JPMorgan Flips Meta to Overweight, Lifts Target to $820 on Muse/Frontier-Model Case

Takeaway: JPMorgan's Doug Anmuth upgraded Meta to Overweight from Neutral this morning and raised his price target 28%, to $820 from $640 — the most concrete institutional validation yet of the AI-agent thesis that's already driven the stock's outsized move this week. The tape reaction is muted because Meta was already the desk's top performer; the note matters more for what it says about durability than for what it does to price today.

What happened: JPMorgan published the upgrade around 6:35am ET today (Sept. 10), confirmed via CNBC ("JPMorgan upgrades Meta after company's recent AI product announcements"), Bloomberg ("Meta Upgraded at JPMorgan as Muse Highlights Better AI Position"), Seeking Alpha and Investing.com. Anmuth's case:

• Meta's superintelligence lab has "essentially delivered on its goal of reaching the frontier within a year," with Muse Spark 1.3 now "competitive with Claude and GPT models." An upcoming model, internally called Watermelon, is flagged as the next unlock across consumer engagement, advertising and internal efficiency.
• The Muse consumer agent hit as high as No. 3 in the U.S. App Store on its second day — evidence the demand signal cited in this week's rally headlines has legs, not just launch-day novelty.
• Anmuth isn't pricing near-term monetization from Muse itself, but sizes the eventual commission/subscription opportunity in "the tens of trillions of dollars" of addressable market.
• The $820 target is built on 23x his 2028 EPS estimate of $35.44 — which he calls potentially conservative — and holds even against capex estimates of $243B (2027) and $284B (2028), both above Street consensus. His view: core advertising still has "meaningful headroom" from better targeting and AI-generated creative, so the elevated spend doesn't break the earnings case.

Mag7 impact — bullish for META, mixed for the pair trade with GOOGL: Meta enters the note already extended: per this morning's tape, META is +10.3% above its 20-day EMA, RSI 70.4, up 6.55% on the day prior with 2.2x relative volume and +7pp versus SPY. That's the desk's own read on why this looks like "an already-paid move that typically digests" rather than fresh fuel for a breakout — the JPMorgan call is confirmatory, not a surprise catalyst, and options are pricing it that way (META ATM IV 35.7%, put/call vol 0.54, no unusual skew visible in the cache). The more interesting read-through is competitive: ARK Invest has been simultaneously trimming Alphabet while adding Meta, and GOOGL remains the tape's clear laggard (RSI 39.4, -3.2% vs EMA20, -1.98% gap), 8.5 points behind META/MSFT on trend despite being the largest single CapEx spender in the group at $185B. A bank-side note that explicitly frames Meta's AI stack as ahead of the pack — rather than merely "in the AI race" — reinforces that dispersion rather than resolving it. Spillover to NVDA/hyperscaler capex names is directionally positive (higher, not lower, 2027-28 capex estimates from a bank now bullish on the name) but second-order.

What to watch next: Whether other sell-side desks follow with target raises in the next 24-48 hours, which would confirm this is a re-rating rather than a one-off; Meta's response (if any) on Watermelon timing, since Anmuth's target explicitly leans on that model shipping; and whether META's RSI cools back under 70 without giving up the +10% trend edge over GOOGL — a orderly digestion would validate the "own the strength, don't chase it" positioning already on the desk, while a sharp reversal would say the move was more headline-driven than fundamentals-driven. Also worth tracking: capex commentary from META's own IR in response to the $243B/$284B estimates, since that number running hotter than consensus is exactly the kind of data point that has hit datacenter-adjacent names (VRT, KLAC, AMAT) this week.