Key Points
Muse is the top app on the Apple App Store with an estimated 3 million weekly users.
Meanwhile, Meta will spend about $145 billion in AI data center costs this year.
Paid subscriptions won't come close to offsetting the company's AI costs.
By all early accounts, Meta Platforms' (NASDAQ: META) new Muse AI agent is a huge success. The app has an estimated 3 million weekly users and has been at the top of Apple's (NASDAQ: AAPL) App Store for several weeks.
Meta shareholders have driven the company's stock up more than 20% since Muse launched, optimistic that the company has figured out how to monetize AI.
Missed AI’s "Act 1"? Act 2 Could Be 15x Bigger. Most investors think they missed the AI boat because they didn't buy Nvidia in 2005. But according to our analysts, we’re only at the end of "Act 1"—the R&D phase. "Act 2" is the global rollout. Continue »
Call me a skeptic, but I'm not sold on the idea that Muse will be the catalyst some investors hope for. For one, Meta will burn through $145 billion in AI data center spending this year. Even selling a heck of a lot of Muse subscriptions won't come close to recouping those costs.
Secondly, virtually every major tech company is moving in the same agentic AI direction. The market is already getting crowded, and it's just getting started.
Image source: The Motley Fool.
Cheap AI doesn't exist
Muse is free to use with a limited number of AI tokens, and there are two paid tiers: $20/month (5 million weekly tokens) and $100/month (3 billion weekly tokens). Those prices align with what most AI companies charge for their monthly AI services right now, but the problem for Meta (and likely its rivals, too) is that these subscriptions won't come anywhere close to offsetting AI expenses.
As mentioned, Meta is spending $145 billion to build AI data centers this year, and some analysts think it could spend even more next year. Recouping those costs through subscriptions will be nearly impossible. That's especially when you consider that Muse is a consumer product, not an enterprise product like many other AI agents. Enterprise products can generate significant revenue because companies have larger budgets than consumers and are more willing to stick with the same service for longer.
Microsoft and Alphabet both recently grew cloud revenue rapidly, thanks mostly to selling cloud-based AI services to companies. Meta isn't a cloud company, and that's sort of the problem. Meta is investing similar amounts of money in AI data centers as its peers but doesn't have enterprise products to offset its high costs.
Meta recently launched Muse for Small Business, but it's more of a workflow automation tool (allowing Muse to connect to a Canva subscription) than the robust tools Anthropic and OpenAI offer. Making matters worse for Meta is that it's very likely that even its current subscription prices aren't high enough to match the amount of AI tokens users are spending.
For $100/month, Muse users (Musers?) receive 3 billion tokens per week. But the total amount of tokens a user gets often exceeds the value of what they pay for them. Earlier this year, a SemiAnalysis report found that a $200/month ChatGPT subscription actually gives a user $14,000 worth of AI tokens. For Claude, $200 essentially gets you $8,000 worth of tokens.
In short, AI companies are heavily discounting their products. A similar disparity is likely occurring with Muse's token costs versus its subscription price, which isn't ideal when it's spending hundreds of billions of dollars on AI.
The consumer-side competition will heat up, too
The other problem I think Meta will face with Muse is that, very soon, it probably won't be a unique product. While Meta beat many of its rivals to the punch in offering a consumer-first personal AI agent that can accomplish useful everyday tasks, it certainly won't be the last. Apple has already inched closer to this area with its newest version of Siri.
And while Siri can't yet book hotels or shop online for you, it likely could soon. If it does, Apple could eventually have the upper hand in this space because of its focus on consumer privacy.
Even with Muse's popularity, Meta consumers don't have much trust in the company's privacy practices. A recent Oppenheimer survey found that just 8% of respondents said they'd trust Meta with their passwords, compared with 30% for Alphabet and 23% for Apple (which took the No. 1 and No. 2 spots).
Even if Muse successfully fends off the competition for a while, it doesn't change the fact that Meta is burning through piles of cash. And with the company still generating the vast majority of its $200 billion in annual revenue from advertising, Muse's subscriptions aren't likely to make much of a dent in Meta's sales either.
Chris Neiger has positions in Apple. The Motley Fool has positions in and recommends Alphabet, Apple, Meta Platforms, and Microsoft. The Motley Fool has a disclosure policy.