Cloudflare acquires Human Native to build a market for AI training data

Cloudflare has acquired Human Native, a UK-based startup building a marketplace to connect content creators with AI developers, as the internet infrastructure firm deepens its push into the economics of the AI-driven web.

The deal brings Human Native’s team and technology into Cloudflare at a moment when the AI industry’s demand for high-quality training and inference data is colliding with growing resistance from publishers, artists, and media companies whose work is being scraped with little control or compensation. Financial terms were not disclosed.

Cloudflare says the acquisition will help it build tools that allow creators and publishers to decide how their content is used by AI systems — whether that means blocking AI bots entirely, optimising content for AI discovery, or offering data for sale at prices set by the content owner. The company is effectively positioning itself as infrastructure for a new layer of the internet: one where data used by AI is tagged, priced, and exchanged rather than quietly extracted.

“Content creators deserve full control over their work, whether they want to write for humans or optimise for AI,” said Matthew Prince, co-founder and CEO of Cloudflare. “Last year we gave publishers and creators the tools to control which bots can access their content, but the real goal has always been to help create a new economic model that actually works for the next phase of the Internet.”

That framing matters. Until now, much of the debate around AI data has been binary — block AI entirely or accept that content will be used without compensation. Cloudflare is arguing for a third path: markets. If successful, this would shift the fight over AI training data away from courtrooms and into pricing, discovery, and licensing infrastructure. It would also give Cloudflare more influence over how AI traffic flows across the web, extending its role from gatekeeper to broker.

Human Native was founded in 2024 and backed by UK venture firms LocalGlobe and Mercuri. The startup focused on helping AI developers discover structured, reliable datasets, while allowing creators to surface their work in ways that preserve attribution and payment. Its team includes alumni from DeepMind, Google, Figma, and Bloomberg.

“We started Human Native with the goal of getting generative AI out of its Napster era,” said James Smith, co-founder and CEO of Human Native. “We believe that creators should have control, compensation and credit when their work is used to power AI products.”

The “Napster era” comparison reflects a growing belief across media and policy circles that today’s AI data practices mirror early peer-to-peer file sharing: technically powerful, economically unsustainable, and legally unstable. By acquiring Human Native, Cloudflare is betting that AI companies will eventually need cleaner, more transparent data supply chains — not just to avoid lawsuits, but to secure higher-quality inputs as synthetic data and scraped content hit diminishing returns.

Investors backing Human Native see the same shift. “High quality data is the key that unlocks real differentiation in AI,” said Ziv Reichert, a partner at LocalGlobe, pointing to rising competition among model builders and the limits of freely available web data. Alan Hudson, founding general partner at Mercuri, said the need to reward creators fairly has only intensified as generative AI has moved from research labs into commercial products.

For creators and publishers, the promise is optionality rather than obligation. They may still choose to block AI systems entirely. But the acquisition suggests that large infrastructure players now see paid data access — not unrestricted scraping — as the more stable long-term model.

For Cloudflare, the move goes beyond content protection. It is an attempt to shape how value is assigned and exchanged in the AI economy, at a layer of the internet most users never see. Whether AI developers accept higher data costs, and whether creators gain real leverage rather than symbolic control, remains uncertain. But the direction of travel is clear: the era of “free” AI data is under pressure, and the plumbing of the internet is starting to reflect that reality.