Meta CEO Mark Zuckerberg during the MetaConnect event on September 17, 2025 in Menlo Park, California.
David Paul Morris | Bloomberg | Getty Images
Mark Zuckerberg is counting on you meta Its new artificial intelligence model, Muse Spark, is aimed at restoring the company’s position in the fast-growing AI market, and its guidance and commentary will become even more important following Wednesday’s first-quarter results.
That’s because the new model, previously codenamed “Avocado,” was announced in early April, just as the second quarter began. Muse Spark marks a turning point in Meta’s AI strategy, moving away from the previous Llama model, which was released for free to the open source community.
Meta will eventually work with OpenAI, Anthropic, google. Analysts say what’s important today is that Meta’s AI tools continue to strengthen the company’s dominant advertising business and that the company is showing its AI technology can compete with market leaders.
As of Sunday, Meta AI is behind Anthropic’s Claude and Google’s Gemini in text, but only Claude is trailing in vision, according to Arena.AI, a site that tracks the quality and performance of top models. In both areas, it currently outperforms OpenAI’s GPT. Claude also tops the Documentation and Code category, while Meta is further down the leaderboard.
Citizens analysts said in a report to clients last week that AI is a “complementary good” for Meta, and said they expected to hear more from the company on its earnings call.
Analysts recommending the stock to buy said in a report that they are “impressed with Meta’s Muse Spark model,” citing the model’s strengths in text and vision. “While the company has integrated Meta AI into its core app, we are waiting for a strategy to drive consumer adoption at scale, similar to other AI chatbots such as ChatGPT and Claude. We believe this will unlock new data and advertising budgets.”
Meta’s advertising business continues to grow due to improved targeting capabilities as AI advances. Analysts expect first-quarter sales to rise 31% from a year ago to $55.6 billion, according to LSEG. This will be the fastest pace of expansion since 2021.
But thanks to the popularity of OpenAI and Anthropic’s AI models and services, whose combined valuations have ballooned to more than $1 trillion, Wall Street is looking for momentum in the AI space beyond just advertising. Meta’s stock has risen 24% over the past year, while Alphabet’s stock has gained 116% in that time, driven by Gemini’s growth.
When Meta announced Muse Spark earlier this month, it was touted as the first major AI model to come out of Meta Superintelligence Labs, led by the company’s chief AI officer, Alexandr Wang. Wang previously served as CEO of Scale AI and joined Meta in June as part of the company’s $14.3 billion investment in the data label startup.
Mr. Zuckerberg followed suit with even more high-profile hires. He brought on former GitHub CEO Nat Friedman and business partner Daniel Gross. Daniel Gross was the CEO of Safe Superintelligence, an AI startup co-founded by Ilya Sutskever in 2024 after leaving OpenAI.
“This leadership change and subsequent nine-month restructuring of Meta’s AI stack represents an aggressive effort to close the gap with competitors such as OpenAI (private) and Google,” Truist analysts wrote in an April 21 report.
“Back to the AI conversation”
Meta said that internal tests published to coincide with Muse Spark’s debut showed the model was less powerful than state-of-the-art AI models such as Anthropic, a way for the company to address early expectations.
Still, analysts are relieved that Meta has finally made its way out, and more models are likely on the way. JPMorgan Chase analysts wrote in a report last week that Muse Spark has “re-introduced the meta into the AI conversation.”
“Investor sentiment towards Meta is becoming increasingly constructive,” the analysts wrote. “Stock prices are under pressure due to rising costs and capital expenditures, concerns about delays in AI models, and adverse legal rulings on social media.”
Meanwhile, Meta is focusing on AI and reducing its headcount.
The company announced Thursday that it will lay off about 8,000 people, or 10% of its workforce, on May 20 in an effort to improve operational efficiency. It comes as Meta is pouring money into AI infrastructure, telling investors in January that AI-related capital spending in 2026 would be in the range of $115 billion to $135 billion, up from $72.2 billion in 2025.
Analysts at Loop Capital wrote in a recent report that Meta’s large investment is reinforcing the negative perception that the company is “desperately spending money to fix flawed AI initiatives.” They said the release of Muse Spark shows that Meta is creating an AI model that can further improve its core online advertising business.
Even if future models of Muse Spark and Meta fail to outperform competing systems, their tests are of “mixed importance” because of the company’s clear advantage in advertising, Loop analysts wrote.
“While the underlying LLM/agent inference models are certainly key to meta, we see the image/video generation models as strategically important with greater impact on short-term engagement and monetization,” they write. “The real hurdle to success is building a model that powers a great product for users, creators, and advertisers.”
Spotlight: Meta and Broadcom strengthen AI chip partnership through 2029.
