NYT Sues Microsoft and OpenAI Over Infringing Supercomputer

What Happened?

On June 25, 2026, The New York Times filed a motion to amend its lawsuit against OpenAI and Microsoft, alleging that Microsoft built a custom supercomputer, among the most powerful in the world, to facilitate copyright infringement by OpenAI. The motion, reported by Ars Technica, seeks to adjust the contributory infringement claim against Microsoft following the Supreme Court's ruling in Sony vs. Cox Communications, which raised the standard of proof for such claims. According to NYT spokesperson Graham James, new evidence obtained during discovery strengthens the case: "Today we ask permission to file an amended complaint that clarifies our contributory infringement claim against Microsoft, based on new law and new evidence."

The motion, filed in the U.S. District Court for the Southern District of New York, includes an amended version of the lawsuit detailing how Microsoft allegedly designed and deployed a custom supercomputing system for OpenAI, with the express purpose of training language models like GPT-4. According to court documents cited by Ars Technica, the system reportedly used tens of thousands of GPUs and ranks among the most powerful in the world, comparable to supercomputers at government research centers. The NYT claims Microsoft not only provided the infrastructure but actively participated in planning its use to scan and process copyrighted articles without a license.

This legal move comes after the Supreme Court in March 2026 ruled in favor of Cox Communications in a related music piracy case, establishing that to prove contributory infringement, the plaintiff must show the defendant acted intentionally to induce the illegal conduct. The NYT acknowledges that this new legal standard requires adjusting its original argument, which relied on a broader theory of vicarious liability. Now, the newspaper seeks to demonstrate that Microsoft had a specific intent in building the supercomputer: to enable OpenAI to infringe copyright on a massive scale.

Why Is This Important?

This case goes beyond a dispute between a publisher and an AI company. It addresses the liability of technology infrastructure providers when their products are used to infringe copyright. If the court accepts the NYT's theory, Microsoft could be held responsible not just for OpenAI's software, but for deliberately creating the computational tools necessary for infringement. This would set a historic precedent in applying copyright law to the era of generative artificial intelligence.

The Supreme Court's ruling in Cox Communications raised the bar for contributory infringement claims, requiring proof of direct intent. The NYT now alleges that Microsoft designed the supercomputer with that specific purpose, an argument that, if accepted, would significantly expand the scope of big tech liability. Compared to previous cases, such as Getty Images vs. Stability AI or authors vs. Meta, this focus on physical infrastructure and provider intent is novel and could have far-reaching implications.

Historically, contributory infringement liability has been applied to platforms like Napster or Grokster, where software facilitated illegal file sharing. However, here the defendant is not a mere intermediary but a hardware builder specifically designed for a task that, according to the plaintiff, is inherently infringing. This echoes cases of photocopiers or video recorders, where courts assessed whether the device had substantial non-infringing commercial uses. The key difference is that Microsoft's supercomputer was used exclusively to train AI models, and the NYT argues that training relied on massive copying of its articles.

What Will Be the Consequences?

If the motion is approved and the NYT succeeds, companies that build hardware or provide cloud services for training AI models could face increased legal scrutiny. This could lead to changes in how AI infrastructures are designed and marketed, with stricter contractual clauses on the use of copyrighted data. For example, cloud providers like Amazon Web Services or Google Cloud might be forced to implement content filters or require customers to guarantee data licenses.

For Microsoft, the consequences are immediate: a possible order to cease using the supercomputer or pay millions in damages. According to analyst estimates, building such a system could cost billions of dollars, and a court order could halt the development of GPT-5 or future models. Additionally, Microsoft could face legal sanctions affecting its Azure cloud business, which competes directly with AWS. For the AI ecosystem, this case could slow investment in specialized infrastructure if investors perceive high legal risk. AI startups relying on third-party hardware could see their compliance costs rise.

In the market, the news has already caused volatility: Microsoft shares fell 2% on the day of the announcement, while OpenAI (valued at $150 billion) could be affected if investors fear disruption in access to the supercomputer. In the long term, the case could incentivize AI companies to develop their own training data or negotiate licenses with publishers, as OpenAI has already done with some media outlets. However, the focus on infrastructure could shift the balance of power toward hardware owners like Nvidia, which could benefit if companies seek to build their own systems to avoid litigation.

What Should Readers Know?

The NYT vs. OpenAI and Microsoft case is a milestone in AI regulation. Unlike previous lawsuits centered on training data, this one focuses on physical infrastructure and the intent of its creator. Readers should understand that the outcome could define the limits of corporate liability in the era of generative AI. Moreover, the use of the supercomputer as evidence shows that computational scale is a key factor in the ability to infringe copyright. As AI models require more computing power, companies providing that power could become legal targets.

It is important to note that the NYT's motion has not yet been approved by the court; the judge must decide whether to allow the amendment. If granted, the case will enter a deeper discovery phase, where internal emails from Microsoft and OpenAI will be examined to determine intent. This process could take months or years, and any appeal could reach the Supreme Court. Readers should follow this case closely because, regardless of the outcome, it will set a legal precedent affecting the entire tech industry. The combination of copyright, AI infrastructure, and corporate liability is a cocktail that will redefine how artificial intelligence is built and deployed in the future.