An Nvidia shareholder has filed a 108-page lawsuit against the company's founder and CEO Jensen Huang and other board members, accusing them of damaging the chipmaker by building its artificial intelligence services on copyrighted and unlicensed material.

Jessica Berliner filed the complaint on July 31, 2026, in the US District Court for the Northern District of Illinois. The case, docketed as Berliner v. Huang (Case 1:26-cv-09153), is a shareholder derivative suit — a legal action in which a shareholder sues a company's own executives on behalf of the corporation itself, rather than on behalf of outside parties like authors or publishers. Courthouse News

The suit accuses Nvidia of "adopting and implementing an unlawful business strategy" of using copyrighted and unlicensed materials — including books, videos and commercial voices — to develop its AI services. Berliner's complaint alleges that Nvidia's large language models were trained using a database of nearly 200,000 pirated books, and that the company unlawfully copied those works. Publishers Marketplace

The defendants, who include Huang and other Nvidia officers and directors, are accused of violating their fiduciary obligations — the legal duty executives and board members owe to the company and its shareholders to act in the corporation's best interest. Publishers Marketplace

What makes this case unusual is who is suing. Most AI-training copyright disputes to date have been brought by creators — authors, musicians, artists — who say their work was taken without permission. Berliner's suit turns that logic inside out: she is arguing that by exposing the company to that very legal risk, Nvidia's leadership failed the shareholders they are supposed to protect.

Nvidia has faced a steady stream of such copyright claims. Three authors sued the company in March 2024, alleging it used their copyrighted books without permission to train its AI. Reuters In March 2026, a 3D artist filed proposed class actions against Nvidia, Meta and Roblox, alleging the companies misused millions of 3D models. Reuters In June 2026, music company Jamendo sued Nvidia, claiming it used Jamendo's music and data to train its Fugatto and Audio Flamingo AI models. Reuters

Berliner's complaint reframes those earlier suits not as separate grievances from outside parties but as evidence that Nvidia's own leadership created a liability for the company it runs. Bloomberg Law

The case is at its earliest stage. No trial date has been set, and Nvidia has not yet filed its response in court.