Why Meta’s Moment in the AI Sun Won’t Last
If Meta Platforms wants to raise money by selling equity, now is the time. Meta shares have gone on an incredible run lately, jumping 27% since Sept. 8, when the company released its Muse personal AI agent. Meta stock is now up 18% so far this year, a bigger gain than that for Alphabet, Amazon or Microsoft—quite the turnaround for a company that not too long ago was in Wall Street’s doghouse for its seemingly out-of-control AI spending.
It’s no secret why investors have fallen back in love with Meta—the company has done a good job of marketing Muse, prompting lots of glowing reviews and headlines about app downloads. Wednesday’s Connect conference, where Meta CEO Mark Zuckerberg unveiled new virtual reality and AI hardware—including a device called Charm designed around Muse—appears to have reinforced the positive sentiment. But let’s not overinterpret the current state of play. In AI, no one stays on top for very long.
Remember how OpenAI’s launch of ChatGPT put Google on the back foot, sparking lots of chest-thumping about the bleak future for the tech giant—until Google’s Gemini eventually caught up? Starting around November of last year, Gemini was the hot new model. But Google’s moment in the AI sun didn’t last. With Gemini appearing to lag behind, some now say Google isn’t even in the frontier AI race!
Meta has ridden the same roller coaster. Back in late 2023, when Meta released its Llama 2 model, it got such a positive reception that some saw it as taking market share from OpenAI’s GPT models. Reaction to Llama 3 in early 2024 was also upbeat. But Meta’s reputation fell into the toilet in 2025 after the company ran into problems with Llama 4.
The current burst of enthusiasm for Meta is a bit nuts. A headline on Bloomberg Opinion on Thursday suggested that the new Charm device “could be Meta’s iPod moment.” Huh? The iPod was a revolutionary device for listening to music, filling a need everyone had. The Charm offers another way for people to interact with an agent, in addition to their phone and Meta’s AI glasses. Who needs a separate device just to talk to Muse? Chances are Charm will quickly be nicknamed Charmless.
Whether Muse’s popularity is enduring is also a big question. We reported earlier this week that more than 500,000 people had tried out Muse a week after launch, which is an encouraging number. But how many of those will still be using it in a month or six months’ time? We have good reason to be skeptical about Muse’s prospects over the long term, given the widespread antipathy consumers are expressing toward AI.
And even assuming Muse remains popular, will it make or lose money for Meta? Don’t forget the agent is free to use within certain limits. A report by Rosenblatt Securities on Thursday suggested that for Muse to be profitable, either computing costs would need “to drop substantially” or an average user would have to spend more than $1,000 a month buying stuff through Muse, “well above current e-commerce averages or spending on Amazon.” (Meta plans to take a cut of transactions done through Muse).
All that means Meta’s current stock rally won’t last. If Zuckerberg is smart, he’ll get Meta to sell a few billion dollars’ worth of stock, raising much-needed cash for its costly AI expansion.