$META’s first earnings-call question gets to the economics of scaling Muse.
Morgan Stanley’s Brian Nowak asked what constrains user growth and what Meta wants to build in enterprise AI.
Zuckerberg said Muse passed 10M setups in the U.S. and Canada in its first six weeks, but compute intensity is one reason it hasn’t launched globally. Local optimization and connectors also need work.
Serving a potential billion-plus people depends on hitting compute-efficiency targets. Mark expects that work to remain a limiting factor over the next year, alongside building the business around the product.
There’s already a product ladder: Spark powers the free tier, Wonder is available in the paid tier, and Meta is developing a more efficient Wonder version for broader distribution.
For enterprise, the near-term offering is model APIs, Muse Code and business agents. Longer term, Mark wants consumer and business agents interacting directly, including for commerce.
CFO Susan Li added the financial framework: diversified recurring revenue through subscriptions, usage-based APIs, messaging and other services. She expects that opportunity to develop over multiple years, with investment scaling behind demonstrated customer demand and product-market fit.
That answer puts two practical conditions on expansion: cheaper compute and evidence customers want the products. Li didn’t provide a new enterprise revenue target or capex figure.