JPMorgan Upgrades Meta to Overweight on Muse AI
JPMorgan lifted Meta to overweight with an $820 target. The Muse AI agent gives the platform its first revenue channel beyond ads. Here is the trade.
The upgrade
JPMorgan moved Meta to overweight from neutral on September 10 and lifted its price target to $820, roughly 30% above the prior close. Analyst Doug Anmuth built the call on Meta's AI stack, not its ad engine. A move of that size on the target is a loud signal even for a mega-cap.
Muse is the catalyst
The trigger is Muse, a standalone AI agent that runs tasks on a user's behalf on Meta's Muse Spark model family. The early traction is the real argument. Muse reached No. 3 in the US App Store on day two, with usage running at 10x comparable earlier cohorts. Anmuth reads that as proof Meta can monetize beyond advertising, through the agent itself and through paid Meta Model API access.
The other side of the bet
Conviction has a price, and it sits in the spending plan. JPMorgan models $243 billion of capex in 2027 and $284 billion in 2028, both above consensus. Meta is funding its AI ambition with an outlay that makes any miss expensive. Until model monetization catches up to the spend, margin is the risk that matters. That single line is where the skepticism of the past few quarters took root.
The trade
For traders, the upgrade rewrites the story. Meta had been the hyperscaler whose AI spend the market booked as a cost. Anmuth inverts it: Muse is the first visible revenue channel outside ads, and an $820 target forces the skeptics to answer. The stock jumped about 5% premarket.
Near term, the tape hangs on confirmation. If Muse engagement holds, the narrative carries; if it fades, the capex line snaps back into focus. The read runs wider than one name. A credible consumer AI channel at a platform of Meta's size underwrites the broader AI trade, from the chip suppliers to the data-center operators whose utilization leans on exactly these capex numbers. Own the stock from here and you are trading less the ad line than the question of whether Meta turns its models into revenue. Each weekly app ranking and engagement print from here is the tape that settles the bet.