- Published on
The New York Times updated its lawsuit, accusing Microsoft of customizing supercomputers to help OpenAI train AI models
- Authors
- Name
- DeepHorus
The New York Times Escalates Legal Battle: Updated Lawsuit Accuses Microsoft of Building Custom Supercomputers for OpenAI's AI Training
The New York Times has significantly amplified its legal assault on the AI industry, filing an updated lawsuit on June 26, 2026, with a startling new accusation. The media giant alleges that technology giant Microsoft specifically designed and built custom supercomputing infrastructure to enable OpenAI to train its powerful artificial intelligence models on vast quantities of The Times's copyrighted journalism without authorization. This move transforms the dispute from a question of data scraping to one of dedicated, technical enablement for large-scale copyright infringement.
The legal confrontation began in December 2023, when The Times filed its original lawsuit against Microsoft and OpenAI. That initial complaint centered on the allegation that both companies had used millions of The Times's articles to train their large language models (LLMs), like GPT-4, without permission or compensation. The new filing represents a major development in the ongoing case, which has become a landmark test for copyright law in the age of generative AI.
The core of the new allegation is highly technical. The updated complaint, as reported by tech news outlet Ars Technica, claims that Microsoft went beyond simply providing general cloud services. It states that Microsoft "designed and built specialized supercomputers" specifically for OpenAI's needs. These systems were allegedly optimized to process the colossal datasets required for AI training, thereby facilitating the direct ingestion of The Times's protected articles.
According to the lawsuit, this customized hardware enabled the systematic web crawling and ingestion of copyrighted works from The Times website. The suit contends that this process was not incidental but a deliberate part of the AI training pipeline, resulting in models that can now reproduce The Times's style, substance, and sometimes specific output when prompted. The newspaper argues this constitutes a fundamental threat to its business model and the value of original journalism.
The implications of this lawsuit extend far beyond The Times itself. If successful, the updated claims could establish a precedent that holding companies liable not just for the act of copying, but for providing the specialized tools and infrastructure that make such large-scale copying possible. This could have a profound chilling effect on the development of future AI systems, forcing developers to more scrupulously vet their training data and the hardware pipelines they use.
Reactions from the accused parties have so far been limited in public detail, but Microsoft has previously stated that it "respects the rights of content creators" and builds its AI services in compliance with the law. OpenAI has argued that its training on publicly available internet data falls under fair use principles. The updated lawsuit, however, directly challenges the notion that such massive, system-level appropriation of journalism can be considered fair.
Industry analysts are watching the case closely as it moves forward in court. A ruling in favor of The New York Times could force a massive restructuring of how AI models are trained, potentially requiring explicit licensing agreements with publishers. Conversely, a decision favoring the AI companies could solidify the current practices of using broad swaths of internet content for training, setting a permissive standard for the burgeoning industry.
As this legal drama unfolds, the outcome will likely shape the future intersection of intellectual property and artificial intelligence for years to come. The New York Times is not merely seeking damages; it is fighting for a principle that could redefine the boundaries of digital creation and innovation in the 21st century.
