Meta closed regular trading on the New York market at $741.24 on Sept. 21 local time, up 11.43% from the previous session. Its market capitalization soared to $1.89 trillion.
The stock's surge stems from Muse's runaway popularity. According to Bloomberg and market research firm Sensor Tower, the Muse application ranked No. 1 in free app downloads on both Apple's U.S. App Store (iOS) and Google Play as of that day.
It reached that mark just two weeks after launch. Muse was downloaded more than 902,000 times within six days of release, outpacing rival models such as OpenAI's ChatGPT as well as betting apps including Polymarket and Kalshi. Because Muse can also be used within Meta's messaging app WhatsApp, actual users are likely to exceed download figures, observers say.
Meta launched Muse, a personal AI agent that performs tasks on a user's behalf, on Sept. 8. While Anthropic and OpenAI have led the market with enterprise AI agents, Meta sought differentiation with an agent focused on personal tasks. For example, the agent can draft and send emails, make travel reservations and assist with online shopping.
Muse's strength lies in its ability to respond based on Meta's own social media properties, such as Facebook Marketplace and Instagram. Users can also connect personal services including Gmail and Calendar. Muse is free to use, but paying $20 or $100 a month grants more tokens. Meta's free AI app, released in April last year, offered image and short-video generation but no personalized services.
Analysts say Meta Chief Executive Mark Zuckerberg's vision of a personalized AI assistant has won over consumers. Despite pouring huge sums into AI, Meta had failed to stand out in the market, but Muse has turned the mood around, they say. Evercore ISI said in a report that "Meta has scored a big success with Muse," calling it "an example showing that Meta's $200 billion AI investment was not in vain." Bernstein analyst Stacy Rasgon said in a CNBC interview, "Most agent use cases so far have not been for ordinary people," adding, "But now a number of agents, including Meta's Muse, are beginning to show the potential for much broader adoption."
Muse's success has also been a boon for CPU makers. Running AI agents depends on inference, and CPUs play a central role in developing inference capability. CPUs are essential to coordinating large numbers of graphics processing units so they work in concert, while also managing various applications. As the agentic AI era arrives, CPUs have gained stature alongside GPUs used for training.
AMD, often called Nvidia's rival, closed at $615.52, up 9.95% from the previous session, pushing its market capitalization past $1 trillion. It became the fourth U.S. chipmaker to join the "trillion-dollar club," after Nvidia, Broadcom and Micron. Along with Intel, AMD is a leading maker of CPUs for AI servers. Intel also finished 12.14% higher at $121.78.
Arm Holdings shares rose 17.16% to $322.90. Arm, long the dominant force in smartphone architecture — the guidelines that determine how chips operate — declared its entry into the server CPU market in March this year when it unveiled its first in-house AI chip for AI data center servers. That marks the emergence of a formidable competitor to Intel and AMD.
