Meta accelerates its repositioning in AI with a major shift

After months of false starts, Meta returns with an AI model capable of standing up to its rivals. Muse Spark is not just a technical advancement, it is a sign of a profound strategic shift. This launch says a lot about Meta’s ambitions. And perhaps even more about the end of an era.

Muse Spark, the new face of AI at Meta

Meta officially announced Wednesday the launch of Muse Spark, its new AI model, now available via its dedicated website and app. This launch marks the strong comeback of the Menlo Park company after more than a year without major AI news.

At the helm of this project: Alexandr Wang, 28, founder of Scale AI and new AI director at Meta. Wang joined the company after Meta acquired 49% of Scale AI for about 14 billion dollars. He now oversees the Superintelligence Labs, the division responsible for pushing the group’s AI ambitions to the maximum.

The markets welcomed the announcement: META stock jumped up to 9%, wiping out losses accumulated since the end of March. A strong signal, showing that investors still believe in Zuckerberg’s ability to turn things around after recent setbacks.

Regarding performance, Meta states that Muse Spark performs the same tasks as its predecessor Llama 4 Maverick, but with significantly reduced computing consumption. 

Internal benchmarks show the model matches or even surpasses on some criteria the solutions from Google, OpenAI, and Anthropic. Independent tests confirm solid results in the medical field, but still point out shortcomings in programming.

A strategic turning point that doesn’t say its name

This launch is much more than a technical update. Muse Spark breaks with one of Meta’s most entrenched traditions: open source. Unlike the Llama series, this new model is closed and proprietary. 

Some partners will be able to access it via API in preview, but Meta has not yet committed to public release. The company mentions potential future open source versions, but without a specific timeline.

This reversal illustrates the increasing pressure Meta is under in the AI race. The Llama 4 model disappointed at its launch in 2025. The Behemoth project, a 2,000 billion parameter behemoth, was abandoned before even seeing the light of day. Faced with these setbacks, management decided: speed is essential, and protect what can still be protected.

To finance this race, Meta is sparing no expense: 135 billion dollars of AI investments are planned for 2026, almost double the previous year. A colossal sum, which does not come without social consequences. 

The Reality Labs division, the one for the Metaverse, which has swallowed 80 billion dollars, has undergone hundreds of layoffs. And according to several sources, new layoff plans are under consideration, possibly exceeding 20% of employees. Meta denies this, but doubt is growing.

This context also sheds light on a background piece of information: Zuckerberg is developing a personal AI agent himself to steer Meta with fewer intermediaries. AI as an internal management tool, as a hierarchical shortcut, perhaps this is the true face of this revolution.

Muse Spark is just the beginning, warn the leaders of Meta. More advanced models are announced for this year. The AI race is back on.