CEO Mark Zuckerberg was not a direct source of income for the company, a family of Meta’s open -source artificial intelligence model. But finally Kadrey v. The new documents submitted to the court in the copyright case called Meta made clear that Meta has shared income with some business partners. These documents presented by plaintiffs and previously released censored sections reveal that the revenue obtained from the users of the platforms that contain meta’s Llama models.
Although there is no clear statement in the documents of which Meta has received payment from host companies, some of the platforms that the company has cooperated with in blog posts published by the company are AWS, Google Cloud, Azure, Dell, Databricks, Nvidia, Groq and Snowflake. While these companies offer their Llama models to developers through their infrastructure, it also facilitates the use of the model by providing additional services and tools.
Mark Zuckerberg mentioned the possibility that Llama models could be licensed at a earning meeting in April last year. At that time, Meta said that it was the expectation of income from major technology companies that use Llama models for commercial purposes. In particular, companies such as Microsoft, Amazon and Google said that Meta would demand income sharing if they sell these services to third parties. The documents submitted in the case in question confirm that such agreements have already begun.
Although it can be downloaded, privatized and operated independently by developers who want Llama models, the use of the model through large cloud service providers continues to be a widespread choice because it provides practicality. Meta earns income, albeit indirectly through this use.
On the other hand, Zuckerberg argues that the main benefit of the Meta’s project is the development of models thanks to feedback and contributions from the open source community. Meta also uses Llama models in its own products. Many products, especially the company’s AI assistant assistant Meta AI, are used in these models.
Another important issue that comes up with income sharing agreements is the accusations of copyright violations directed to Meta. Kadrey v. In the commodity case, it is claimed that the company has trained hundreds of Tarabayt Pirate E-Books. The plaintiff argues that Meta has downloaded these contents with hidden torrent methods, and in this process, due to the nature of the torrent infrastructure, it also offers these pirate books to other users. This means that the company not only consumes pirate content, but distributes this content.
Meta continues its artificial intelligence investments in 2025
Meta plans to significantly invest in its work in the field of artificial intelligence in 2025. According to a statement in January, the company’s capital expenditures for 2025 will be between 60 and 80 billion dollars. This amount represents an increase of approximately twice compared to 2024. It is stated that most of the expenditures will be divided into the enlargement of data centers and engineering teams working in the field of artificial intelligence.
In addition, the news that the meta AI assistant has been worked on a paid subscription system that would offer additional features to the AI assistant. It is stated that this subscription service is considered as part of the strategy of generating more income than artificial intelligence -based solutions. All these developments show that Meta’s open -source artificial intelligence strategy restructuring both in terms of technological development and financial sustainability.