Muse has gone viral: Is this a real inflection point for Meta or just a false boom?

It has been over two weeks since Meta launched its new product in the US region on September 8. Dolphin Research previously released a brief comment on it, but as user evaluations keep increasing, its popularity has remained unabated in recent days. Even today, Muse still ranks No.1 on the free app download chart for iOS in the US.

This has directly driven the volatility in the stock prices of Muse's backend industrial chain, front-end business ecosystem and related mapped targets. Today, Dolphin Research will first conduct a detailed discussion on the product side, and the sorting out of changes in the backend industrial chain is coming soon.

I. The skyrocketing popularity of Muse is not unexpected

2C Agent is an AI application track that never lacks popularity, but compared with 2B Agent which has been thriving in paid revenue, the implementation of business models for lifestyle-focused 2C Agents has never achieved a real breakthrough.

Therefore, over the past six months, major tech giants have quietly adjusted their direction for 2C Agents, shifting their internal resource focus to the workflow scenarios on the consumer side. The progress of general lifestyle 2C Agents has been visibly slowed down due to commercialization constraints.

Only Meta has been consistently emphasizing the importance of developing personal AI lifestyle assistants. The management has mentioned the concept of "everyone having an AI assistant" on earnings calls for at least two years. Muse is built with long-term efforts from the company's core team, ranking as the core project with the highest internal resource priority, and it is also a key product layout of Meta this year.

Of course, this is also a direction that Meta, which lacks productivity scenarios by nature, has no choice but to bet on. Otherwise, the new billion-user-scale entry point ChatGPT will seize users' attention, which will cause long-term damage to its own social ecosystem.

So why did Muse achieve instant success? What are its advantages?

1. Cognition Rewriting: Chat Q&A Type vs. Active Task Type

As a task-based personal AI assistant, Muse is fundamentally different from users' existing product cognition of ChatGPT in its underlying design. Although ChatGPT can also handle some tasks, and can execute more long-task workflows when paired with Codex, most users still habitually classify ChatGPT as a chatbot in non-work scenarios, with interactions limited to question-and-answer forms.

But Muse, positioned as a lifestyle assistant, was never designed as a Chatbot from the very beginning. In its promotion and introduction, Muse demonstrates its functions and positioning by giving examples of different tasks it can complete.

However, on the underlying large model, Muse is still supported by the public Muse Spark large model, without carrying out lightweight processing such as dedicated model tuning and parameter streamlining. At present, Muse Spark 1.3 ranks in the global first tier in terms of intelligence scoring, which to a certain extent guarantees the capabilities of complex instruction understanding, cross-platform operation and multi-scenario adaptation, and underpins the product experience.

2. Balance between User Data Memory and Privacy Security

Speaking of task-based AI assistants, several similar products have actually been in testing or full launch in North America before Muse was released. Typical examples include Instinct and Grok bot.

Instinct is developed by a startup, released in February and is still in the invitation testing phase, which has led to a sharp rise in its valuation in the primary market. But compared with Muse developed by large companies, Instinct has a relatively bold style, is less cautious about user privacy and security, and has a larger authorization scope. It may feel "smarter" during experience, but the platform has not adopted relatively comprehensive protection measures for risks brought by the AI's autonomous decision-making.

Grok bot was launched in August, which is also an AI robot that handles tasks for users. Users can authorize Grok bot to access Gmail, Google Calendar, Notion, Figma, or other software such as finance and AI code assistants to execute tasks.

However, compared with Muse, Grok bot is more of a workflow-focused AI assistant, which may not show significant differentiation from the currently well-known Openclaw-type platforms.

Detailed comparison of other indicators can be referred to the figure below:

Dolphin Research believes that Muse's greater advantage may lie in its "years of accumulation of user data", and it has achieved a proper middle balance between the utilization of these data and "user privacy protection".

