Meta’s Muse Spark AI Shows Early Promise, but Investors Seek Clarity on Zuckerberg’s Strategy

Meta’s newly unveiled Artificial Intelligence model, Muse Spark, has shown early promise, yet investors are looking for clearer strategic direction from CEO Mark Zuckerberg when the company reports first-quarter earnings on Wednesday.

Unveiled in early April under the codename Avocado, Muse Spark marks a strategic departure from Meta’s previous open-source Llama models. The company now plans to eventually monetize the Technology by offering pAId access to developers—similar to the business models of OpenAI, anthropic, and Google.

Analysts emphasize that for now, Meta’s primary goal is for its AI tools to continue strengthening its dominant advertising business while dEMOnstrating that its AI can compete with market leaders. According to Arena.AI, which tracks model quality and performance, Meta AI currently trails Anthropic’s claude and google’s Gemini in text-based tasks but matches Claude in vision-based tasks. It outperforms openai’s GPT in both areas but lags behind Claude in document and code categories.

Citizens analysts described AI as a “complementary good” for Meta in a CLIent note, expressing anticipation for more details on the Earnings Call. They praised Muse Spark’s text and vision capabilities but noted that Meta has yet to outline a clear strategy to drive consumer usage comparable to ChatGPT or Claude—something that could unlock new data and ad budgets.

Meta’s ad business continues to benefit from AI-powered targeting. Analysts expect first-quarter revenue of $55.6billion,up31$1 trillion. Meta’s stock has risen 24% over the past year, compared to Alphabet’s 116% gain driven by Gemini.

Muse Spark is the first major model from Meta Superintelligence Labs, led by Chief AI Officer Alexandr Wang, formerly CEO of Scale AI. Meta has invested $14.3 billion in the data-labeling startup. Zuckerberg has also hired former GitHub CEO Nat Friedman and AI entrepreneur Daniel Gross, previously CEO of Safe superintelligence, co-founded by Ilya Sutskever.

Truist analysts noted that the leadership overhaul and nine-month rebuild of Meta’s AI stack reflect an aggressive push to close the gap with rivals. Unlike Llama, Muse Spark is closed-source, signaling a shift toward high-performance, specialized infrastructure.

Although Meta’s internal testing suggests Muse Spark is less powerful than cutting-edge models from Anthropic and others—an APParent effort to manage expectations—analysts are relieved that Meta has entered the race. JPMorgan wrote that Muse Spark “has brought Meta back into the AI conversation.”

Meanwhile, Meta is reducing headcount to focus on AI. The company announced a 10% workforce reduction (about 8,000 employees) effective May 20, as it ramps up AI infrastructure spending. In January, Meta projected 2026 AI-related capital expenditures between $115billionand$135 billion, up from $72.2 billion in 2025.

Loop Capital analysts noted that Meta’s heavy spending has fueled perceptions of “a company desperately spending to fix problematic AI initiatives.” However, they believe Muse Spark demonstrates that Meta is producing models capable of improving its core ad business. Even if future models underperform rivals, benchmarks are of “mixed importance” given Meta’s ad advantage.

“The real bar for success is building models that power excellent products for users, creators and advertisers,” they wrote, adding that image and Video Generation models hold greater near-term engagement and monetization potential.