Meta's long-awaited AI model is finally here. But can it make money?

Meta Platforms Inc., led by CEO Mark Zuckerberg, unveiled its new proprietary AI model named Muse Spark on Wednesday, September 17, 2025, at its Meta Connect event in Menlo Park, California. This release marks the company's first major AI model in nearly a year, following a massive investment and strategic overhaul, and is a direct attempt to compete in the lucrative AI market currently dominated by rivals like OpenAI, Anthropic, and Google. The central challenge now facing Meta is transforming this significant technological investment into a profitable business stream, a feat it has yet to achieve despite being one of the industry's biggest spenders on artificial intelligence. The launch of Muse Spark represents a pivotal strategic shift for Meta. After the underwhelming reception of its previous open-source Llama models, particularly Llama 4, Zuckerberg initiated a major corporate restructuring. This included a landmark investment of over $14 billion in June to hire Scale AI's Alexandr Wang and his top engineers, forming the elite Meta Superintelligence Labs. The company has signaled an even more aggressive financial commitment, planning capital expenditures between $115 billion and $135 billion for the year, nearly double its 2025 figure. Analysts view Muse Spark as a necessary demonstration to investors that Meta's vast spending has yielded a substantive product, moving the company away from its open-source roots toward a proprietary, monetizable offering. Meta's path to monetization, however, is fraught with challenges in a crowded and mature field. The company plans to eventually offer paid API access to Muse Spark, following an initial private preview. Yet, it enters a market where competitors are already entrenched with popular services and multi-trillion-dollar valuations. Analysts argue Meta's clearest advantage lies not in chasing the general developer community but in leveraging its unparalleled user base of over 3 billion people across Facebook, Instagram, and WhatsApp. The primary business opportunity is seen in enhancing its core advertising engine, which generated 98% of its $200 billion revenue last year. By using AI to create more engaging ads and improve targeting, Meta could directly boost advertiser ROI, creating a compelling reason for them to pay for new AI services. Early technical analysis suggests Muse Spark excels in image and video processing, capabilities highly valuable for the video-centric advertising on platforms like Reels. Ultimately, the success of Muse Spark is a critical test of Meta's AI ambitions. While Zuckerberg has long harbored visions extending beyond advertising, the immediate imperative is commercial. The model must not only compete on technical benchmarks with established leaders but also carve out a unique value proposition. For a company of Meta's scale, developing its own frontier model is also a matter of strategic sovereignty, ensuring it remains a key player and is perceived as a leading AI company. As one analyst put it, the team has delivered a state-of-the-art model; the pressure is now back on Zuckerberg and his leadership to answer the pivotal question: what will they do with it?