Cutting-edge models and AI agents open up room for upside! JPMorgan turns bullish on Meta (META.US), setting target price at $820

According to Wisdom Financial APP, Meta (META.US) shares have fallen 1% so far this year, underperforming the S&P 500 Index's 12% gain. However, Meta’s share price has already rebounded 20% from recent lows, while in the same period, the S&P 500 Index dropped 1%. JPMorgan pointed out in a recent research report that Meta stock still has considerable upside potential, as the company is currently at an early stage of launching cutting-edge models and AI-driven products outside its advertising business, particularly the Muse AI agent and Meta Model API access services.JPMorgan raised its rating on Meta shares from "Neutral" to "Overweight" and raised its December 2027 price target from $640 to $820, implying an upside of about 30% by the end of next year,

In the report, JPMorgan stated that advanced models will become the core of Meta’s product and commercialization pipeline over the next several years and will propel the company toward superintelligence. In the summer of 2025, when Meta begins rebuilding Meta Superintelligence Labs (MSL), its goal is to deliver frontier-level models within a year. The bank believes that the MSL team has largely achieved this goal, and the model release pace is quickening—from the launch of Muse Spark 1.1 this July to the recent release of Muse Spark 1.3, the latter of which is already competitive with the Claude and GPT series models.

JPMorgan expects Meta’s next-generation model (codenamed Watermelon) to unlock more opportunities for Meta in consumer products, business intelligence, FoA (Family of Apps) user engagement and advertising, as well as internal operations and efficiency. The bank has long believed that Meta is well positioned to offer consumer-driven AI products to its approximately 4 billion user base, and at-scale distribution capability is a significant competitive advantage.

Meta’s Muse AI agent has already shown robust early growth momentum, briefly reaching No. 3 in the U.S. App Store on its second day online, and early usage levels reached ten times that of the training cohort. While commercialization is not Muse’s near-term priority, JPMorgan believes there are meaningful opportunities in both the commission/take-rate and subscription models, with the potential total addressable market (TAM) possibly reaching tens of trillions of dollars.

Importantly, Muse is built around safety and privacy protection, which is an advantage for Meta. JPMorgan believes this could become a key barrier for other agents to earn user trust and perform more complex tasks involving passwords, credit cards, and other personal information. Over time, the bank expects interactions among agents to become more common, and merchants and enterprises to gain greater intelligence and higher transaction volumes. As Muse acquires more users, it will become a significant use case for Meta’s computing resources, but the bank also sees substantial opportunities for Meta Model API services aimed at developers and businesses.

As Meta’s modeling capabilities advance from Muse Spark 1.3 to Watermelon and subsequent models, supported by highly competitive pricing, the appeal of Model API access services should continue to grow, further driving usage. Although Meta is now highly focused on cutting-edge model construction and AI product development, JPMorgan believes its AI infrastructure capacity is also extremely valuable and will support future computing demands.

Meta has stated that the company will maximize computing capacity utilization in 2026 and 2027 and will have more flexibility from 2028 onwards. JPMorgan expects Meta’s capital expenditures in 2027 to reach $243 billion, up 70% year-on-year, and $284 billion in 2028, up 17% year-on-year, both well above market consensus. The bank believes that factors that could affect 2028 capital expenditures include the pace and hurdles in building data centers, adoption and commercialization rates of AI products, and supply constraints.

Currently, JPMorgan expects Meta to face greater free cash flow (FCF) pressure in 2027 and 2028, with negative free cash flow each year in the range of $65-70 billion. However, at the same time, the bank has not included new AI product commercialization revenue in its forecasts. More importantly, the bank believes there remains huge growth potential for Meta’s core advertising business, mainly driven by improvements in AI-powered content recommendations and user engagement, more accurate ad targeting and retrieval, as well as improvements in AI content creation.

Overall, JPMorgan believes Meta’s massive distribution capability and user engagement continue to drive network effects, while its precise ad targeting ability provides advertisers with enormous value. The company is seeing significant improvements through AI-driven ad sorting and recommendation capabilities, thereby boosting user engagement, increasing return on ad spend (ROAS), and driving revenue growth. JPMorgan notes the company has already demonstrated a credible path to commercializing its AI-driven capital expenditures via consumer-focused and enterprise-facing agents as well as Model API access services.