Superpower Daily

Superpower Daily

Your reading space

Sign in
Garry Tan Says Regulators Shouldn’t Curb Alleged Model Distilla...GSA Sets October Shift to Zero-Fee, Usage-Based OpenAI PricingPentagon Negotiates $5B Fluidstack Loan for AI Power and Coolin...TSMC Reports Record August Revenue as AI-Chip Demand Holds Stro...Anthropic Says It Removed Claude Accounts Tied to Nine Influenc...Enflame Shares Rise 206% in Shanghai Debut, Backing Its Next AI...Garry Tan Says Regulators Shouldn’t Curb Alleged Model Distilla...GSA Sets October Shift to Zero-Fee, Usage-Based OpenAI PricingPentagon Negotiates $5B Fluidstack Loan for AI Power and Coolin...TSMC Reports Record August Revenue as AI-Chip Demand Holds Stro...Anthropic Says It Removed Claude Accounts Tied to Nine Influenc...Enflame Shares Rise 206% in Shanghai Debut, Backing Its Next AI...
Swipe

JPMorgan Upgrades Meta, but Warns AI Compute Could Hit Cash Flow

The bank sees early demand for Meta’s new personal AI agent and more upside from AI products, while forecasting that the infrastructure behind them could drain cash before direct monetization arrives.

By 2 min read
JPMorgan Upgrades Meta, but Warns AI Compute Could Hit Cash Flow
JPMorgan Upgrades Meta, but Warns AI Compute Could Hit Cash Flow

Listen to this story

The audio brief

About 1:37
0:001:37
Read transcript
JPMorgan has upgraded Meta to Buy and set an $820 price target, citing early traction for Muse, the company’s new personal AI agent. But the bullish call comes with a major cash-flow warning: the bank expects Meta’s AI infrastructure spending to produce an annual free-cash-flow deficit of $65 billion to $75 billion in both 2027 and 2028, before counting any revenue from AI products. Muse launched on September 8 in the United States across iOS, Android, and muse.ai, with free access and subscription plans. Meta says the agent can take actions, including sending email and booking travel, through a dedicated secure virtual machine. JPMorgan analyst Doug Anmuth said Muse reached as high as number three in the U.S. App Store on its second day, while early usage was running at ten times the level of Meta’s training cohorts. That signal supports the upgrade, but subscriptions are not the only potential payoff. JPMorgan also sees AI improving Meta’s core advertising business through recommendations, engagement, ad targeting, retrieval, and AI-generated content. Meta has additionally made Muse Spark 1.3’s maximum reasoning available through Muse Code and the Meta Model API, for longer-running tasks across multiple workflows in one thread. The key question is whether early Muse interest becomes durable demand, and whether those broader AI gains arrive before the compute bill puts sustained pressure on cash flow.

Story brief

3 key points

Meta’s new AI bet is already affecting the investment case: JPMorgan now rates the company Buy and assigns an $820 target, but models a $65 billion–$75 billion annual free-cash-flow deficit in 2027 and 2028 before AI-product revenue. Muse’s early distribution signal—reportedly reaching No. 3 on the U.S. App Store and usage 10× training cohorts—supports the upgrade, while the costs highlight the unresolved question:...

  1. 01

    Muse launched September 8 across iOS, Android, and muse.ai, offering free access alongside subscription plans.

  2. 02

    The agent can send email and book travel through a dedicated secure virtual machine.

  3. 03

    JPMorgan’s $820 target implied roughly 30% upside from Meta’s Wednesday close.

JPMorgan upgraded Meta to buy from neutral and raised its price target to $820, citing early traction for the company’s Muse personal AI agent and its broader push into AI products. The bank’s bullish call comes with a costly condition: it expects AI compute demand to put deep pressure on Meta’s free cash flow in 2027 and 2028.

Meta introduced Muse on September 8 as a personal AI agent powered by its Muse Spark models. It is rolling out in the United States on iOS, Android and muse.ai, with free access and subscription plans. Meta says Muse can take actions such as sending email and booking travel through a dedicated secure virtual machine.

JPMorgan analyst Doug Anmuth said Muse reached as high as No. 3 in the U.S. App Store on its second day, with early usage running at 10 times Meta’s training cohorts. He argued Meta is still at an early stage of releasing frontier models and AI-driven products beyond advertising, including Muse and Meta Model API access.

Anmuth’s forecast reflects a direct tradeoff in the bank’s thesis: successful AI products would also require more computing capacity to train and run them. The projected cash-flow deficit excludes potential revenue from AI products, making it a forecast of the investment burden before Muse or other AI offerings demonstrate direct returns.

JPMorgan is not treating Muse subscriptions as Meta’s only route to returns. Anmuth also pointed to potential AI-driven gains in Meta’s core advertising business through content recommendations and engagement, ad targeting and retrieval, and AI content creation.

Meta’s recent model update also made Muse Spark 1.3 with maximum reasoning available through Muse Code and the Meta Model API; Meta says it is designed for longer-running work across multiple workflows in one thread. For JPMorgan’s call, the important next evidence is whether early Muse use becomes durable demand and whether Meta’s AI improvements begin to strengthen its existing business as spending rises.

Sources

  1. about.fb.comIntroducing Muse: The World’s First Personal AI Agent Built for Everyone
  2. research.meta.aiIntroducing Muse Spark 1.3
  3. cnbc.comJPMorgan upgrades Meta after company’s recent AI product announcements

Loading discussion...

Related coverage

Continue reading