Meta struggled to monetize AI products in near term

Ines Lin, Taipei; Jingyue Hsiao, DIGITIMES Asia 0


Meta regained market focus with the widespread popularity of its open Llama. Nevertheless, as its AI investments are still in the early stages, it will take years to see returns.

Despite the growth in sales, profit, and daily active users over the past quarter, Meta experienced a drop in stock prices after it released its latest financial results, which may be attributed to Mark Zuckerberg's remarks on the long-term nature of AI investments and other products such as VR devices and smart glasses, impacting market sentiment.

According to Meta's earnings call, AI and Reality Labs are key areas for the company's business development. Capital expenditures, particularly on servers and data centers, will likely rise in the coming quarters.

Expanding infrastructure investments can help train larger and improve underlying foundation models. The in-house MITA chip has been used to support core inference workloads and is now being deployed for tasks such as advertising and organic content recommendations starting in 2024, with hopes for broader support in AI training and certain generative AI applications.

Meta CFO Susan Li said the company's future spending plans will hinge on several key factors. These include advancements in foundation models, the uptake of generative AI products, and enhancements in hardware capabilities. Li emphasized that Meta's focus extends beyond immediate concerns such as Llama 4, encompassing long-term considerations such as Llama 7 and the necessary infrastructure.

Meta remains heavily dependent on digital advertising, constituting 98% of its sales. Meta is believed to be utilizing an open-source strategy to attract more users before integrating advertising into AI products to generate revenue.

However, data suggests that user engagement tends to drop as the volume of advertisements increases. If Meta embeds ads into its AI products, especially when the industry is still emerging, it could inadvertently deter users rather than attract them.

Meta remains bullish on AI's potential but has refrained from implementing overly aggressive monetization tactics. In addition to open-sourcing models, the company has introduced the Meta AI assistant, which integrates across platforms such as Facebook, Instagram, and WhatsApp. The strategy aims to familiarize users with AI technology and grow the developer and user ecosystem before exploring monetization methods.

Meta's business models are akin to those of Googe. However, unlike Meta, Google has search engines, productivity apps, cloud services, and Pixel devices, and the company can change its open-source strategy and begin to charge for its AI products.