DeepSeek Pushes Forward in Global AI Race

Chinese AI company DeepSeek has released a preview of its highly anticipated V4 model, claiming it can compete with leading closed-source systems from major US tech firms including Anthropic, Google, and OpenAI.

DeepSeek says V4 marks a major improvement over prior models, especially in coding capabilities—a key feature that has become central to AI agents and helped drive the success of tools like ChatGPT Code Interpreter and Claude Code. The release represents another significant milestone in China's AI ambitions and its broader tech competitiveness.

A Year of Disruption

The V4 announcement comes exactly one year after DeepSeek shocked the US AI industry with the release of R1, a model it claimed was trained at a fraction of the cost of leading American systems. This efficiency narrative raised eyebrows across Silicon Valley and prompted serious conversations about the true cost barriers in frontier AI development.

Domestic Hardware Focus

The release is also notable for explicitly highlighting compatibility with Huawei technology, marking DeepSeek's commitment to developing AI systems that can run on domestic Chinese hardware. This addresses a critical vulnerability in China's tech supply chain, particularly following US restrictions on advanced chip exports.

Unresolved Questions

DeepSeek has not disclosed V4's training costs or what hardware it was trained on. Meanwhile, US officials continue to accuse the company of using banned Nvidia chips, and Anthropic has claimed that DeepSeek misused Claude to improve its own products. These allegations remain contested.

What's clear is that the global AI competition is intensifying. Whether through cost efficiency, domestic chip integration, or raw capability, the race for AI dominance is reshaping tech geopolitics in real time.

Source: The Verge