Meta expects its generative AI products to earn revenue between $460 billion and $1.4 trillion by 2035. This information came from court documents unsealed on Wednesday. The company also predicts $2 billion to $3 billion in revenue in 2025 alone.
These forecasts are part of a legal case. A group of authors sued Meta, claiming the company used their books without permission to train its AI models.
How Meta Plans to Make Money from AI
The documents fail to clarify what Meta considers a “generative AI product.” Nonetheless, the company is already generating revenue from AI through various avenues. It has established revenue-sharing agreements with platforms utilizing its Llama AI models, and it recently introduced a new API that allows developers to customize and experiment with Llama tools.
Mark Zuckerberg, Meta’s CEO, said Meta AI could eventually show ads or offer paid features through a subscription model.
Massive Spending on AI Development
Meta is spending heavily on building its AI tools. In 2024, the budget for generative AI was over $900 million. That figure may pass $1 billion in 2025. These numbers don’t include the cost of data centers. Meta plans to spend up to $80 billion on infrastructure next year.
Controversy Over AI Training Data
The lawsuit also highlights how Meta gathered training data for its AI. In 2023, the company considered spending more than $200 million to license high-quality data. About $100 million would have gone toward buying books. Instead, the authors claim Meta chose to use pirated ebooks.
Meta denies these claims. The company argues that it followed fair use rules. In a statement, Meta said its AI tools help boost innovation and creativity.
“We disagree with the authors’ claims,” the company said. “We will continue to defend our work and support AI development that benefits everyone.”
Check out other related news:
