FLASH
Shkruar nga Pamfleti
A global battle that no longer takes place on the ground, but on servers, algorithms and open models
The release of China's new DeepSeek R1 artificial intelligence model has shaken the foundations of the global technology industry. Offered for free and without copyright, this "open-source" model has caused a financial earthquake in American stock markets, challenged US technological dominance and called into question the future of Silicon Valley's traditional "kept secret" model.
R1 has been considered the "Chinese answer" to OpenAI's GPT-4, and its release is directly linked to the loss of about $1 trillion in US technology markets, a clear sign of investor panic due to the transformation of AI into a free and available commodity for all.
Low cost, maximum efficiency
According to reports, the R1 was built for less than $6 million, using Nvidia's H800 chips, a cost far lower than the hundreds of millions that GPT-4 is alleged to have cost. This makes the Chinese model extraordinary in terms of economic efficiency and strategic impact.
Meanwhile, other companies like Alibaba with its “Qwen3” series and Mistral AI in France are following the same path, turning open-source into a new global paradigm. The US, if it does not adapt, risks being
Is the Chinese model the end of the American monopoly?
While OpenAI maintains secrecy and tries to justify its $500 billion “Stargate” project, the real innovation seems to be happening outside of America’s borders. The slow progress of projects like Google Gemini, Meta Llama, and Anthropic’s Claude has added to the frustration.
Ironically, the US has been stifling China's AI development for years through sanctions and chip bans. But this may have backfired: the restrictions have forced China to create its own, cheaper and more efficient alternatives.
Through its open strategy, China is following a similar logic to the success of Android and Google Play: development through community and global access. Its goal is clear, to dominate the market with speed and scale, until only one dominant model remains that can then be capitalized on through advertising and premium services.
Western experts are increasingly realizing that the fight for AI dominance will not be won through isolation, but through diffusion, adoption, and open innovation.
China's restrictions
However, China also has its own problems. The harsh censorship of the Internet raises serious questions about the reliability and adaptability of Chinese models in the global market. The example of the RedNote (Xiaohongshu) app clearly shows the clash between American users and the Chinese moderation system, especially on topics such as Taiwan and Xinjiang.
A silent revolution in progress
Despite the challenges, China has regained ground without American chips and without OpenAI, only through the strategy of “AI for all”. Even in the US, figures like Elon Musk with Grok-1 are moving towards more open models. Even OpenAI itself is considering this direction.
And here's the great paradox: America's newfound dominance in AI may come precisely as an unintended consequence of China's "socialist AI." /Adapted by "Pamphlet" from "Al Jazeera"
Lini një Përgjigje