Meta

Overview

Meta has documented content relationships for both model development and information delivered within Meta AI. In January 2023, Shutterstock announced an expanded partnership under which its images, videos, and music would support Meta's development, training, and evaluation of machine-learning capabilities.

Separately, Meta's March 2026 news-partnership announcement names News Corp, Le Figaro, Prisa, and Süddeutsche Zeitung. That announcement describes providing timely information and links in Meta AI responses, helping users reach publishers' articles for more detail.

These relationships establish different kinds of demand. The Shutterstock announcement explicitly concerns model development and training. The news announcement describes access used when answering questions; it does not establish that the named publishers' material is used to train models. Keeping those uses separate is essential when interpreting Meta's place in the data market.

What the partnerships illustrate

Our assessment is that the same broad category, licensed content, can describe very different products. An illustrative visual-data collection may be valued for diversity and annotations during training, while a news feed may be valued for freshness and source links during answer generation. Their quality criteria and delivery requirements are not the same.

A seller should therefore identify the intended use before deciding how to package an archive or describe its value. The existence of one partnership does not establish that another collection is wanted, or that terms from one type of content would apply to another.

Assessing a potential fit

The cited announcements describe established partnerships rather than an open purchasing program with published rates. Prospective suppliers should prepare information about formats, coverage, update frequency, provenance, and the uses they can support. For current-information products, source attribution and ongoing delivery can matter alongside the underlying content. For training material, diversity and structure may be more central. Confirm the relevant team and commercial route before treating Meta as an active prospect for a specific asset. This profile records documented demand without estimating deal value, assuming universal intake, or extending a runtime-content agreement into an unsupported claim about model-training rights.