Bitget App
Trade smarter
Buy cryptoMarketsTradeFuturesStocksEarnInstitutionAI & More
Up 36% in September! Meta achieves its best monthly performance since 2013, with market value approaching $2 trillion

Up 36% in September! Meta achieves its best monthly performance since 2013, with market value approaching $2 trillion

华尔街见闻华尔街见闻2026/09/25 02:06
Show original
By:华尔街见闻

Since September, Meta's stock price has rebounded significantly. The rapid rise of Muse has reignited market expectations for its AI commercialization. Muse reached 902,000 downloads within six days of launch and is quickly expanding into shopping, travel, and other scenarios. The market is beginning to bet that AI agents could unlock new growth opportunities beyond advertising. However, security and privacy risks, as well as high computing power costs, remain key factors in realizing these expectations.

Meta is currently experiencing a strong rebound driven by AI expectations, with the popularity of its personal AI assistant Muse serving as a key catalyst in this rally.

Since the beginning of this year, Meta's stock price has been under considerable pressure, but it has rebounded rapidly since September. As of September 24th, the stock has surged about 36% this month, marking its strongest monthly performance since 2013. Its market capitalization is once again approaching $2 trillion, just about 1% away from this threshold.

Behind this market reversal, the AI commercialization potential brought by Muse has become an important driver. Muse officially launched on September 8th and is capable of independently carrying out multi-step tasks such as shopping, travel arrangements, and email management. Within the first six days of launch, downloads reached 902,000, surpassing the 773,000 achieved by Meta AI during the same period, and quickly topped the free app charts in the US and Canadian Apple App Stores.

However, market optimism is heating up rapidly. AI agents are beginning to penetrate real consumer scenarios such as shopping and reservations, which may disrupt the traditional Internet platforms' reliance on user search, account, and usage habits; at the same time, Muse's own security vulnerabilities, privacy concerns, and high computing power costs also present real-world constraints on this AI boom.

Up 36% in September! Meta achieves its best monthly performance since 2013, with market value approaching $2 trillion image 0

From Weakness to a Strong Rebound This Year, Meta Sees an AI Catalyst

Meta was not previously a hot AI stock in the market. Before August 18th this year, its share price had fallen 18% from the beginning of the year, ranking among the 50 worst performers in the S&P 500 index year-to-date. The revenue guidance released at the end of July was below market expectations, which also dampened investor sentiment.

Subsequently, two factors triggered a turnaround in the market. On one hand, Meta agreed to pay up to $18 billion to settle lawsuits related to social media, easing long-standing legal uncertainties to some extent; on the other, the launch of Muse quickly gained user attention, providing the market with a new AI growth narrative.

Since its late August low, Meta's share price has rebounded by around 43%. On September 10th, JPMorgan analyst Doug Anmuth upgraded Meta from Neutral to Overweight, believing the company is at an early stage of launching advanced models and AI products beyond advertising, with significant room for growth remaining.

This means the market's focus is shifting from Meta's traditional advertising business growth to whether AI can bring new users, revenue, and valuation space.

Muse Moves from 'Chatting' to 'Execution' as AI Agents Enter Consumer Scenarios

Compared to traditional chatbots, Muse emphasizes directly completing tasks for users. Users can have it handle multi-step workflows such as placing shopping orders, travel bookings, and drafting emails. Muse currently offers free tiers, with paid plans at $20 and $100 per month.

After launching, Meta has continued to expand Muse's capabilities, including testing phone call features to US businesses, opening connectors for developer integration, releasing a Mac version of Muse, and integrating Shop Pay for agent-based checkout. During Meta Connect, the company also introduced the hand-held device "Muse Charm" for operating Muse, and released a new version of its smart glasses without a camera.

The business ecosystem is also expanding in parallel. Meta has established a grocery sales partnership with Instacart’s parent company Maplebear and a collaboration with Expedia, further extending Muse from a pure AI assistant to a platform connecting consumer services.

JPMorgan expects that Muse could become one of the most widely used consumer AI applications after ChatGPT, with daily downloads now surpassing the total of Meta’s other apps. However, the bank predicts that Meta will continue to prioritize user growth and engagement in the short term, and formal commercialization may not occur before 2027.

Up 36% in September! Meta achieves its best monthly performance since 2013, with market value approaching $2 trillion image 1

AI Agents Disrupt 'Consumer Inertia', Traditional Platforms Begin to Feel Pressure

Muse's popularity is also prompting the market to re-examine the traditional consumer internet model. Goldman Sachs' trading desk recently pointed out that the potential impact of agent-based AI on "consumer inertia" stocks has become a hot market topic, with its high switching cost enterprise basket index dropping 2.58% in a single day.

