Ziff Davis CEO: if OpenAI can find $750 billion for data centers, it can find money for publishers, too
Ziff Davis CEO argues that the Justice Department’s claim that licensing would cripple AI development is unfounded. He cites existing publisher‑lab agreements and the massive compute budget of OpenAI to show licensing costs are manageable. The piece calls for a streamlined licensing framework to be…
The U.S. Justice Department recently released a Statement of Interest in a copyright infringement case in the Southern District of New York. Ziff Davis, a major digital media publisher, is one of the plaintiffs. In a separate commentary, the CEO of Ziff Davis argues that the government’s concerns about licensing agreements between artificial‑intelligence (AI) labs and publishers are misplaced and that a robust licensing framework can coexist with the rapid growth of AI technology.
Government Claims About Licensing and AI
The Department of Justice’s brief contends that if frontier AI labs—such as OpenAI, Anthropic, and others—were required to negotiate licensing agreements with publishers, it would become "significantly more difficult to develop a robust AI industry." The brief acknowledges that many publishers have already entered into deals with these labs, yet it insists that licensing would pose a logistical and financial burden.
In a footnote, the brief admits it takes no position on whether a licensing regime would be financially or logistically feasible. It also notes that both mainstream and independent publishers have already entered into licensing agreements. The question then becomes: where is the real difficulty?
Existing Licensing Platforms Make the Process Simple
Over the past 18 months, several platforms have emerged to streamline licensing for AI training data. Cloudflare, Tollbit, Really Simple Licensing, and the News Media Alliance have built systems that simplify royalty payments and contract management. These platforms are modeled on well‑established licensing bodies such as ASCAP and BMI, which have successfully managed rights for music creators for decades.
These proven frameworks demonstrate that licensing can be both efficient and fair. They allow publishers of all sizes to negotiate terms that reflect the value of their content while providing AI labs with the data they need to train models.
Cost of Licensing vs. Cost of Compute
The Department’s argument that only the largest technology companies can afford licensing fees is misleading. OpenAI alone has projected that it will spend $750 billion on compute infrastructure—data centers, chips, and power—by 2030. This figure dwarfs the $20 billion spent annually on music royalties in the United States.
Even if publishers were to receive a comparable share of licensing revenue, the cost would be a negligible fraction of the overall expenses of an LLM company. For instance, a $20 billion royalty stream would represent less than 3% of a $750 billion compute budget. Yet for content creators, that sum could provide meaningful income and support continued production of high‑quality journalism.
The Real Barrier to AI Competition Is Capital, Not Licensing
OpenAI’s massive capital requirements mean that only a handful of firms can afford to develop frontier models from scratch. Startups in the AI space are increasingly building on top of existing models rather than competing directly. This trend would be stifled if licensing costs were treated as an insurmountable hurdle.
Moreover, without a licensing framework, publishers may resort to paywalls and other restrictions to protect their content. This would raise costs for consumers, limit access to free information on the open web, and reduce revenue streams for writers and independent creators.
What Happens Next?
The Justice Department’s brief has sparked debate among publishers, AI labs, and policymakers. Ziff Davis and other media companies are calling for a clear, streamlined licensing regime that protects intellectual property while enabling AI innovation. The outcome of this debate could shape the future of both the publishing industry and the AI sector.
As the conversation continues, stakeholders must balance the need for robust AI development with the rights of content creators. A fair licensing system could provide a win‑win scenario, ensuring that AI firms have the data they need while publishers receive fair compensation for their work.
Why it matters
The discussion determines whether AI’s rapid growth will be hampered by licensing hurdles, potentially affecting innovation and the livelihoods of content creators.
Key points
- Justice Department claims licensing will hinder AI growth
- Existing licensing platforms simplify agreements
- OpenAI’s $750B compute budget dwarfs potential licensing costs
- Licensing can benefit both AI firms and publishers
- Without licensing, publishers may impose costly paywalls
Frequently asked questions
Why does the Justice Department believe licensing will slow AI development?
The brief argues that negotiating licensing agreements would add logistical and financial burdens to frontier AI labs, making it harder for them to develop robust models.
What platforms exist to streamline licensing for AI?
Platforms such as Cloudflare, Tollbit, Really Simple Licensing, and the News Media Alliance offer turnkey solutions for royalty payments and contract management.
How does OpenAI’s compute budget compare to licensing costs?
OpenAI plans to spend $750 billion on compute by 2030, which far exceeds the $20 billion annual music royalty market, making licensing a relatively small expense.
What could happen if publishers impose paywalls?
Paywalls could raise costs for consumers, limit free access to information, and reduce revenue for writers and independent creators.
