Citi expects Meta's month-old Muse to earn $23 billion a year from transactions by 2030

Start with the smaller line, because it is easier to check. Subscriptions worth $4.5 billion a year ^[3] would take 3.75 million people on the $100 plan ^[17]. That is more than twice the number who open Muse on a given day now ^[18]. Bank of America Institute data found that the 3% of households paying for any AI service spend a median $20 a month ^[12]. At that price the same line needs 18.75 million payers ^[19].

Transactions carry about 84% of the forecast ^[21]. Citi describes Muse as a tool people use to search, compare and transact online ^[7]. Personal agents, "with Muse benefiting from its 1st-mover advantage, are emerging as the new front door to the internet," Citi said ^[6]. In that framing Muse sits where a search box sits, at the start of a purchase, and Meta's revenue is a cut of what follows ^[16]. The reports do not give the commission rate or the purchase volume Citi assumes. Spread $23 billion across today's 1.8 million daily users and it comes to roughly $12,800 a user a year ^[22], so the estimate depends on a user base many times its current size.

The usage record is thin. Muse has 1.8 million daily users from 6.6 million downloads, about 27% ^[23], in its first month ^[8]. OpenAI said in February it had more than 900 million weekly users and upwards of 50 million paying subscribers ^[13], roughly one payer for every 18 users. Apply that ratio to Muse and 3.75 million subscribers on the $100 plan would need about 67.5 million weekly users ^[20].

Citi's line holds if daily users multiply and merchants accept Meta's terms. Muse fails slowly if users arrive but do their buying elsewhere, leaving Meta with a feature that deepens engagement across its apps ^[10] and earns little directly. It fails faster if OpenAI's ChatGPT Dots, Google's Gemini and Anthropic's Claude widen their own agent offerings ^[9] quickly enough that being first counts for little. "That leaves the $100 plan and merchant commissions to carry a product whose costs rise with every download the charts are celebrating," PYMNTS wrote ^[16].

I think Citi is modelling the right revenue line. With free use the default ^[15], commissions are the part of Muse that grows as free use grows. The $27 billion is still a price set on a product launched four weeks before the note ^[1][8]. The case weakens if the next usage update shows downloads climbing while daily users stay near 1.8 million, pulling that 27% ratio down ^[23]. It fails if Meta cannot show purchases moving through the agent at all.