On September 8, Meta launched a free app called Muse. Three weeks later, the company was worth roughly $457 billion more than it had been at the end of August. The stock closed September 25 at $751.66, up 31.3% for the month, its best monthly performance in 13 years. This is the story of how a single product launch moved nearly half a trillion dollars, what it reveals about the AI economy, and the bill that is still coming due.
Muse is Meta’s personal AI agent, available in its own app and inside WhatsApp. Ask it to book a trip, sort your email, or buy something, and it does the task itself rather than just answering questions. It launched in the United States on September 8 and in Canada on September 18, the only two countries where it is available, according to Tech Xplore. It is free to use with limits, and Meta offers optional subscriptions at $20 or $100 a month for heavier use. Restricted to users over 18.
The adoption numbers did the early talking. Within two weeks, Muse had been downloaded 1.8 million times on iPhones in North America, compared with 1.3 million downloads for ChatGPT’s app when it launched in May 2023, according to data firm Apptopia. It climbed to first place among free apps in the U.S. App Store, ahead of ChatGPT itself, reaching about 2.8 million downloads by September 22. The stock’s biggest single day came on September 21, when it jumped 11.3%, its largest one-day gain since April 2025, adding roughly $192 billion in market value in one session. Wells Fargo raised its price target from $640 to $796 the same week, pointing to the launch and better odds that Meta’s AI products can actually make money.
Then came the Connect conference on September 23, where Meta announced Muse would soon arrive in its smart glasses and revealed partnerships with retailers including Walmart, opening the door to commissions on purchases Muse makes for users. That is the sentence that turned a fun app into a business model in investors’ minds: an agent that shops for you can take a cut of everything it buys.
And the market immediately started repricing everyone in Muse’s path. On September 22, Expedia fell 3.7% and Booking Holdings dropped 3.9%, as investors absorbed the news that Muse connects to more than 500 airlines through a Duffel integration and can search, compare, and book travel without anyone visiting a travel site. Banks felt it too: the S&P 500 Financials index slid 2.4% to its lowest since July, with JPMorgan and Wells Fargo each down more than 2.5% and Allstate and Charles Schwab dropping over 5%. Wolfe Research sketched a $30 billion to $50 billion addressable market shifting toward agents, and some analysts warned up to $50 billion in market value was at risk across travel and finance. The logic is uncomfortable for incumbents: if your business is essentially a tax on people not wanting to do boring tasks themselves, an agent that does the tasks is an existential audit.
Step back and the arc of Meta’s year makes this rally even more striking. Shares had fallen as low as $525.72 in late March after court rulings about the impact of its networks on young people’s mental health, and dropped again in late July when earnings were weighed down by AI spending that could reach $145 billion this year. The company guided 2026 capital spending of $130 billion to $145 billion while second-quarter free cash flow fell to just $784 million. That tension has not gone away. It has just been drowned out by applause.
So what are the lessons? First, distribution is destiny. Meta did not win September with a better model; it won by putting its agent in front of billions of people across Facebook, Instagram, and WhatsApp at zero customer-acquisition cost. Second, the market will pay a platform multiple the moment it believes the revenue mix can change. Investors spent years valuing Meta as an advertising business with an expensive AI hobby; Muse let them imagine it as the toll booth for agentic commerce. Third, disruption gets priced in days now, not years. Travel and banking stocks did not wait for Muse to actually take bookings before falling; the mere plausibility was enough. And fourth, gravity still exists. Meta shares fell 3.3% on Friday, a reminder that a 31% month needs constant good news to sustain.
Here is my honest take: the $457 billion is half substance, half story. The substance is real. Two million-plus downloads in two weeks, the top of the App Store, and a genuine new monetization path through commerce commissions are the strongest evidence yet that Meta’s AI spending can convert into revenue. The story part is the assumption that it all scales smoothly: that users keep the agent once the novelty fades, that regulators stay quiet, that the $130 billion-plus spending plan does not swamp the returns. September rewarded the believers. The next few quarters will test whether the believers did their math.
Watch three things next. First, Meta’s third-quarter earnings: does management disclose Muse user numbers, or keep them hidden? Second, whether the Walmart-style commerce partnerships start producing reported revenue, because commissions are the bridge from downloads to dollars. Third, regulators. An agent that books travel, moves money, and shops for billions of people will attract scrutiny the way social feeds once did. The download chart was September’s story. The earnings call is October’s.
Either way, every boardroom in travel, banking, and retail just got the same memo: the agent economy is not a forecast anymore. It is a download chart.
