Meta Platforms Inc.
What Happened: An unredacted court filing submitted by attorneys for the plaintiffs in the copyright lawsuit Kadrey v. Meta reveals that the company earns a share of the revenue from businesses that host its Llama AI models. The lawsuit indicates that the company reportedly trained its Llama models using pirated e-books.
The filing does not specify which hosts pay Meta, but the company has previously listed partners including Amazon's AWS
Meta did not immediately respond to Benzinga's request for comment.
Meta CEO Mark Zuckerberg suggested the possibility of monetizing Llama models during an earnings call in April, hinting that companies reselling these services should share a portion of their revenue with Meta. However, Zuckerberg later backtracked and in a July blog post, the Meta CEO stated, "selling access" to Meta's openly available Llama AI models "isn't [Meta's] business model."
The recent revelation of these revenue-sharing agreements contradicts the company's latest stance and suggests that Meta has begun to act on its previous strategy.
Why It Matters: This revelation comes just days after Zuckerberg announced that Meta's Llama model had reached a milestone of 1 billion downloads. Zuckerberg celebrated this achievement on Threads, stating that each download represented a "vote of confidence in open source AI". The new court filing, however, raises questions about Meta's revenue strategy and the company's transparency with its partners and users.
Zuckerberg stated that Meta benefits from Llama mainly through AI research-driven improvements. Llama models power various Meta products, including its AI assistant, Meta AI. Despite ongoing legal battles, Meta continues to advance its AI investments. The Zuckerberg-led company also intends to launch a trial paid subscription service for Meta AI, offering users access to advanced versions of its technology.
Meta holds a momentum rating of 78.90% and a growth rating of 77.28%, according to Benzinga's Proprietary Edge Rankings. The Benzinga Growth metric evaluates a stock's historical earnings and revenue expansion across multiple timeframes, prioritizing both long-term trends and recent performance.