(1) User Data Accumulation

With exclusive user social data and ecological advantages, Muse has three substantial differences.

a. Long-term Memory: Muse has the ability to memorize long-term context, and tasks continue to advance normally even after users close the App. In its functional modules, Goals is responsible for tracking long-term targets, and Artifacts precipitates results into reusable itineraries, checklists and panels;

b. Real Account Connection: Muse accesses Gmail, Calendar, OpenTable, Spotify, Amazon, as well as Meta's own IG, WhatsApp and Marketplace through three methods: built-in connectors, official APIs and browser operations;

Especially by connecting its own social accounts, Muse has relatively high reading permissions. After user authorization, it can directly read public or semi-public data such as Instagram posts, follower data and comments. It also supports reading WhatsApp chat records and message content, automatically sorting out users' social messages, extracting to-do items and summarizing important information.

c. Active Recommendation: Based on real account connection and long-term context memory, Muse can actively put forward task suggestions according to the calendar, conversations and connected accounts, forming the rudiment of a progressive Agent, and also reserving space for future advertising monetization.

(2) User Privacy Protection

Privacy security and risk control systems are core links that may cause problems for AI applications. For Meta, as a large company, the compliant use of user private data can easily become a target for criticism of its products.

Mark Zuckerberg claimed in the middle of the year that the development progress of AI assistants fell short of expectations, and the launch was postponed to September precisely because the problems of AI authorization and risk control needed to be solved. At present, Muse is only launched in the US region, which is very likely because the team is continuously optimizing privacy protection and AI authorization mechanisms.

To balance functional experience and privacy security, Muse's isolation and protection of user privacy are mainly reflected in two aspects:

a. Isolated Storage: Muse provides each user with an independent and exclusive Secure VM (secure virtual machine), with the Sentinel agent guarding all external actions. The platform sets up a dedicated secure storage partition to strictly separate the data accessible to the model from users' core private data (such as account passwords and payment credentials).

All data interactions and function calls of third-party platforms are transferred through encrypted secure APIs to avoid exposing sensitive data to the large model.

b. Hierarchical Authorization: The reading of real account data mentioned above requires user authorization. In Meta's own ecological accounts, the function of hierarchical user authorization is provided. For example, users can separately turn off the permission to read Instagram private messages, and only authorize the reading of public posts, or separately turn off the permission to read WhatsApp messages.

The figure below shows the security architecture diagram disclosed by Meta.

(1) Gray line process: A normal instruction goes from the Main UI to the user's virtual machine VM, the hatch daemon (Agent ontology) calls the external large model for reasoning, then the hatch safety (risk control review) scores and evaluates the risk. All actions and network requests of the Agent in the VM are directed by the eBPF agent to Sentinel for inspection.

(2) Blue line process: When encountering sensitive actions such as payment, Sentinel will intercept it and push it to the Approvals in the user control interface. After the user's permission, Sentinel issues the license and obtains payment credentials such as passwords from the Secure Credential Storage safe.

The Agent does not see the user's password and card number throughout the whole process, but the setting of a resident virtual machine for each user will increase component costs.

II. How much incremental value can Muse bring to Meta?

At present, Muse is mainly in public beta for US users aged 18 and above. It adopts a tiered pricing model based on Token consumption demand, with no restrictions on functions. In addition, it is expected that Muse will later charge a certain commission on the transaction volume of merchants, while retaining its most familiar advertising display business.

1. Subscription: Hard to be optimistic in the short term

Although the first batch of users gave positive feedback and North America has stable subscription habits, Dolphin Research still believes that the direct monetization (payment) on the C-end needs to be treated cautiously in the short term:

According to Sensor Tower data (only counting the iOS end), Muse's DAU is nearly 250,000 people at present, and its single-day revenue reaches 2,000-3,000 US dollars per day. Although the current data is only from a partial test, we can roughly perceive the payment situation:

Calculated based on an average of 2,000 US dollars per day for two weeks and a minimum paid package of 20 US dollars per month, the number of paying users is 1,500. Assuming that the iOS end accounts for 1/3 of the total paying population, the total paying population across all ends is 4,500 people. With an average DAU of 150,000 in the past two weeks, the payment rate is only 3%.

However, considering that the package is calculated at the minimum standard and the iOS payment proportion is generally more than 1/3, the actual payment rate is significantly lower. At the same time, by comparison, ChatGPT covers more relatively complex productivity scenario demands, and its payment rate is only 5% under the high stickiness feature of 1 billion weekly active users.

