Meta’s new shopping agent, Muse, may never carry ads. That’s the case laid out in a new MBI Deep Dives analysis, and it has less to do with restraint than with a paid media loop most marketers haven’t fully mapped.
The no-ads argument
The October 1 newsletter piece, “Why Muse May Never Need Ads”, argues that keeping ads out of Muse is itself a positioning move. If users hand over logins, payments and errands, the product has to feel like it works for them, not the highest bidder.
The post comes from Abdullah Al Rezwan, who writes under the pseudonym Mostly Borrowed Ideas. It is an opinion, not a Meta announcement, and the author discloses a personal portfolio.
Meta’s public commitments leave that door open. Muse conversations and virtual-machine data stay out of Meta’s ad systems. But Meta AI chats have fed ad personalisation since December 16, 2025.
The retargeting loop
Here’s the piece’s most distinctive claim: when Muse shops on a merchant’s site, Meta says the visit appears as the person’s own activity. That means a merchant’s site tag can pick up the visit and serve retargeting ads later on Facebook and Instagram.
The loop needs two assets: a tag on the merchant site and a large owned ad surface. Meta has both. So Meta can keep Muse clean and still make money on the ad platforms it already owns. MBI calls this a shot at “aggregating the aggregators”.
Why Meta can afford to wait
Meta reported $59.36 billion in Q2 2026 ad revenue. That existing machine makes delaying Muse monetisation far cheaper for Meta than for almost any rival. Digiday’s competing reading expects ads in Muse, pointing to heavy AI capex and limited early revenue. Most investors MBI speaks to still see Muse ads as almost inevitable. The author is not so sure.
What marketers should watch
- No new inventory yet. If this thesis holds, Muse adds no separate placement to buy; spend stays on existing Meta surfaces.
- Labelling gets messy. Agent traffic can look like a buyer or a bot, and Meta treats it as the user’s own activity. Measurement teams need a clear flag.
- Commerce players are diverging. DoorDash is testing iMessage ordering with 20,000 US iOS users, while Airbnb plans an agent next year.
- Amazon is the holdout. It blocked Muse, but Amazon Ads also announced a ChatGPT partnership on September 10, 2026.
A practical move
Don’t wait for a Muse placement. Check whether your site tag captures agent-driven checkout visits, and decide how those users enter your retargeting pools. If API-based checkout bypasses the tag, your measurement plan needs a fallback.
The contrast is real: ChatGPT says advertising passed a $1 billion annualised run rate on August 31, 2026, while Comscore put ChatGPT at 50% of US AI prompt volume in June, down from 70% in January. Gemini and Claude now hold 30% and 11% respectively. Whether an ad-free agent and an ad-funded one can both hold large audiences is the next set of numbers worth tracking.
Source: PPC Land
