Meta Paid Up After Scale AI Demanded Millions for Its Data - And Got It

In late 2023, Scale AI cut off Meta’s access to a dataset that Meta had quietly relied on to train its AI models. Scale noticed heavy use from Meta servers and flagged the activity. Rather than backing down, the startup issued a simple message pay up or lose access.

Meta initially resisted. However, as pressure mounted internally, especially from the team building its Llama AI models, leadership changed course. Engineers warned that pulling the data mid-training would derail ongoing development. That threat pushed Meta into negotiations.

By early 2024, the companies reached an agreement. Meta paid tens of millions of dollars. In return, it regained access to Scale’s data and avoided a major disruption to its AI roadmap.

Other recent Meta news –

• Meta Faces Antitrust Heat with Massive Scale AI Bet – Here’s What You Need to Know
• Meta Bets $15B on Scale AI, Taps Founder for New AGI Lab
• Meta Showcases V-Jepa 2, a World Model That Helps Robots Predict Before They Act – Here’s How

This incident shows a shift in how AI startups handle data deals. Until now, Big Tech companies have often used training sets without compensating the original providers. Scale’s move sets a precedent, startups are willing to block access unless they get paid.

OpenAI, Anthropic, and others are now keeping a closer eye on who’s using their data and how. The Meta-Scale clash shows that the balance of power is shifting. The days of scraping datasets without oversight might be over.

You may also be interested to read –

• Anthropic Finally Reveals How It Built Its Multi-Agent Claude Research System – Know All About It Here
• Google’s Android Enterprise Turns 10 – Here’s What’s Coming Next
• Horizon3.ai secures $100 million Series D to automate security for enterprise and federal networks