Meta forecasts substantial revenue growth from its generative AI products, projecting earnings between $2 billion to $3 billion by 2025 and an impressive $460 billion to $1.4 trillion by 2035. This information was revealed in recently unsealed court documents submitted by authors suing the company for allegedly training its AI systems using their works without permission.
While the documents do not specify what Meta defines as “generative AI products,” it is evident that the tech giant is diversifying its AI revenue streams. Meta has entered revenue-sharing arrangements with platforms that host its open Llama AI models and recently launched an API for customisation and assessment of these models. During a recent earnings call, CEO Mark Zuckerberg mentioned that Meta AI might implement advertising and offer a premium subscription tier for advanced features.
The legal documents also shed light on Meta’s significant investment in AI. The budget for its “GenAI” initiatives exceeded $900 million in 2024 and is expected to surpass $1 billion this year, excluding the costs related to the infrastructure required to support AI model training. The company has signalled intentions to invest between $60 billion and $80 billion in capital expenditures by 2025, primarily for the development of extensive new data centres.
There are suggestions that Meta’s operating budget could have been even higher if it had pursued licensing agreements with the authors suing it. Reports indicate that discussions took place regarding spending over $200 million to acquire training data for the Llama models, which included approximately $100 million earmarked for book acquisitions. However, it appears Meta opted for alternative methods, allegedly involving large-scale piracy of ebooks.
Despite attempts to seek comment from Meta regarding these matters, the company has not yet provided a response.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
