 |
| Chamath Palihapitiya |
Venture capitalist Chamath Palihapitiya recently proposed a scenario on X where both the United States and China emerge as equals in the AI race.
Here is what Chamath said:
Thought exercise:
USA and China both win the AI race.
Our AIs are roughly equally capable: what we can build, they can also build. Whatever next gen weapons we design, they can also design.
This creates a detente where we each stick to our own hemispheres and leave each other alone.
Each of us rely on a few countries for resources, know how and capital. Those chosen countries exclusively work with US or China but not both.
Every other country in the world then becomes a vassal state of either the US or China. With no native AI or critical resources, they lobby to become a protectorate of one or the other.
Huge shift in the global order if this happens.
The AI rivalry is currently a tale of two different strengths. The United States leads in "frontier AI," dominating a large majority compute capacity through firms like OpenAI, Google, and Meta. American models like Claude 4.5 and Gemini remain the gold standard for complex reasoning and software engineering. However, China is chasing it via highly efficient open-source models like DeepSeek R-1, which promises to rival Western tech while remaining accessible and affordable for global markets.
While the US focuses on breakthroughs, China is going after the "diffusion" race by embedding AI into its massive manufacturing sector. Beijing’s "AI Plus" initiative leverages a workforce of 105 million manufacturing workers, dwarfing the 13 million found in the US. This allows China to deploy AI in the "physical economy" at a scale and speed that the US currently cannot match.
The Looming Energy and Chip Gap
Energy infrastructure has become a critical battleground because AI data centers are incredibly power-hungry. US electricity demand for data centers is projected to hit 426 terawatt-hours by 2030, threatening to overwhelm the national grid. In contrast, China produces more than twice the electricity of the US and could possess 400 gigawatts of spare capacity by 2030. This "electron gap" gives China a significant advantage in sustaining long-term AI growth.
On the hardware front, US export controls previously hobbled China's access to high-end chips. However, recent relaxations allowing modified Nvidia to ship H200 chips, combined with the rise of domestic alternatives like Huawei’s Ascend 910C, may narrow the hardware divide. Analysts suggest these shifts could accelerate China’s compute power by a few years.
Usage, Trust, and the "Diffusion Marathon"
Social adoption patterns suggest China may have a smoother path to total AI integration. According to the Edelman Trust Barometer, 72% of Chinese citizens trust AI, a figure much higher than in Western nations. This trust translates to active use: an estimated 150 million working-age Chinese use AI daily, compared to 58 million Americans. This massive user base acts as a productivity multiplier across the Chinese economy.
Experts describe this era not as a sprint to a single breakthrough, but as a "diffusion marathon." While the US holds the edge in elite talent and private capital, China’s 2.6:1 lead in AI users and its purported dominance on the factory floor provide a formidable counterweight. The result is a world where innovation and application are increasingly balanced between the two powers.
Market Impacts of a Divided World
If Palihapitiya’s vision holds true, the economic consequences will be massive, with trillions of dollars at stake. US tech giants like Microsoft and Nvidia saw record gains in 2025, but Chinese firms like Alibaba and Tencent are now seeing 20-30% gains as their open-source models gain traction. Global AI infrastructure spending is expected to reach $600 billion in 2026, split between Western cloud leaders and a booming Chinese data center market.
Geopolitically, this parity could lead to "bifurcated" supply chains—separate ecosystems for US-friendly and China-aligned nations. For investors, this requires a dual strategy, balancing exposure to American innovators with Chinese "scalers." As the US deregulates and builds out nuclear power (and maybe Fusion power in near future) to keep pace, the global order is shifting from a single-leader model to a tense, two-sided stability.