence it represents. Meta cut thousands of employees in May, explicitly linking those cuts to AI investment. It unveiled Muse Spark as proof the investment was producing results.
It then failed to deliver the API that would turn those results into revenue. That sequence, compressed into roughly six weeks, is exactly the kind of execution gap that investors will scrutinize as Meta’s capex commitments keep rising.
Zuckerberg has staked more than capital on the AI bet. He has staked the company’s workforce composition and its competitive positioning for the rest of the decade.
A delayed API on a flagship closed-source model is a small but specific piece of evidence about whether the execution is matching the ambition. Markets are patient with vision. They become less patient when products fall behind schedule.
Related: Mark Zuckerberg sends stunning message to Meta employees
