US Anthropic ban is best advert for Chinese AI

Bankers working for JPMorgan Chase and Goldman Sachs in Hong Kong must have been miffed when they were shut off from using artificial intelligence (AI) models from Anthropic, a pioneering American firm in the field.
Goldman Sachs and JPMorgan Chase pulled the plug in April and last week respectively, based on a strict interpretation of Anthropic’s terms of use, which reflect Washington’s stringent restrictions on China’s access to frontier American AI models.
The banks’ decisions are seen as a potential blow to the competitiveness of the city’s financial sector. In a report, the Financial Times warned that, “Preventing access to the world’s most advanced AI models represents a threat to Hong Kong’s revival as an international financial centre, given their rapid adoption in other parts of the world, particularly for coding.”
A US Commerce Department directive this month means all foreigners are to be denied access to Anthropic’s most advanced Fable 5 and Mythos 5 models, including Anthropic’s own foreign employees.
The export control directive gave Anthropic just 90 minutes to shut off access, citing national security concerns.
As events unfold, China’s AI industry may emerge as the real winner. If foreigners can’t use America’s most advanced AI models, most will just switch to the “good enough” ones developed by Chinese firms, which charge a fraction of the cost.

DeepSeek’s flagship V4 Pro model is reportedly 60 times cheaper than Anthropic’s Fable 5. It topped a global bang-for-the-buck ranking last month after slashing prices.
China’s MiniMax’s M2.7, Xiaomi’s MiMo V2.5 Pro and Alibaba’s Qwen3.7 Max were all ranked highly for efficiency and value by the American third-party benchmark firm Artificial Analysis.
These are not just “good enough” models, either. As Agathe Demarais, a senior fellow in geoeconomics at the European Council on Foreign Relations, wrote in the Financial Times: “The performance of Chinese models like DeepSeek’s V4 Pro, GLM, Kimi, MiniMax and Qwen is increasingly on a par with that of US AI.”
More importantly, Chinese AI models are open source, meaning clients can freely customise them to meet their own needs. Unlike with the United States, they don’t have to fear Chinese regulators shutting them down without prior warning on national security grounds.

“Demand for Chinese models has already overtaken that for US models on OpenRouter, a popular platform for accessing different AI models,” Fortune magazine’s Asia editor Nicholas Gordon reported this month. The four most-used models last week were DeepSeek, MiniMax, Tencent and Xiaomi.
“The Chinese open-source models have proved popular not just within China but also in many other developing countries around the globe, where they are seen as a good tradeoff between price and performance,” Gordon added.
The US ban on Fable and Mythos, he points out, “may also end up vindicating China’s broader move towards tech self-sufficiency”.
In contrast, the Europeans are behind in AI and deeply alarmed by the latest US move. Bruno Retailleau, a former French interior minister and 2027 presidential candidate, warned that relying on foreign technology means a nation “can be unplugged overnight”. He called on Europe to stop being naive and to “rearm our technological power” by heavily supporting native champions.

Alistair Carns, a British member of parliament and former armed forces minister, said the US restrictions immediately affected British researchers, companies and hospitals that had been piloting the AI models. He framed the sudden cut-off as an example of Europe forfeiting its technological leadership.
So, China isn’t hurting; rather, it’s Anthropic and its clients, perhaps even the AI industry as a whole in the US. The firm is seeking a mega IPO, and the export directive couldn’t have come at a worse time. Its CEO, Dario Amodei, might have been a tad too convincing about how powerful his AI models are, which can reportedly penetrate and outsmart most computer systems. He portrays himself as the lone conscience of the AI industry, demanding its ethical usage and warning the US government to prevent authoritarian states from getting hold of the models.
Well, his fearmongering might have been too effective. This time, Washington really listened.
“Telling the authorities that you’ve built a weapon and then being shocked when they treat it like one is, all things considered … bad corporate management,” Patrick Boyle, a former financial trader and popular YouTube commentator, wryly observed about Amodei.
I guess Hong Kong bankers don’t have to worry too much about lagging behind now.