A DeepSeek engineer has criticised Anthropic and OpenAI over calls to slow or “pace” the development of advanced artificial intelligence, warning that concentrating frontier AI in a handful of proprietary US companies could deepen inequality and restrict access to the technology.
Liu Shengyu, a DeepSeek engineer involved in the company’s V4.1 models, said he did not trust Anthropic or OpenAI to make cutting-edge AI “open and affordable”, contrasting their closed-source approach with DeepSeek’s open-source strategy.
“I especially don’t want Anthropic to master the most advanced artificial intelligence or AGI,” Liu wrote on social media on Monday. He argued that such concentration could make social mobility increasingly difficult.
Liu went further by comparing the prospect of Anthropic controlling the most advanced AI with a scenario in which Adolf Hitler’s Germany had acquired nuclear weapons before the Allies. He said his belief in more inclusive and open-source AI was among the reasons he had chosen to remain at DeepSeek.
The remarks came after Anthropic CEO Dario Amodei called on frontier AI companies to “pace” development, warning that AI capabilities were advancing faster than safety measures.
OpenAI CEO Sam Altman and xAI founder Elon Musk have also backed calls for greater caution, although the companies differ on how AI development should be governed.
The debate has taken on a geopolitical dimension as Washington and Beijing compete for leadership in advanced AI. Amodei has warned that a Chinese lead in artificial intelligence could pose a serious threat to the United States and argued that Washington should maintain restrictions on exports of advanced chips and chipmaking equipment to China.
Chinese commentators have challenged that position. The Dongbuyaqiao Research Institute, an independent Chinese research platform, suggested Amodei’s intervention could help US companies gain greater influence over AI rules before international standards are established.
A commentary published by China’s state-run Global Times described the debate as a “quiet AI cold war”, arguing that proposals to slow development could help preserve America’s technological lead while limiting China’s role in global AI governance.
China’s foreign ministry also criticised what it described as confrontation and excessive competition in AI. Spokesman Guo Jiakun said on Monday that “fearmongering, confrontation and vicious competition” would hinder efforts to establish sound global AI governance.
US President Donald Trump has meanwhile rejected calls for an industry-wide slowdown. He has argued that restraining US AI development could give China an advantage, while dismissing warnings that artificial intelligence could take over the world or destroy humanity.
The dispute is unfolding as investors reassess the enormous valuations attached to AI companies and the pace of spending on data centres and advanced computing.
US technology stocks fell on Monday, with Nvidia down 3.4 per cent and Micron Technology losing more than 5 per cent. Broadcom and AMD each fell more than 4 per cent, while the Nasdaq Composite declined 0.56 per cent.
AI-related shares in Hong Kong also weakened on Tuesday. The Hang Seng AI Index fell 1.65 per cent after dropping 3.3 per cent on Monday, while Z.ai and MiniMax fell 5.69 per cent and 7.05 per cent respectively.
The clash between Liu and US AI executives reflects a broader dispute over who should control the development of the world’s most powerful AI systems, how quickly they should advance and whether access to frontier models should remain concentrated among a small number of companies.
Neither DeepSeek nor Anthropic immediately responded to requests for comment.
