Meta shares rise sharply with announcement of new Muse Spark model

Meta announced on April 8, local time at Estados Unidos, the launch of Muse Spark, its first large-scale language model developed by Meta Superintelligence Labs. The announcement led to the company’s shares rising by up to 9% on the same day, recovering losses recorded at the end of March. The model is now available on the Meta AI app and website, with integration planned for platforms such as Instagram, WhatsApp and Facebook in the coming weeks.
The company highlights that the Muse Spark performs similar tasks to the Llama 4 Maverick, but with less processing power consumption. Esse launch represents a milestone in Meta’s strategy to compete with rivals such as OpenAI, Google and Anthropic. The initiative is part of a larger effort to advance AI capabilities applied to the company’s products.
Market reactions to the announcement of the new model
Investors responded positively to the news. Shares of Meta recorded significant gains during the trading session, erasing a sequence of recent declines. Analistas noted that the movement reflects expectations surrounding the company’s heavy investments in artificial intelligence.
The rise occurred in a context of general optimism in the technology sector. The stock’s performance helped reinforce confidence in Meta’s long-term plan for AI.
Technical details of Muse Spark
Muse Spark is the first model produced under the leadership of Alexandr Wang, who took over as chief AI officer after Meta invested US$14.3 billion in Scale AI. The company claims that the model performs comparable or better in several writing and reasoning benchmarks when compared to previous generations.
However, reports indicate that the launch has been delayed from the initial schedule due to test results against competitors. Meta has included comparisons in presentation materials that position Muse Spark as competitive across multiple performance metrics.
- The model prioritizes computational efficiency for everyday tasks.
- It is designed for applications such as shopping and travel planning.
- Meta plans future updates to the Muse series.

Meta investments in artificial intelligence
The company plans to allocate between US$115 billion and US$135 billion in capital expenditures throughout 2026, an amount that represents almost double the amount forecast for the previous year. Most of these resources should be directed to AI infrastructure and superintelligence laboratories.
These investments include building data centers and hiring specialized talent. Meta has rebuilt its AI stack from the ground up over the past nine months to support the development of more advanced models.
The Muse Spark marks the beginning of a new phase of proprietary models, unlike the previous more open approach with the Llama family. The company indicates that open-source versions may be released in the future.
Alexandr Wang’s profile in Meta’s AI leadership
Alexandr Wang, billionaire in the technology sector, leads the development of the model at Meta Superintelligence Labs. Antes After joining the company, he founded and ran Scale AI, a company specializing in data labeling for AI training.
His hiring came as part of the $14.3 billion transaction that gave Meta a significant stake in Scale AI. Wang held the title of the world’s youngest self-made billionaire until October 2025.
Meta highlights that Muse Spark is a powerful foundation for advancements in areas such as science, mathematics, and health. The company is already working on developing the next generation of models.
Integration of the model into Meta products
Muse Spark currently powers the Meta AI assistant in the app and dedicated website. Nos In the coming days, the integration should extend to other company services, including AI augmented reality glasses.
Meta describes the model as aimed at helping users with practical day-to-day tasks. Essa approach seeks to make AI more accessible within the company’s ecosystem of social networks and tools.
The launch comes as Meta seeks to demonstrate a return on the billions invested in artificial intelligence capabilities. The company remains focused on building personalized assistants for millions of users around the world.
Comparisons with competing models
According to data presented by Meta, Muse Spark approaches the performance of cutting-edge models in reasoning and writing tests. Ainda Thus, it has limitations in areas such as coding when compared to offerings from direct rivals.
The company openly acknowledges these differences and claims that the model has been optimized for efficiency and integration with its products. Independentemente of benchmarks, the focus is on practical usefulness for end users.
Experts are closely monitoring Meta’s next steps in this technological race. Muse Spark is seen as the first in a series that could define the company’s position in the AI sector in the coming years.
Company’s long-term strategy
Meta continues to expand its infrastructure to support increasing demands for model training and inference. The resources allocated in 2026 should allow significant advances in available computing capacity.
This strategy reflects the company’s determination not to fall behind in the development of artificial intelligence technologies. The release of Muse Spark serves as an early indicator of the results of this concentrated effort.
The company plans to continue investing in research and development to improve model performance across multiple dimensions. Usuários can now test Muse Spark directly on Meta platforms.






