Meta has launched Muse Spark, its first major new AI model in over a year, marking a significant shift from open-source AI to proprietary models and signaling a new phase in its artificial intelligence strategy [1]. The release comes nearly 10 months after Meta spent billions of dollars to recruit Scale AI co-founder Alexandr Wang as the centerpiece of Mark Zuckerberg's AI overhaul, and follows the creation of Meta Superintelligence Labs, a new elite unit formed after hiring Wang and several top engineers and researchers [1]. In June, Meta spent more than $14 billion on these hires, and in January, the company announced plans to invest between $115 billion and $135 billion in capital expenditures for 2026, nearly double its 2025 capex figure [1].
Morningstar analyst Malik Ahmed Khan commented that Meta needed to demonstrate to investors and operators that it has been working on substantial AI developments, calling this release 'the first step.' He emphasized that the next challenge is making the model work and figuring out how to monetize it [1]. Despite heavy spending, Meta has yet to show new revenue streams from its AI investments, unlike rivals OpenAI, Anthropic, and Google, which have led the AI boom with powerful models and popular chatbots, generating significant value and embedding their technologies across products and services [1].
Muse Spark is proprietary, marking a departure from Meta's previous open-source Llama models. The company has stated intentions to eventually release some open-source versions, but the current focus is on proprietary offerings. Arun Chandrasekaran, an analyst at Gartner, described this as a 'major shift' and noted it signals Meta's intention to move away from the Llama brand, especially after Llama 4 failed to captivate developers following its April release [1].
Meta plans to offer third parties paid API access to Muse Spark after an initial 'private API preview' with select parties, following the model of other frontier AI labs. However, Meta is entering the paid AI API market late, as OpenAI and Anthropic are collectively valued at well over $1 trillion, and Google has already embedded its Gemini models across its portfolio and cloud offerings [1].
CONCLUSION
Meta's launch of Muse Spark represents a pivotal shift toward proprietary AI models and a bid to monetize its massive investments. While the company has demonstrated progress, analysts highlight the challenge of turning these efforts into revenue, especially as competitors have already established lucrative AI businesses. The market is watching closely to see if Meta can successfully commercialize Muse Spark and justify its substantial capital expenditures.