In late July 2026, the global AI computing power market heated up once again. NVIDIA's latest quarterly earnings showed its AI chip revenue doubled year-over-year, far exceeding market expectations and driving tech stocks higher worldwide. Meanwhile, several cloud service providers and AI startups announced billion-dollar data center construction plans, pushing AI infrastructure investment into an "arms race" phase.
Computing Power Demand Explodes, Industry Chain Thrives
As large models evolve from text generation to multimodal and complex reasoning, computing power demand is growing exponentially. According to industry research firm IDC, the global AI computing power market is expected to surpass $1.2 trillion in 2026, up more than 40% year-over-year. Behind this figure lie the rapid deployment of applications such as autonomous driving, medical imaging, and industrial intelligence, as well as sustained corporate investment in AI-native applications.
Against this backdrop, chip makers like NVIDIA and AMD have seen orders booked into 2027. NVIDIA's latest-generation Blackwell architecture GPUs are shipping in far greater volumes than the previous generation, while AMD's MI400 series has also been widely adopted by cloud vendors. Domestically, Chinese computing power chips such as Huawei Ascend and Cambricon are accelerating their iterations and gradually expanding market share with policy support.
Capital Markets Vote with Their Feet, AI Computing Power Becomes the Strongest Theme
In the secondary market, the AI computing power sector has become one of the biggest investment themes of 2026. As of end-July, the Nasdaq technology segment had risen 28% year-to-date, with AI chips, optical modules, and liquid cooling temperature control leading the gains. In the A-share market, computing power concept stocks have also performed actively, with many hitting record highs.
In terms of capital flows, global hedge funds and asset managers have been increasing their positions in AI computing power-related assets. A recent Goldman Sachs report noted that AI computing power is one of the few tracks with extremely high earnings visibility, projecting that related companies could achieve a compound earnings growth rate of over 50% in the next two years. Meanwhile, many brokerage analysts believe computing power investment is spreading from cloud vendors to telecom operators, local governments, and other diversified players, creating a "demand resonance" effect.
Industry Insight: Computing Power Is More Than Technology; It Is New Infrastructure
"AI computing power is becoming infrastructure similar to electricity and transportation," said Wang Ming, a technology analyst at TF Securities. "Just like the electricity revolution of the last century, computing power will reshape the cost structure of every industry. We are bullish on computing power service providers and hardware manufacturers with independent and controllable capabilities."
However, some experts have flagged risks. Li Tong, chief strategy analyst at China International Capital Corporation, pointed out that the current valuation of the computing power sector is relatively high, with some stocks having priced in two years of future earnings. She advised investors to pay attention to new business models such as computing power leasing and edge computing, as well as the actual deployment results of AI in vertical industries, rather than merely chasing hardware concepts.
Outlook: Investment Opportunities in the Intelligent New Infrastructure Wave
On the policy front, China's "computing power network" construction is accelerating. According to the "National Integrated Computing Power Network Construction Action Plan" released by the National Development and Reform Commission in July, the country will add 20 large-scale intelligent computing centers and double total computing capacity by 2028. The EU and the United States have also introduced computing power support plans, intensifying global competition in this field.
For ordinary investors, several key themes are worth attention: first, AI chip and server manufacturers, including NVIDIA's supply chain and domestic GPU makers; second, data center infrastructure such as optical modules, power supplies, and cooling equipment; third, computing power operators and cloud service providers, especially those capable of training large models.
It should be noted that AI technology evolves extremely rapidly. