China Catches Up to the United States in AI Race

A new report by Stanford University reveals the growing competitiveness in the AI sector, with China rapidly closing in on the traditional leaders. The study highlights the growing presence of new global players and improvements in AI models in various regions.

"Chinese models are catching up in performance to the United States, but new players are emerging in the space around the world," says Vanessa Parli, Research Director, Stanford's Human-Centered AI Institute (HAI), to Wired. 

In recent years, the AI market has been dominated by US companies such as OpenAI and Google, both of which pioneered advanced AI models. However, Stanford University's AI Index Report 2025 highlights that the competition has evolved, with companies from other parts of the world, especially China, making significant inroads in the field. China's rapid rise has been reflected in the DeepSeek-R1 model, developed by Chinese company DeepSeek, which has achieved performance comparable to the most advanced models from OpenAI and Google.

The report adds that AI is no longer just a competition between companies in the United States and China, but has become a global race, with players in other regions such as the European Union, the Middle East, Latin America, and Southeast Asia starting to produce relevant AI models. This shift in market dynamics has profound implications for the future of the technology, both in terms of innovation and global competitiveness.

The field is also being transformed by the emergence of "open-weight" AI models, which can be freely downloaded and modified, which has increased accessibility and competition in the market, reads the report. Meta, with its Llama model, has been one of the main drivers of this trend, followed by other companies such as DeepSeek and Mistral. OpenAI, meanwhile, has also announced plans to launch open-source models in 2024.

The Stanford report also highlights advances in the efficiency of AI hardware, which has improved by 40% in the last year. This increase in efficiency has reduced the costs of running advanced models, allowing AI models to run even on personal devices, opening up new possibilities for widespread adoption.

In addition to technical advances, the report notes that the AI sector is experiencing a significant increase in demand for skilled machine learning professionals. This increase in labor demand, coupled with growing private investment in AI, which reached a record US$150.8 billion in 2024, reflects the growing importance of this technology in the global economy. Governments are also increasing their investment in AI, with a significant increase in AI-related legislation from 2022.

However, the report warns of the growing rate of incidents related to the misuse or misbehavior of AI models, which has prompted research to improve the safety and reliability of these systems. In addition, although some AI models already surpass human capabilities in specific tests, such as image classification and language understanding, the report notes that the goal of developing general AI that surpasses human capabilities remains a long-term challenge.