The core market concern is that if AI agents can directly perform search, shopping, reservation, and other tasks for users, the usage patterns and account stickiness that users have developed around specific platforms may be weakened.

Meta is also proactively expanding the impact. By integrating services like Maplebear and Expedia, Muse is entering more concrete links in the consumer chain. For traditional internet platforms, whether users will directly open the app in the future or let an AI agent handle tasks on their behalf may become a new competitive variable.

However, security and privacy issues with Muse are also being quickly exposed.

Researchers have revealed that Muse's macOS version contains a zero-day vulnerability, with the potential for local applications or terminal commands to obtain tokens used to authenticate user accounts and further control the account. In terms of privacy, some users claim that Muse reads Mac notification content and pushes related suggestions without explicit authorization.

Additionally, code analysis has found that Muse’s automated browser has mechanisms to avoid bot detection, modifying some browser characteristics to reduce the likelihood of being identified. Amazon has since blocked Muse from accessing its retail website, and the related automated activities have become a focus of external attention.

Another Hurdle for AI Agents: Computing Power Costs

Beyond security issues, Muse’s business model also faces the challenge of infrastructure costs.

According to the product design, each user theoretically has an exclusive cloud virtual machine, with around 2 virtual CPUs, 8GB of memory, and 100GB of solid-state storage, and it must operate 24/7.

A rough calculation shows that if Muse reaches 100 million users, Meta might need to deploy about 1.6 million AMD EPYC processors, 100PB of memory, and 10,000PB of solid-state storage, with power demand of around 1.6 gigawatts. Meta’s overall electricity planning for this year is about 7 gigawatts, which also needs to meet the demands of model training and advertising businesses.

The revenue side currently cannot cover such massive investments. Assuming 10 million users subscribe to the $20/month paid tier, annualized revenue is about $2.4 billion; yet Meta’s capital expenditure this year is as high as $130-145 billion. If the dedicated virtual machine model is maintained, Muse’s early infrastructure investment will still be considerable.

This is also what the market needs to further verify: the number of users an AI agent can bring does not necessarily translate to profit generation.

Valuation Not Aggressive, but AI Investment Is Testing Cash Flow

Meta's current valuation remains relatively moderate. Based on a Bloomberg consensus of analyst estimates, the company's forward 12-month P/E ratio is about 21, slightly lower than the Nasdaq 100 index's 22.

Analysts expect Meta’s revenue to grow 26% to $254 billion by 2026, with net profit rising 33% to $80.6 billion; by 2027, revenue and profit growth rates are expected to slow to 20% and 9% respectively.

At the same time, AI infrastructure investment is clearly squeezing cash flow. Meta’s free cash flow last year was $46 billion, but due to surging capital expenditures, it is projected to drop to -$6.4 billion in 2026 and further to -$29.2 billion in 2027.

Currently, over 90% of analysts tracked by Bloomberg maintain a Buy rating on Meta, but the stock price is already close to their average target price. After this round of rapid appreciation, market expectations for Meta’s AI prospects have significantly increased. Whether Muse can achieve stable monetization beyond user growth will be key for further valuation expansion in the next phase.

0
0

Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.

Understand the market, then trade.
Bitget offers one-stop trading for cryptocurrencies, stocks, and gold.
Trade now!

You may also like

UBS: US Companies’ Capex Willingness Rebounds from Bottom, Further Rate Hikes May Have Limited Impact on Economy

UBS believes that traditional U.S. investments have been subdued for a long time and are at a low level, with limited room for further rate hikes to have an impact. Meanwhile, corporate operating cash flow has increased by 25%, providing sufficient internal funds, which may weaken the transmission effect of interest rates on Capex. As capital expenditure willingness rebounds from historic lows and non-AI investments show signs of stabilization, U.S. corporate investment may be entering a recovery phase.

华尔街见闻•2026/09/25 08:56
UBS: US Companies’ Capex Willingness Rebounds from Bottom, Further Rate Hikes May Have Limited Impact on Economy

As Nscale submits its IPO application, Bernstein "diagnoses" AI computing power cloud: CoreWeave (CRWV.US) "underperforms the market," IREN (IREN.US) "outperforms the market"

Last Friday, the emerging British AI computing power company Nscale officially submitted its S1 prospectus. This long-anticipated document offers a new reference point for Bernstein to analyze the business model, competitive barriers, and potential risks of the new generation of computing power cloud sector.

智通财经•2026/09/25 08:26