Therefore, Dolphin Research still suggests not being overly optimistic about the contribution of subscription payment. We estimate that most users will stay at the free tier (100 million Tokens quota per week). Even under a very optimistic assumption in the final state, with 3 billion users, 80% penetration rate, 2% payment rate (the monthly active user payment rate is lower than the daily active user payment rate), and an average ARPPU of 20 US dollars per month, the annual new revenue is 11.5 billion US dollars. For Meta's current revenue volume of 250 billion US dollars, the increment is only 5%.

This calculation does not include the underlying costs, including computing power costs and exclusive virtual machine component costs. Therefore, the final profit brought by subscriptions is only a negligible contribution.

2. E-commerce: How to break through ecological restrictions?

According to actual tests at present, e-commerce shopping is a relatively mature scenario that has the potential to expand commercial space (other high-frequency usage scenarios in the top ranking are mainly productivity-focused scenarios such as accounting processing, document processing and tax filing), which mainly relies on the integration of Muse with Meta's existing products and business ecosystem.

In addition to Muse, Meta currently also provides Meta Business Agent and Meta Business Agent Platform.

Meta Business Agent is an "AI shop assistant" that can be called at any time, which can be activated by direct subscription and payment. The AI shop assistant can perform general functions such as answering questions, recommending products, booking services and screening potential customers, and is more suitable for small and medium-sized merchants and individual merchants. In Q2 2026, more than 1 million small and medium-sized enterprises used Meta Business Agent through WhatsApp and Messenger every week.

Meta Business Agent Platform is an API-based enterprise-level development platform, where large enterprises can customize AI shop assistants according to their own businesses and then connect to their own systems. The targeted large enterprise customers need to have certain IT configuration capabilities.

Therefore, the 2C-end Muse can be connected with the 2B-oriented Business Agent, so that the Agent (AI assistant) on the Muse end can dock with the Agent (AI shop assistant) on the merchant end, thus forming a complete e-commerce service chain from content seeding to transaction completion.

However, there is an obvious loophole in the above commercialization process:

Since Meta does not have its own e-commerce platform, if users have clear shopping targets and the merchants are independent station merchants, a three-party closed-loop ecosystem can be formed. But for merchants hosted on e-commerce platforms, this becomes a four-party ecosystem. Whether Muse is allowed to perform access and payment actions depends on whether the e-commerce platforms acquiesce to such operations.

Theoretically, Muse's shopping experience intercepts users' operations such as search, browsing and order placement on e-commerce platforms, which will directly affect the advertising revenue of e-commerce platforms.

Therefore, on September 20, Amazon announced that it would block Muse, which means that Muse cannot perform operations such as browsing for price comparison and making payments on Amazon, and a violation warning window will pop up on the user side (as shown in the figure above). It is worth mentioning that not only Muse, Amazon has similar restrictions on the shopping Agents of OpenAI and Perplexity.

However, small and medium-sized e-commerce platforms, especially those with a high proportion of self-operated businesses, are more inclined to join Muse, essentially taking a fancy to Meta's super App traffic. The e-commerce platforms that have joined at present include Shopify and Etsy, which are platforms with their own inventory and doing 1P business. Essentially, they are the first large merchants themselves. Meta has a very strong subjective willingness to import full-scale social traffic to Muse, which is undoubtedly a considerable dividend for Shopify.

This loophole mainly affects the transaction commission commission or advertising effect in the above four business models, but its impact on the call of B-end Agents is not too great.

If we map this scenario to the Chinese market, Tencent's WeChat ecosystem is a perfect counterpart that has such private domain data, a relatively large public domain business ecosystem, and no troubles of ecological restrictions (Muse - WeChat Mini Program intelligent assistant, independent station merchants - WeChat Shop, Shopify - Mini Program). But Tencent may not be so radical in terms of cross-service privacy authorization in the short term.

At the same time, there are also considerations on the cost side (according to research information, when Muse performs simple tasks, the cost per user may be as high as 50