All three companies emphasized the need for improved government oversight and faster implementation of AI regulations, proposing a unified regulatory framework managed by the Department of Commerce to avoid fragmented state-level regulations.

DeepSeek's pricing strategy is particularly alarming, as its models are reportedly 20-40 times cheaper than similar US offerings, putting pressure on American companies to adapt their business models.

In response to these challenges, Anthropic has called for stricter US export controls on AI chips to prevent China from gaining a technological edge, while Google has advocated for balanced regulations to maintain US competitiveness.

In March 2025, several major US AI companies, including OpenAI and Anthropic, expressed concerns to the US government about the rising sophistication and competitive pricing of Chinese AI technologies, particularly models like DeepSeek R1 and Ernie X1.

OpenAI has raised alarms about DeepSeek R1, describing it as a state-subsidized and state-controlled model that poses potential risks to global AI development and security, similar to concerns surrounding Chinese telecommunications giant Huawei.

Concerns have been raised regarding DeepSeek's data-sharing requirements with the Chinese government, which could enhance state surveillance and compromise sensitive US systems.

These companies warned that America's leadership in artificial intelligence is diminishing, primarily due to advancements from Chinese models that are significantly cheaper and increasingly effective.

Baidu's Ernie X1 and Ernie 4.5 are also notable competitors, with Ernie X1 matching DeepSeek R1's performance at half the cost, and Ernie 4.5 outperforming OpenAI's GPT-4.5 on certain benchmarks.

Anthropic highlighted biosecurity risks associated with its model, Claude 3.7 Sonnet, which could potentially assist in developing biological weapons, contrasting this with safety measures in US AI models.

US companies stressed the importance of better government investment in AI infrastructure, including the establishment of 50 gigawatts of AI-dedicated power capacity by 2027 to meet anticipated energy demands.

Anthropic's emphasis on biosecurity issues associated with DeepSeek underscores the broader security and infrastructure challenges posed by the competitive landscape in AI.