“The centerpiece of our vision for what we’re building.” said Meta CEO Mark Zuckerberg on Wednesday.
Meta’s Muse gains rapid traction
Data from Sensor Tower provided to CNBC shows that it took Anthropic’s Claude 670 days to hit 9 million U.S. downloads.
The agent’s growth has fueled investor interest, even as Meta’s stock slumped 6% on Wednesday despite a solid earnings report. Meta CEO Mark Zuckerberg said that the speed of adoption required the company to work with payment processor Stripe to increase credit limits.
Consensus forecasts suggest that Muse and related revenue could help drive Meta’s total revenue from $255 billion this year to $430 billion in 2029. Deutsche Bank’s Benjamin Black stated in a recent note that Muse has put Meta “back in the frontier discussion,” and he maintains a $820 target and a buy rating on the stock.
Apple targets agentic AI development
Apple is working to enter the market, with Bloomberg reporting the company is developing an autonomous AI agent capable of controlling devices, interacting with applications, browsing the web, and writing code.

Mark Gurman of Bloomberg reported on 19 November that the system is designed to perform complex tasks, including debugging code and calling services. While Apple has previously been reported to be working on a proactive Siri AI agent for 2027, this new report describes a more capable system styled after computer use.
The listing describes the goal as creating an agent that can use a computer like a capable person, and the leader will be responsible for the delivery of the platform and its AI capabilities to ensure they are “capable, reliable and safe.” The report indicates the agent may not be generally available until late next year.
Evercore ISI analyst Amit Daryanani said that Apple possesses a competitive advantage through its integration of personal data, applications, and Apple Pay. The firm’s analysis highlights several potential revenue and user metrics for Apple’s AI strategy:
| Metric Category | Estimated Figure |
|---|---|
| Compatible iPhone users worldwide | 550 million to 560 million |
| Active iPhone users | 1.4 billion |
| Muse user base | 10 million |
| Instinct user base | 100,000 |
Evercore’s financial modeling for Apple’s AI strategy includes a subscription scenario projecting $21 billion in annual revenue from 125 million customers paying $14 per month. Using a different transaction-based approach, an annual projection of $45 billion is possible if 125 million monthly users each spend $600 on qualifying purchases and Apple takes a hypothetical 5% fee. To support these agents, Apple could distribute workloads across its Private Cloud Compute infrastructure, on-device processing, and external cloud providers.
Agentic AI tools drive infrastructure build-out
The rise of autonomous agents has created a growing market that includes heavy hitters such as OpenAI’s Dots, Meta’s Muse, and Google’s Gemini Spark. Google introduced a new feature for its Gemini AI assistant on Thursday, enabling it to perform tasks across various applications and operate Mac computers.

SpaceXAI's Grok Bot is also a potential rising player. “Agentic AI tools are what is driving this next era of AI infrastructure build-out because they are driving usage in a way that expands the amount of AI compute that's necessary,” Bob O'Donnell, president of Technalysis Research, said.
While agents may benefit small businesses, they pose threats to proprietary application systems and large e-commerce sites. Companies selling personal agents also aim to gain e-commerce power by acting as brokers, shoppers, and task managers to capture fees from online exchanges.
Beyond commercial implications, Apple has shown caution regarding agentic AI, recently warning developers that such tools “could put users at risk” by unwittingly allowing Full Disk Access. To mitigate this, Apple rolled out new controls earlier this week to provide individuals with better oversight regarding the data AI agents can access.
Privacy is a major concern. “You could have the AI agent collecting its own information and also combining what it learns with information it buys from data brokers. So it can get extremely, extremely specific about you in ways that I don't think users are aware of,” Calli Schroeder, senior counsel at Electronic Privacy Information Center, said.
Earlier coverage: Apple Bets on Siri AI Apps as the Main Draw for Its Smart Glasses