#How Did Export Controls Impact AI Model Competition?
The US government implemented export controls on Anthropic's advanced AI models, aiming to safeguard American technology leadership. However, this move inadvertently opened the door for Chinese AI companies to make significant inroads in the US market. Since June 2026, models like Mythos 5 and Fable 5 have faced restrictions, leading to a rapid rise in the competitiveness of Chinese alternatives.
Data indicates that usage of Chinese AI models among US firms has soared, now accounting for over 30% of AI token activity, which is a dramatic increase from just 11% prior to February 2026. This growth is not merely a trend; it represents a profound shift in preferences.
#Why Are Chinese Models Gaining Popularity?
The primary reason for the rise in usage of Chinese models like DeepSeek and Z.ai is their significant cost advantage. On average, these models are priced 60-90% lower than equivalent options from top US companies. For instance, select models are available for around 5% of what similar Anthropic products charge. This pricing strategy has made them an appealing option for businesses looking to reduce expenses.
Notably, DeepSeek has seen the highest adoption rates, but Z.ai's GLM-5.2 model experienced an unprecedented 27-fold increase in daily usage shortly after its launch. Companies are not just experimenting with these alternatives; businesses like Lindy have made the decision to switch their entire operations from Anthropic to DeepSeek due to substantial cost savings.
#What Do these Changes Mean for US AI Policy?
The export controls, initially intended to protect US AI advancements, have had the unintended effect of pushing companies toward more accessible and cost-effective Chinese models. The temporary suspension of access to key US models has instilled a sense of uncertainty among Chief Technology Officers (CTOs) in the US, making them wary of potential limitations on accessing these crucial resources.
In contrast, Chinese open-source models provide a level of flexibility that allows companies to download, self-host, and manage their AI infrastructure independently of outside permissions. This independence supports strategic planning and operational stability for businesses adopting these models.
#How Should Investors Respond to This Shift?
For investors, the implications are clear. Companies transitioning to Chinese AI models can potentially see significant improvements in their profit margins. A reduction of 60-90% in AI inference costs can lead to direct enhancements in profitability. Investors should closely monitor earnings reports from AI-dependent companies that have quietly switched to these more affordable providers.
The growing interest and competition from Chinese giants like Alibaba and ByteDance also raise questions about potential retaliatory export regulations from China. Should Beijing impose restrictions on AI model access, companies that have shifted to Chinese technology might find themselves facing the same risks they sought to avoid, potentially leading to future market instability.