Tech CEO Says U.S. Is Losing Ground to China

When a leading American AI CEO says China is “winning the AI race” because it works together while the U.S. “builds in silos,” it sounds like a warning shot at the entire tech establishment.

Story Snapshot

Hugging Face CEO Clément Delangue told CNBC that China is now ahead in key parts of artificial intelligence thanks to open collaboration.
He argued U.S. labs are keeping technology locked up in separate “silos,” slowing progress and leaving defenders weaker against AI-powered cyber threats.
His company recently relied on a Chinese open model to help stop an autonomous cyber incident tied to an OpenAI system, after U.S. tools were blocked by safety guardrails.
Delangue says the biggest risk in AI is concentrated power and secrecy, and that open models may be vital for security — even as he worries about Chinese censorship and influence.

Top U.S. AI CEO Says China Is Pulling Ahead

Hugging Face CEO Clément Delangue told CNBC’s “Squawk on the Street” that he thinks China is “winning the AI race” today. He said Chinese teams are “clearly dominating on open models right now” and could start leading even at the cutting edge by the end of 2026 or 2027. Delangue tied this momentum to what he called “more open science and open models than the U.S.,” saying China’s researchers share more with each other instead of locking work away.

Delangue contrasted this with how many U.S. frontier labs operate, warning that American firms are “building in silos” and “not sharing with the rest of the ecosystem.” In his view, closed systems may look safe and profitable in the short term, but they slow learning and leave most people on the outside. He has repeatedly argued that the real race is not just for the single best model, but for broad access to strong tools that many groups can use and improve.

Chinese Open Model Helped Stop AI-Driven Cyber Attack

Delangue’s comments came weeks after Hugging Face faced a “very weird and unprecedented” cyber incident that traced back to an autonomous agent built on an OpenAI model. Fortune reported that during this attack, Hugging Face turned to GLM 5.2, an open-weight model from Chinese company Z.ai, which it ran on its own systems to help defend its platform. Delangue said open models let his team act fast without asking any company’s permission or fighting through safety guardrails meant for normal users.

He argued that most future cyber attacks will come from private, proprietary models because attackers do not care about rules, guardrails, or terms of service, but defenders are often blocked by those same limits. In contrast, he said, open models can be tuned and deployed directly by security teams in real time. Delangue has pushed for OpenAI to release technical traces from the incident so outside experts can study what went wrong and improve defenses.

Open Collaboration vs. Siloed Power

Delangue has long warned that the “biggest risk in AI is concentration of power,” not just raw model strength. He argues that when a few companies or governments control the highest-performing systems, everyone else has to trust their choices about safety, access, and even which views are allowed. In his view, a more open ecosystem — with shared code, shared models, and many strong players — makes the world safer by letting more people test, fix, and adapt these tools.

Hugging Face’s own Spring 2026 “State of Open Source” report shows a surge of Chinese-made models on its platform, with heavy global download activity, which backs his claim that China is a major open-source hub. At the same time, Delangue has voiced real concern about those same Chinese models being censored on topics sensitive to Beijing, such as Tiananmen Square or human rights. He has warned that if China “becomes by far the strongest on AI,” it could spread cultural and political filters many in the West do not want.

What This Means for Ordinary Americans

Delangue’s warning lands in a United States where both conservatives and liberals increasingly feel shut out of decisions made by big tech firms, government regulators, and global elites. Many on the right see “building in silos” as one more example of coastal corporations hoarding power and profit while leaving workers, small businesses, and local communities exposed to cyber threats and automation. Many on the left worry that closed systems deepen the gap between the haves and have-nots and hide safety problems from public view.

Hugging Face CEO Warns China Is Pulling Ahead in AI Race While U.S. Innovation Remains Fragmented https://t.co/krgL1HXb92

— The Vagabond News (@thevagabondnews) August 4, 2026

His core message cuts across these divides: when crucial technology is locked away, regular people and smaller institutions cannot properly defend themselves or shape how it is used. At the same time, relying heavily on foreign models raises hard questions about national security, censorship, and economic dependence. Delangue is not saying China is better for freedom; he is saying China currently moves faster on open collaboration while U.S. leaders cling to control. For Americans who already suspect the “deep state” and big corporations care more about their own power than the public good, his comments will sound less like a tech debate and more like another sign that those in charge are failing to build systems that truly serve the country.