Meta’s long-awaited AI model is finally here. But can it make money?

Almost 10 months after Meta spent billions of dollars to bring in Scale AI’s Alexandr Wang as the centerpiece of Mark Zuckerberg’s AI overhaul, the company finally revealed its first new model on Wednesday. One big question is — will users pay for it?

While rivals like OpenAI, Anthropic and Google have spearheaded the artificial intelligence boom with powerful models and popular chatbots as well as other services, Meta has been a hefty spender on AI but has yet to show any new revenue streams from it.

In June, Meta shelled out more than $14 billion to hire Wang and some of his top engineers and researchers, soon creating Meta Superintelligence Labs as a new elite unit. And in January, the company told Wall Street it plans to pour between $115 billion and $135 billion this year into capital expenditures, nearly double its 2025 capex figure.

“It’s been a year of basically no releases and a lot of hiring, and then the capex worries for this year are pronounced,” said Morningstar analyst Malik Ahmed Khan, in an interview. “I think Meta had to show investors and operators they have been working on something of substance. That’s the first step.”

Meta’s second step, Khan said, is making the model work and figuring out how to monetize it.

Muse Spark, Meta’s newly released model, is proprietary, a sharp change from its predecessor family of models called Llama, which consisted of open-source offerings, though the company said it does plan to eventually release some open-source versions. Zuckerberg shook up his company’s strategy after the April release of Llama 4, which failed to captivate developers.

Arun Chandrasekaran, an analyst at Gartner, described the move as a “major shift” and said it “signals an intention to move away” from the Llama brand.

Taking a cue from other frontier AI labs, Meta aims to eventually offer third parties paid API access to Muse Spark after an initial “private API preview” with “select parties.”

But Meta is very late to the game. OpenAI and Anthropic are collectively valued at well over $1 trillion, thanks to the popularity of their models and services, and Google has embedded Gemini across its portfolio of apps and products, while also selling access to the Gemini models via its cloud unit.

Meta’s AI technology, to succeed, has to be good enough to compete with top models while also providing a novel business opportunity.

Andrew Boone, an analyst at Citizens, said Meta’s clear advantage is the more than 3 billion people who use Facebook, Instagram and WhatsApp every month. And the business opportunity for Meta has nothing to do with trying to attract developers, who currently swarm to OpenAI, Anthropic, Gemini and a host of Chinese models, but rather to focus on its core market: advertising.

“That’s the crown jewel, that’s what needs to continue to improve,” said Boone, who recommends buying the stock.

Khan shares that sentiment.

“I believe that would be the killer use case from Meta’s perspective,” Khan said, with the goal being to “make ads more engaging and improve targeting.”

Advertising accounted for 98% of Meta’s $200 billion in revenue last year. The company has made numerous efforts to diversify its business, most notably spending tens of billions of dollars to try to make the metaverse happen. But Meta’s ad model is the one thing that’s consistently worked, and the company’s investments in AI have served to improve its targeting capabilities and provide better tools for marketers.

Khan said that as advertisers see returns on investment from their Meta spending, they reinvest that money back into more ads on the platform. So it makes sense that they’d be willing to pay for AI services if they can get even better results.

Meta declined to comment about its API plans beyond its initial announcement.

Based on the technical benchmarks Meta released comparing Muse Spark to rivals, the new AI model appears to excel in areas related to image and video processing, said Doris Xin, CEO of AI startup Disarray. Those are important characteristics for advertisers seeking to make dynamic campaigns for an audience that’s grown accustomed to viewing short-form videos on Reels or gawking at cat photos on Facebook and Instagram.

“Compared to like Claude and Gemini, I think it definitely feels like it has more of a consumer bent,” Xin said about Muse Spark.

Zuckerberg, however, has long had ambitions that go well beyond advertising. His approach with Llama was targeted at developers and getting the best and brightest minds in AI using Meta’s tools even if they weren’t paying for them.

With the switch to proprietary models, the pitch to developers becomes more difficult. Joseph Ott, CEO of AI startup Samu Legal Technologies, said he’s unsure about where he would find value.

“The only reason I would use Llama is that I could fine-tune it,” Ott said, referring to the practice of customizing AI models.

Many developers use so-called open-weight AI models, like those provided by Chinese tech companies, as a basis to train AI models to meet their specific use cases. Ott said it’s unclear what would make Meta’s Muse Spark stand out against free or cheaper alternatives and the leading proprietary AI models.

Ulrik Stig Hansen, co-founder of AI and data training startup Encord, said it’s important for Meta to develop its own AI foundation models to avoid any future dependencies on third parties. As one of the few companies with the resources and computing infrastructure necessary to create and maintain big AI models, Meta wants to ensure that it remains relevant in the hottest market on the planet.

“It is about AI sovereignty and being a player in the game,” Hansen said. “They want to be perceived and known as an AI company.”

As for Meta’s massive investment in Wang and his team, Boone said the latest benchmarks suggest that Zuckerberg got what he wanted, and now it’s “back on Mark.”

“We just gave you a state-of-the-art frontier model,” Boone said, referring to the team behind Muse Spark. “What are you going to do with it?”

WATCH: Meta unveils its new AI model: “Muse Spark.”