You could say that the software litigation landscape of 2025 was dominated by “The Great AI Copyright Reckoning.” The software industry moved from theoretical anxiety to high-stakes courtroom battles over how data is ingested and how code is generated. Here are a few thumbnails of the major cases as well as a quick look at the trends we’re seeing in 2026.
- Perhaps the most impactful software-adjacent case of the decade was the $1.5 billion settlement of Bartz v. Anthropic.
Anthropic faced massive statutory damages for allegedly training its models on pirated libraries (LibGen and PiLiMi).
• The Outcome: A landmark $1.5 billion settlement in September 2025.
• Impact: This set a massive “floor” for the cost of unlicensed training data and accelerated the trend of AI developers seeking formal licensing deals with content owners.
- The Transformative Use split: Kadrey v. Meta
Decided almost concurrently with the Anthropic motions, the court in Kadrey offered a “mixed bag” for developers.
• The Ruling: While the court found AI training to be largely “transformative” (leaning toward Fair Use), it opened the door for market dilution arguments.
• Impact: It signaled that “training” might be legal, but “competing” with the original data source via AI output is where the legal liability lies.
- The Server Test Survives: McGucken v. Valnet
The Supreme Court declined to review the Ninth Circuit’s “Server Test” in April 2025.
• The Ruling: Confirmed that embedding content (linking to software/images hosted elsewhere) does not constitute a “display” right violation under the Copyright Act.
• Impact: A major win for SaaS platforms and web-based software that rely on embedding and API-led architectures.
2026: Emerging Trends and Predictions
This year, software litigation is shaping up to be the year of “Agentic Liability” and “Regulatory Harmonization.” If 2025 was about inputs (data training), 2026 is about outputs and autonomy.
- The Rise of Agentic Liability
As software moves from “chatbots” to “agents” that can execute code and sign contracts (Agentic AI), a new legal frontier is emerging.
• Trend: Courts are beginning to grapple with “Electronic Agency.” If an AI agent makes a $1M procurement error, is the software developer or the user liable?
• Prediction: 2026 will likely see the first major “Autonomous Error” case reaching a high court, likely involving a breach of contract claim triggered by an AI agent.
- The “State Patchwork” vs. Federal Preemption
With the Colorado AI Act (effective June 2026) and California’s AB 2013 (training data disclosure), software companies face a fragmented regulatory map.
• Trend: A “flight to the highest standard.” Large software firms are beginning to adopt California/EU standards globally to avoid managing 50 different codebases for compliance.
- “Shadow AI” and E-Discovery
A major trend for 2026 is the litigation surrounding “Shadow AI”—employees using unapproved AI tools that leak trade secrets.
• Trend: Corporate litigation is increasingly targeting software vendors for “inducement of trade secret misappropriation” if their tools make it too easy for employees to bypass corporate silos.
The shift from “Copyright” to “Compliance” seems to be the early defining theme of 2026. Software firms that cannot provide a “clean” provenance for their training data or a clear “human-in-the-loop” audit trail for their agents will likely see their valuation discounted due to increased litigation risk.
