A company that has filed more than 6,000 copyright lawsuits against individual defendants—quickly settling most for amounts carefully calculated below the cost of defense—has now set its sights on Meta, a corporation worth more than $1 trillion.
Strike 3 Holdings’ recent lawsuit against the tech giant alleging widespread piracy for training of artificial intelligence represents a dramatic strategic shift that will test whether technical evidence sufficient to extract settlements from resource-constrained individuals can survive the scrutiny of one of the world’s most well-funded legal departments. The case also promises to reveal whether Strike 3’s business model represents legitimate copyright enforcement or a sophisticated form of legal extortion that finally picked the wrong target.
For full disclosure, I’ve been retained as an economics expert by individual defendants in several other cases in which Strike 3 has alleged copyright infringement. I offer my thoughts based on my experiences.
Strike 3’s Allegations
Strike 3 claims Meta infringed at least 2,396 of its copyrighted works since 2018 by using the BitTorrent protocol to download content for training AI models, including LLaMA and Movie Gen. BitTorrent operates by allowing users to simultaneously download and upload file pieces from multiple connected users, creating a distributed network for large file transfers.
The complaint alleges two specific strategies by Meta:
- The company systematically downloaded Strike 3’s films to acquire massive volumes of video content for AI training; and
- Meta actively distributed or “seeded” Strike 3’s popular content to exploit BitTorrent’s reciprocal “tit-for-tat” mechanism, which rewards users who upload content by providing them faster download speeds when acquiring other files.
Strike 3 contends this alleged activity was willful infringement for commercial gain, positioning Meta to profit substantially from AI programs trained on pirated content. The company seeks statutory damages that could reach $150,000 per willfully infringed work under 17 U.S.C. 504—which could amount to nearly $360 million in damages if Strike 3 prevailed and obtained the maximum damages.
The Economics of Strike 3’s Business Model
My expert analysis in previous Strike 3 cases has revealed a litigation strategy that depends heavily on settlement revenue, rather than traditional content distribution. Legal scholar J. DeBriyn of the UCLA Entertainment Law Review defines a “copyright troll” as a plaintiff who seeks damages not to be made whole, but as a primary revenue stream. This model typically involves filing numerous lawsuits against individual defendants and using the threat of high statutory damages and expensive litigation costs to compel settlements.
Deposition testimony from a Strike 3 corporate representative in a prior case indicated the company receives between $7.2 million and $9.6 million annually in settlement revenues. My analysis of the firm’s claimed production costs and subscriber numbers suggests that without this litigation income, the business would likely be unprofitable.
Quantitative analysis of federal court data supports this assessment. Using Kaplan-Meier survival analysis, a statistical method for measuring time until case resolution, I found telling patterns in case duration:
- Non-pornography copyright cases: 215 days median time to dismissal
- Pornography copyright cases (excluding Strike 3): 177 days
- Malibu Media (a similar litigant): 140 days
- Strike 3: 113 days
Strike 3’s cases resolve nearly twice as fast as typical copyright lawsuits, consistent with a “sue-then-settle” strategy that avoids adjudication on the merits. Judge Royce C. Lamberth of the U.S. District Court for the District of Columbia has observed that serial litigants like Strike 3 “drop cases at the first sign of resistance.”
A Strategic Reversal and Evidence Problems
The Meta lawsuit inverts Strike 3’s established model. The company’s profitability from sue-then-settle has historically depended on resource asymmetry, where an individual defendant’s cost to defend exceeds the settlement demand. Against Meta, with its market capitalization exceeding $1 trillion, this asymmetry reverses. The threat of litigation costs that compel settlement from individuals represents a pittance for Meta.
Strike 3’s strategic shift creates a credibility problem. Success against Meta depends entirely on the integrity of Strike 3’s technical evidence, generated by proprietary systems called “VXN Scan” and “Cross Reference Tool.” My analysis of the evidentiary data Strike 3 produced in other cases, however, reveals serious flaws.
In one matter, I reviewed two different Excel files of alleged infringement data provided by Strike 3. The files were materially different, with one version inexplicably excluding all works listed in the complaint and obscuring data patterns by removing records of multiple downloads. The data showed illogical patterns inconsistent with single-user behavior, such as a defendant downloading books in multiple foreign languages using Cyrillic, Chinese, and Thai scripts on the same day they allegedly downloaded adult content. The defendant couldn’t read Cyrillic, Chinese, or Thai.
Evidence sufficient to coerce a settlement from defendants lacking resources for a forensic audit will face much higher scrutiny from Meta. The case will test an evidentiary system that my analysis suggests is unreliable. In other words, Meta’s scrutiny of Strike 3’s process for collecting evidence would likely demonstrate that it is unreliable, thereby blowing up the basis for its suits against individual defendants.
The Discovery Strategy Hypothesis
The apparent mismatch between Strike 3’s model and Meta’s resources raises the question of why the company would pursue a case where the defendant can easily afford to successfully challenge the plaintiff’s technical evidence. The answer may lie in weaponizing the pre-trial discovery process.
Discovery requires parties to exchange information, including internal documents, emails, financial records, and employee testimony. For a company Meta’s size, this process is enormously expensive. It creates risk exposure by forcing the revelation of internal communications that could generate negative publicity, regardless of whether they prove liability.
Strike 3 filed its complaint after another case, Kadrey v. Meta, had proceeded through discovery. That process revealed Meta employees used BitTorrent to download pirated books for AI training and took steps to conceal this activity. Strike 3’s complaint explicitly references these findings, quoting an internal Meta communication: “not sure we can use Meta’s IPs to load through torrents pirate[sic] content, ahah.”
Armed with this knowledge, Strike 3’s strategy becomes clearer. The lawsuit represents a calculated gamble that Meta will determine the cost and risk of protracted discovery exceed the cost of a substantial settlement, regardless of the merits.
Implications for Copyright Policy
This case represents a collision between mass intellectual-property enforcement and mass data acquisition for technological innovation. The outcome will test whether procedural costs and the reputational risks of discovery can extract settlements from well-resourced defendants, forcing closer examination of the data-acquisition practices underlying AI development.
From an economic standpoint, copyright damages should make plaintiffs whole or deny liable parties wrongful profits. Arrangements providing windfall profits to plaintiffs result in resource misallocation. As Judge Richard Posner noted, attempting to force settlements using high litigation costs against opponents lacking resources to resist constitutes abuse of process.
The Meta case will determine whether Strike 3’s technical evidence can survive rigorous scrutiny and whether copyright enforcement serves its intended purpose of stimulating artistic creativity, or merely generates revenue through systematic threats of litigation. The result may influence how courts evaluate similar claims in an era when AI development increasingly relies on vast datasets of potentially copyrighted material.
