Chris Curl,
Editor
Oct. 8, 2026
The AI industry has discovered a convenient definition of unfair competition.
Build a model using the accumulated work of writers, programmers, researchers, artists, and generations of human civilization, and you are advancing the frontier.
Let a competitor learn from your model’s answers, and suddenly the conversation becomes theft, national security, and government intervention.
That contradiction deserves attention.
The fight between Chinese open-model developers and American companies like Anthropic and OpenAI is about more than technical leadership. It is about whether intelligence becomes an accessible tool, or a service controlled by a handful of companies collecting rent from everyone else.
Protecting America or Protecting Margins?
Models like Kimi and DeepSeek give developers alternatives to proprietary platforms. Many are more accurately described as open-weight than fully open-source, but the practical advantage remains: businesses can download, customize, and run them on their own infrastructure.
That threatens an expensive business model.
American frontier labs are committing enormous resources to compute and infrastructure. They need customers to pay enough to justify those commitments. Capable, lower-cost alternatives make that harder.
There are legitimate security concerns around foreign technology. But protecting American AI companies from competition is not automatically the same thing as protecting American entrepreneurs.
One wants to sell intelligence at a premium.
The other wants to use it affordably.
An American startup does not owe a frontier lab a return on its capital spending. It needs the best tools for the job, not an obligation to subsidize someone else’s infrastructure bill.
Who Created the Knowledge?
The allegations against Chinese developers deserve scrutiny. Unauthorized access, contractual violations, and secretly routing customer data through another provider are specific issues that should be judged on evidence.
But they do not erase the larger question.
These models did not invent human language, mathematics, science, literature, or software. They learned from a civilization that generations of people built.
The labs contributed valuable engineering. That does not give them ownership of the intellectual foundations beneath it.
Not every training source raises the same legal questions, and learning from human work is not legally identical to extracting a competitor’s service. Still, the moral tension is obvious: humanity supplies much of the knowledge, while a small number of corporations seek to control (and charge for) access to the resulting intelligence.
If there is a case that someone is being ripped off, the public deserves a place in that conversation.
Competition Is the Opportunity
Defending open models does not mean ignoring China’s restrictions or security risks. It means defending a principle: entrepreneurs should have access to competing tools, with safeguards based on actual risks rather than incumbent convenience.
Cheaper intelligence can help American businesses build better products, automate costly workflows, and compete with much larger companies.
That is an opportunity… not something to shut down because it threatens a model provider’s margins.
For investors, the distinction matters. AI adoption can accelerate while the price of intelligence falls. The technology can succeed even if some of the companies selling it struggle to justify their valuations.
That is what Digital Dispatch follows: the difference between a transformative trend and the business models trying to capture it.
Subscribe to Digital Dispatch to understand where AI’s economics are heading, and where opportunity emerges when the gatekeepers lose their grip.
Click here to get started.
Keep coming back,
Chris Curl
Editor, Bizarro World