Meta’s first model from its Superintelligence Labs division still lacks a public API for developers nearly two months after launch. The Wall Street Journal reported on June 3 that persistent bugs and infrastructure problems have delayed the rollout multiple times, with no confirmed release date on the calendar.
What’s actually going on with Muse Spark
Muse Spark debuted on April 8, 2026, under the leadership of Chief AI Officer Alexandr Wang and represents a notable strategic pivot for Meta. Unlike the Llama series, Muse Spark is a closed-source model without open weights. That’s a deliberate choice, one that prioritizes monetization over community contribution.
A Meta spokesperson confirmed on June 4 that the API is currently being tested with early partners and is expected to launch sometime in June 2026.
For a company that poured $14.3 billion into Scale AI to bolster its artificial intelligence capabilities, the inability to ship a functional API on schedule raises fair questions about execution.
Why developers and investors should care
The closed-source nature of Muse Spark makes developer adoption particularly consequential. With Llama, Meta could afford slower commercial rollouts because the open-source community was doing much of the evangelism and adoption work organically. Developers were downloading the weights, building applications, and effectively marketing the model for free. Muse Spark doesn’t have that luxury. Its entire value proposition to Meta depends on controlled access, enterprise partnerships, and eventually, paid API usage.
For Meta’s stock and investor sentiment, the company has bet heavily on AI as its next growth engine, with the $14.3 billion Scale AI investment being just one piece of a much larger spending spree.
What this means for the broader AI race
The early partner testing that Meta confirmed is a standard part of any major API rollout. If Meta can deliver a stable API within June as its spokesperson suggested, the damage is likely contained.