US restrictions on closed AI models drive shift to open-source systems

US restrictions on closed AI models drove a shift to open-source alternatives. The combined OpenRouter share for Google, Anthropic, and OpenAI fell from 55% to 33%.

Categorized in: AI News Government
Published on: Jul 09, 2026
US restrictions on closed AI models drive shift to open-source systems

The US government's sudden move to restrict access to advanced AI models from Anthropic and OpenAI has triggered a surge in interest in open-source alternatives, particularly from China, creating new considerations for government agencies that rely on or regulate AI.

In early June, the Trump administration ordered Anthropic to block non-Americans from using its most powerful closed models, Mythos 5 and Fable 5. Faced with the complexity of screening users, the startup pulled the models offline entirely. Shortly after, OpenAI agreed to let the government approve every customer for its newest model, GPT-5.6.

Haitham Mengad, co-founder of Stems Labs, a startup focused on AI-powered music creation, said, "Fable has been a game-changing model for me. Honestly, when they took it off, it was the first time that I realized… it's almost like a drug."

The de facto bans blindsided the tech world and thrust a long-simmering debate to the fore: open versus closed AI. Most of the best-known AI models - like OpenAI's ChatGPT or Anthropic's Claude - are "closed," meaning the company keeps the underlying code and data locked away. Users access the AI via a subscription, but the company controls who gets in and can shut down access at any time.

Open-weight models work differently: the developers release the model's core files for anyone to download, modify and run on their own computers. Once released, no one - not the company, not a government - can take them back.

Open models gain ground

Open models were already gaining fans because using closed AI keeps getting more expensive. Around the same time as the US restrictions, China's Zhipu AI (also known as Z.ai) released GLM-5.2, an open model that performed nearly as well as top offerings from Anthropic and OpenAI on several benchmarks.

AI analyst Andrew Curran noted, "GLM-5.2 is free to download, fine-tune, and run on an enterprise's own servers, putting pricing pressure on frontier labs at the same time that access looks shaky."

On OpenRouter, a platform that routes requests across different AI models, Google, Anthropic and OpenAI's combined share of usage dropped from 55 percent to 33 percent between January and June. China's open DeepSeek now leads by a clear margin.

"You want to be as flexible as you can be. Maybe a year and a half ago some large company might say we bought Anthropic or we bought OpenAI, and now no one, no one buys only one," said Oren Michels, co-founder and CEO of Barndoor AI.

Security perceptions shift

Early suspicions about Chinese AI models as a security threat are fading, at least somewhat. Once an open model is downloaded and run on a user's own hardware, the company that made it - Chinese or otherwise - has no access to the data or control over how it is used.

"I don't think there's any risk, to be honest," Mengad said. "The fears are more psychological, emotional than rational."

Among Western companies, France's Mistral stands largely alone in championing open models. US tech giant Meta, once a vocal open-source advocate, has stepped back from that.

What governments might do next

The US crackdown could eventually extend to open models as they become more powerful. Ethan Mollick, a professor at the University of Pennsylvania and a leading voice on AI, said, "If Mythos-level models are considered risky, China will also not want them to be open." That suggests governments everywhere, not just Washington, may want to keep top-tier AI locked down.

Why this matters for government professionals

The sudden unavailability of frontier models shows how quickly critical AI tools can disappear when a single company or government changes policy. For agencies developing AI for Government strategies, the episode underscores the risk of vendor lock-in. Government procurement teams must weigh the reliability and support of closed models against the sovereignty and cost advantages of open-weight alternatives. Building flexibility into AI stacks - and monitoring how both Washington and Beijing regulate model access - has become a practical necessity, not a hypothetical exercise.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)