A leading Chinese investor believes the global race for artificial intelligence is increasingly being shaped not by algorithms alone but by the infrastructure that supports them. Allen Zhu Xiaohu, managing director at GSR Ventures and one of China’s most prominent early stage tech investors, said during a recent podcast that China is positioned to surpass the United States within the next decade. His reasoning points to a fundamental but often overlooked advantage: China’s rapid expansion of data centres and robust electricity supply, both of which are essential for training and operating large AI systems.
Zhu’s comments reflect a growing recognition in the tech world that AI development now depends heavily on physical capacity. While breakthroughs in models and algorithms attract public attention, the ability to store, process and power enormous volumes of data has become just as important.
Infrastructure becoming the real battleground
According to Zhu, it is far easier for China to catch up in the development of AI models than it is for the United States to expand its data centre and power infrastructure at the same pace. China’s national focus on high density computing facilities has accelerated construction across multiple provinces, supported by government incentives and local partnerships. These facilities provide the computing muscle required to train large language models, run complex simulations and support real time AI applications across industries.
By contrast, Zhu argues that the United States faces greater challenges in scaling its infrastructure due to regulatory hurdles, higher construction costs and an energy system that is already under pressure. As demand for AI computing grows, bottlenecks in power supply and grid capacity are becoming more visible, raising concerns about whether the US can continue to expand rapidly enough to meet future needs.
China’s aggressive buildout of power and computing facilities
China’s edge, Zhu says, lies in its coordinated approach to building large scale infrastructure. New data centres have been rising in regions with abundant renewable energy, such as wind rich northern provinces and hydro powered southwestern regions. This strategy not only increases computing capacity but also helps reduce the environmental impact of energy intensive AI operations.
The country’s ability to construct massive facilities swiftly gives it a structural advantage. With energy supply closely tied to state planning, developers can secure the electricity needed to operate high performance computing clusters without facing the lengthy approval processes common in other markets. This has created a foundation for rapid scaling, allowing China to support the growing appetite for AI across tech, finance, manufacturing and public services.
A veteran investor’s perspective on AI’s future
Zhu, often described as a unicorn hunter for his track record of backing billion dollar start ups such as Didi Chuxing and the social platform RedNote, has watched China’s tech ecosystem evolve over decades. His assessment carries weight because he has witnessed multiple waves of innovation and understands the structural factors that shape long term success.
In the podcast, he noted that while talent and research remain crucial, the next leap in AI will depend on nations’ ability to build the computing backbone required to support increasingly powerful models. He emphasized that countries investing heavily in stable power systems and scalable data infrastructure will hold the advantage in the coming decade.
What this means for global AI competition
Zhu’s remarks highlight a broader shift in how experts view the global AI contest. Instead of focusing solely on breakthroughs in algorithms or the number of start ups in each country, attention is moving toward the physical capabilities that enable those breakthroughs. Nations able to deliver cheap and abundant energy, paired with large scale data facilities, will be better positioned to lead.
As China continues to expand its computing footprint, the global balance in AI development may shift. The United States still leads in research and private sector innovation, but infrastructure constraints could slow its progress. For China, the combination of state driven planning, rapid construction capacity and growing private investment may create an environment where large scale AI becomes more accessible.
Zhu’s analysis underscores a simple but powerful insight: the future of AI belongs to those who can power it.
