The release of Elon Musk's Grok 3 AI chatbot has reignited discussions on the global AI race, but according to Tian Feng, former dean of Chinese AI software giant SenseTime's Intelligence Industry Research Institute, its significance lies not in its technical breakthroughs but in the contrasting development approaches of China and the U.S.
While the U.S. relies on massive funding and computing power to push AI forward, China is taking a more cost-effective and efficiency-driven route, as exemplified by DeepSeek.
Tian told CGTN in a phone interview that U.S. companies, backed by extensive capital and cutting-edge chips, continue to drive AI progress through sheer computational power. This approach, as seen with Grok 3, prioritizes innovation at any cost. The U.S., benefiting from its financial leadership, is less concerned with cost reduction and more focused on maintaining a technological lead over China.
In contrast, China follows a strategy of maximizing efficiency within existing resources, exemplified by DeepSeek's highly optimized training and inference processes. DeepSeek V3's training cost is just about 1/10 of OpenAI's GPT-4, and the expert predicts AI computing costs will drop by another order of magnitude within a year. This efficiency-first approach makes AI more accessible and accelerates its adoption as a practical tool across industries.
The Role of Open-Source in AI's Future
The expert believes that open-source development plays a crucial role in AI's evolution. Unlike Grok 3, which is not open-source, DeepSeek freely shares its research and even allows commercial use without licensing fees. This openness attracts global developers, creating a \"flywheel effect\" where broader participation leads to faster iterations, commercialization, and scientific breakthroughs.
According to Tian, China must build a robust AI community through open collaboration to establish a strong AI ecosystem. Open-source initiatives will allow Chinese researchers and developers to work together, fostering innovation at a larger scale.
AI+Manufacturing: China's Unique Advantage
Beyond AI models, Tian sees China's manufacturing sector as a major opportunity for AI integration. He points out that Tesla's gigafactories provide Elon Musk with rich real-world data, giving him an edge in AI development.
However, Tian believes China's industrial diversity is potentially 100 times greater than that of the U.S. and has a bigger scale.
Integrating AI into manufacturing will be a game-changer for China. As Tian puts it, \"No scale, no intelligence.\" AI's evolution, especially towards Artificial General Intelligence (AGI), depends on large-scale data and adoption. With its massive industrial base, China is well-positioned to lead AI-driven manufacturing advancements.
Two Diverging Paths for AI's Future
Looking ahead, Tian sees AI development following two primary paths:
1. High-cost, high-performance models – the U.S. will continue investing heavily in top-tier AI models for cutting-edge applications, leveraging its financial strength to widen the gap with competitors.
2. Affordable, mass-scale AI – China will focus on making AI more accessible and efficient, maximizing performance with existing hardware and optimizing costs.
As the global AI race intensifies, DeepSeek has already become a benchmark, with major U.S. and Chinese models now measured against it. Tian believes DeepSeek's efficiency and cost advantages position it to remain highly competitive in the evolving AI landscape.
Ultimately, the future of AI will be shaped by these diverging strategies – one driven by unlimited resources, the other by maximizing efficiency. The question is, which approach will prove more sustainable in the long run?
Reference(s):
cgtn.com
