JPMorgan Upgrades Meta to Overweight, $820 PT

JPMorgan analyst Doug Anmuth upgraded Meta Platforms Inc.

(NASDAQ: META) from Neutral to Overweight with a price target of $820 (from $640).

“Meta shares are up +20% from their recent lows (vs. SPX -1%), but still down -1% YTD vs. the SPX up +12%.

We believe there’s still meaningful upside potential as Meta is in the early stages of releasing frontier models and AI-driven products beyond advertising, notably Muse AI agent and Meta Model API access.

Frontier models are at the core of Meta’s product and monetization pipeline over a multi-year period, and its path toward superintelligence.

Meta’s goal in the summer of 2025 as it began to rebuild MSL was to deliver models at the frontier within a year.

We believe the MSL team has essentially delivered with an accelerated pace of model rollouts from Muse Spark 1.1 this past July to more recently Muse Spark 1.3, which is competitive with Claude and GPT models (see Figure 1).

We expect Watermelon to unlock additional opportunities for Meta around consumer products, business intelligence, FoA engagement and advertising, and internal operations and efficiency.

We have long believed that Meta is well-positioned to deliver consumer-driven AI products to its base of ~4B users, and that scaled distribution is a significant competitive advantage.Meta’s Muse AI agent has strong early traction, reaching as high as #3 in the US app store on its second day (see Figure 2) and early usage running at 10x that of training cohorts.

While monetization is not the near-term priority for Muse, we believe there are meaningful opportunities across a take-rate/commission model and subscriptions, within a TAM potentially in the tens of trillions of dollars.

Importantly, Muse is built around security and privacy—an advantage for Meta— which we believe could be a key hurdle for other agents to gain trust and navigate more challenging tasks involving passwords, credit cards, and other personal information.

Over time, we expect more agent-to-agent activity, with merchants and enterprises gaining increased intelligence and transaction volume.

Muse should become a significant use of Meta’s compute as it gains traction, but we also believe there is meaningful opportunity in Meta Model API for developers and enterprises.

Model API access should become increasingly compelling as Meta’s models strengthen from Muse Spark 1.3 to Watermelon and beyond, supported by highly competitive pricing which should drive increased volume.

While Meta has been heavily focused on frontier model buildout and AI product development, we believe there is also strong value in its AI infrastructure capacity which will support future compute requirements.

Meta has indicated it will maximize compute capacity in 2026 and 2027, with more flexibility in 2028 and beyond.

We project 2027 capex of $243B (+70%) and 2028 capex of $284B (+17%), both well above consensus.

We believe potential factors influencing 2028 spend could be data center build status and pushback, AI product adoption and monetization, and supply constraints.

We currently anticipate greater FCF pressure in 2027 and 2028—in the range of negative $65B-$70B each year—but we’re also not yet building in new AI product monetization.

And importantly, we believe there’s still meaningful headroom in core advertising from AI-driven improvements tied to content recommendations and engagement, better ad targeting and retrieval, and AI content creation.”

This research note is reproduced verbatim from the issuing firm. Price Target never edits, paraphrases or alters analysts’ words — we only republish them in one place.