The release of Deepseek's R1, a powerful new open source AI model in China, has sent shockwaves through the global technology industry. It was offered free and royalty-free, disrupt financial markets, challenged the US domination in artificial intelligence, and piqued the fear that Silicon Valley's strict, protected business model could no longer be retained.
Deepseek's open source launch is widely seen as a key trigger behind the trillion-dollar technology sale in the US, indicating deep investor fears over AI and China's growing competitiveness. Openai's GPT ‑ 4 is called “Chinese answer,” and R1 is becoming unstable with investors, changing global AI geopolitics.
The report suggests that the computational cost of the R1 was less than $6 million using NVIDIA's H800 chip. The full development costs remain undisclosed, but this points to a significantly more cost-effective model than its own counterpart. This suggests that R1 may have been constructed for part of Openai's GPT‑4 cost. This cost-effectiveness, combined with open access, makes DeepSeek's model uniquely destructive.
Chinese companies like Alibaba are free to release the QWEN3 Embedding series, followed by French Mistral AI (along with Europe's first Reasoning LLM). The US is risking losing the ground unless it embraces an open source strategy. Ultimately, early internet giants like Google and Facebook used free, user-centric services (such as Gmail and Maps) to encourage adoption before monetization.
In areas where secrets are the norm and models are often firmly fixed, providing valuable tools seems counterintuitive. However, Openai looks cautious once a pioneer of GPT‑4. CEO Sam Altman defends the $500 million Stargate project designed to shut down AI leadership. However, the actual expansion beyond ChatGpt was slow, with only the new shopping feature being released. US competitors (Google Gemini, Meta Llama, Anthropic Claude) are not yet pushing for faster or cheaper innovation.
The first US domination increased with progressive profits, supported by slowing China's export curbs and exports of Nvidia chips and other technologies. However, Nvidia's Jensen Huang warned that these restrictions could backfire, inspired China's chip industry, and ultimately weakened US control.
Open sourcing has become a strategic workaround for China. Legal, scalable, global cooperation. It reflects how Android thrives through external developers. AI has been improved through iteration, and Chinese companies are now leveraging the open force ecosystem to improve and extend the model, just like the Google Play model, without having to take charge of all the costs.
Yann Lecun, Meta's lead AI scientist, explained that the rise of Deepseek was a victory for open source, not only for China to overtake the US. Still, geopolitical interests are clear. Free access is a monetization pass made from a unique model. If open source achieves parity, the commercial model loses leverage.
China's industrial power lies in speed and scale. By saturating the market with low-cost, capable models, we put pressure on our competitors until only the dominant widely adopted models remain worthwhile. It is monetized via ads, data, or premium add-ons, and is often featured by Google and Facebook.
US investors are keenly aware. The reported $1 trillion DIP following the release of Deepseek reflects systematic concerns. For China, open sourcing is another aspect of the national industrial strategy. Subsidize, control and assert merciful intentions through “AI for AI.”
Open sourcing is not risk-free. If US technology is freely available, global rivals, including Chinese companies, can reuse it and outperform it. The opposite could also be true.
China is also facing restrictions. Its strict internet censorship regime raises questions about how open source models trained in that environment can adapt to global content demand. This has already emerged in Rednote (Xiaohongshu), a Chinese social media app that has recently escaped from a potential tiktok ban. Although intercultural exchanges have been largely positive, tensions have emerged, particularly around moderation and censorship of content on politically sensitive topics such as Taiwan and Xinjiang.
These constraints could be at a disadvantage against China's AI model when competing for trust and relevance in international markets.
Nevertheless, open source AI has re-adjusted the world's AI situation by allowing China to compete without access to cutting-edge US chips. Even the US is beginning to realize that, from Elon Musk's Grok-1 to Openai's evolving attitude, long-term AI domination relies on not only its own control, but also adoption, accessibility, and massive innovation.
Ultimately, the path to AI supremacy in the US may not be to defend the model behind closed doors, but to embrace the principles of openness and decentralization that China is currently using to rebuild the world's arenas.
Ironically, the next leap in US technological superiority could be the (UN) intended consequence of China's so-called “socialist AI” approach.
The views expressed in this article are the authors themselves and do not necessarily reflect Al Jazeera's editorial stance.
