Court documents revealed by attorneys involved in a Meta lawsuit defined Meta’s ambitions to make multi-billion dollars in generative AI.
Meta is eyeing a massive revenue in the generative AI industry, with internal estimates pointing to an expected revenue of $2 to $3 billion by 2025 and a staggering revenue of $460 billion to $1.4 trillion by 2035.
During the proceedings of Kadrey v. Meta, the court documents were revealed on Wednesday by Attorneys representing the book authors who sued Meta for copyright infringements. These documents stated that Meta is monetizing its AI products through multiple channels.
However, the document didn’t explicitly state what these ‘generative AI products’ were.
Some clues are there, though.
For instance, revenue-sharing deals tied to its open-source Llama models, a newly launched API for customizing Llama, and the anticipated monetization of its Meta AI assistant, which could soon include in-assistant ads and paid premium features, as hinted by CEO Mark Zuckerberg during the company's Q1 2025 earnings call.
Meta has earlier given hints of its ambition to spend $60 to $80 billion in 2025, focusing on the expansion of new data centers.
Here’s the twisted part: the documents also reveal that Meta allegedly had planned a $200 million budget in 2023 to acquire the training data for its Llama models. Expectedly, $100 million of this budget was supposed to be for book authors whose books would be used for training its AI.
Shockingly, the budget was allegedly diverted to some other project.
As the official court document states:
“Here, there is no dispute that Meta torrented over 267 TB of pirated copyrighted data, i.e., Data from known illegal websites, nor is there any dispute that Meta acquired at least hundreds of thousands of additional pirated works via direct download from such sites.”
It’s quite clear that the storm of lawsuits for Meta isn’t going to slow down easily. How harmful it will be for its revenue goals, we will have to wait and see.
