OpenAI and Anthropic, two of the world’s most prominent artificial intelligence developers, are reportedly exploring plans to use investor funds to help cover potential multibillion-dollar legal settlements tied to ongoing copyright infringement cases.
According to a report by the Financial Times on Wednesday, the companies have held internal discussions about deploying part of their venture capital and corporate funding to offset legal liabilities if courts rule that their AI models unlawfully used copyrighted materials. The report underscores growing financial and legal pressure on the AI sector as a wave of lawsuits from authors, media organizations, and entertainment companies challenge how generative AI systems were trained.
Copyright owners have filed a series of high-stakes lawsuits against technology giants, including OpenAI, Microsoft, and Meta Platforms, accusing them of scraping and reproducing protected works to train AI systems without authorization. The suits, many of them class actions filed in U.S. federal courts, allege that large language models like OpenAI’s GPT and Anthropic’s Claude have effectively copied vast quantities of text, books, and visual media from the internet without compensation.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026): big discounts for early bird.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab: From Technical Design to Deployment (next edition begins Jan 24 2026).
The financial exposure from such cases could be enormous. Industry lawyers estimate potential liabilities across the AI sector could exceed tens of billions of dollars, depending on how courts interpret “fair use” exemptions and whether AI outputs are deemed derivative works.
OpenAI’s Insurance Limits and Risk Strategy
The Financial Times said OpenAI has sought to protect itself by purchasing insurance coverage for emerging AI-related risks, reportedly through Aon, a major global insurance broker. Sources familiar with the policy told the paper that OpenAI secured coverage worth up to $300 million, though another insider disputed that figure, suggesting the true amount is “significantly lower.”
Regardless of the exact figure, both sources agreed that the coverage falls far short of the potential financial exposure the company faces from ongoing and future lawsuits.
Kevin Kalinich, Aon’s global head of cyber risk, told the FT that the broader insurance market is not yet equipped to handle the scale of liabilities facing AI model developers.
“There’s not enough capacity for [AI model] providers,” he said, citing the novelty and unpredictability of AI-related claims.
Given those constraints, OpenAI is reportedly considering a form of “self-insurance” — setting aside a portion of its investor capital in a ring-fenced “captive” vehicle to manage risk internally. Captives are a common mechanism used by large corporations to insure against unique or hard-to-price risks, such as cyberattacks, environmental liability, or product failure.
The report noted that OpenAI’s investor pool includes heavyweights such as Microsoft, Thrive Capital, and Andreessen Horowitz. Any move to use investor funds for insurance or settlements would likely require their approval, given the scale of the potential financial commitments.
Anthropic’s Legal Exposure and Recent Settlement
Anthropic, another leading AI developer backed by Amazon and Google, faces similar legal battles and has reportedly begun using its own capital reserves to handle potential liabilities. The Financial Times cited a person familiar with the company’s finances who said Anthropic is partly funding a recent $1.5 billion preliminary settlement of a copyright class action brought by a group of authors in California.
The settlement, preliminarily approved by a federal judge last month, marks one of the largest copyright-related deals in AI’s short history and could set a precedent for how future claims against AI firms are resolved.
A Legal and Financial Reckoning for AI
The mounting legal challenges highlight an unresolved issue at the heart of the AI revolution — whether developers can legally use publicly available data, including copyrighted material, to train models that now power billion-dollar businesses.
AI developers have argued that their training practices fall under the U.S. legal doctrine of “fair use,” which allows limited reproduction of copyrighted works for purposes such as research and innovation. But publishers, authors, and artists say the models have created a new kind of industrial-scale copying that deprives them of compensation while generating enormous profits for tech companies.
The lawsuits also expose the gap between AI companies’ insurance protections and their real-world financial exposure. Traditional insurers have struggled to model AI-related risks due to the lack of historical precedent and the fast pace of technological change.
If OpenAI and Anthropic proceed with self-insurance strategies, it would represent a significant shift in how AI companies allocate investor capital — diverting funds meant for product development into legal defense reserves. Analysts warn that such moves could slow innovation and signal the growing financial strain the AI industry faces as it collides with copyright law.
For investors, the question now is whether AI companies can remain attractive amid rising legal uncertainty. The lawsuits are believed to be a fundamental challenge to the AI business model, and if courts decide that these models rely on unlawful data use, the entire industry will have to rethink how AI is built and trained.