Eric Schmidt, the former CEO of Google, called OpenAI and Anthropic “modern colonial trading companies” during a recent technology conference appearance. He said they take data from developing countries, process it into AI models, and sell those models back to the same countries. The comment is a sharp criticism of the economic structure of the AI industry.
His argument is simple: the leading AI labs operate like extractive empires. They gather raw materials from one part of the world, build a valuable product, and then charge the people who supplied the raw material for access to that product.
What Schmidt Is Claiming
The colonial trading companies of past centuries followed a clear pattern. They took resources from one place, processed them elsewhere, and sold the finished goods back at a profit. Schmidt says OpenAI and Anthropic do the same thing with data.
“They are modern colonial trading companies.”
The data comes from people around the world. The training happens inside powerful companies. The finished models are then sold back to the same people who supplied the data. This one-way flow is what makes the arrangement colonial.
The Business Model in Action
- Data collection is global. The training data for large AI models includes contributions from users all over the world, especially in lower-income countries.
- Model development is centralized. The real value is created inside a small number of companies in wealthy nations.
- Access is sold back. Users in data-producing countries must pay for the services built on their information.
- Profits stay concentrated. The companies retain most of the financial gain, as well as the power to set the terms.
Schmidt’s point is that this structure creates a permanent imbalance. Countries that provide the raw data do not gain ownership of the final model. They remain suppliers, not partners.
A New Kind of Dependency
For many countries, the colonial comparison is hard to ignore. They do not control the models. They do not set the rules. They are not included in the decision-making process around AI. They are, in effect, on the outside looking in.
That dependency has real consequences. Without access to open models, countries that want to use AI must accept the terms of foreign companies. They cannot easily build local alternatives because the most advanced capabilities are kept behind corporate walls.
Why Open Source Matters
Schmidt reportedly points to open source as a way out. If AI models are open, the underlying technology is not controlled by one company. Developers anywhere can inspect the code, fine-tune the model, and run it with less external control.
Open source changes the relationship. It turns a country from a passive consumer into an active participant. It also prevents the colonial dynamic from becoming permanent by letting local talent build on shared foundations.
That matters because it changes who holds the power to innovate. Open models allow people to train on their own data, host AI on their own infrastructure, and offer services without asking permission from a foreign company.
What Schmidt’s Warning Means
The key takeaway is that AI’s biggest risks include economic power, not just safety. If the leading AI companies operate like colonial trading companies, countries need to be more careful about how they share data and license AI tools.
Schmidt’s words are also a reminder that open source is not just an engineering preference. It is a political position. It shifts the balance of power away from a few centralized AI companies and toward communities that are currently supplying data without gaining control.
The debate is no longer hypothetical. The AI industry is already building the global structures Schmidt warns about. Whether it can avoid the colonial outcome depends on the choices made by companies, governments, and users.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.