Mark Zuckerberg says over 10 million people used Meta’s Muse—and demand was so high, he had to get Stripe to raise credit limits

More than 10 million users in the U.S. and Canada have set up Meta’s Muse AI agents since Sept. 8, and one-quarter of adopters use them daily; shopping is the most common use, with average purchases above $200. Meta plans to charge businesses fees on sales facilitated by Muse, but the agent is compute-intensive: Q3 capex was almost $35 billion and free cash flow was negative $2.1 billion. Meta expects full-year capex of $135 billion to $145 billion, versus $72.2 billion last year, and says it aims to improve Muse’s cost efficiency by more than 10x over the next year.

Analysis

The strategic prize is not simply a new transaction fee: if Muse becomes the discovery layer, Meta could influence product ranking and capture purchase-intent data before retailers own the customer interaction. That could strengthen ad targeting and give Meta leverage to charge merchants, but it risks cannibalizing some ad spend or provoking retailers to limit access. The Personal Agent Protocol is therefore a distribution negotiation, not yet evidence of an open standard or durable monetization.

Near term, user and transaction figures are company-reported and do not establish incremental purchases, repeat use, merchant willingness to pay, or positive contribution margins. The key missing unit economics are fee rate, conversion lift, inference cost per completed task, and who bears refunds or disputes. A 10x efficiency target is a claim, not a margin forecast; serving more users could absorb the savings.

Over 1–3 months, watch for disclosed merchant fees, partner implementation, and evidence that usage persists beyond launch. Over 6–18 months, broad merchant participation could make Meta a commerce gatekeeper; resistance or poor agent reliability would favor retailer-owned channels. Shopify may gain if it supplies merchant infrastructure, but faces disintermediation risk if Meta owns discovery and customer data. Walmart could gain incremental demand while surrendering some control over merchandising and customer relationships. Avoid treating either partner’s participation as proof of net benefit.

Key Decisions for Investors

• META: Treat Muse as strategic upside, not yet an earnings driver. Do not add solely on adoption claims; revisit after Meta reports monetization and credible per-task economics. Falsify the bull case if fees fail to launch, repeat usage weakens, or AI costs pressure cash generation beyond guidance.
• Set an alert for the next 1–3 months: verify fee structure, merchant opt-in, incremental conversion versus existing Meta ads, and inference cost per completed task. Without those disclosures, keep the commerce thesis on watch rather than assigning a revenue multiple.
• SHOP: Avoid a reflexive short because it is a protocol partner. Track whether Shopify merchants retain checkout, customer data, and attribution; persistent loss of those controls would worsen the platform economics, while paid merchant tooling or incremental sales could offset the risk.
• WMT: Monitor implementation and whether agent-referred sales are incremental and economically attractive after any fees. Broad participation without evidence of incremental demand or retained customer economics would weaken the partnership read-through.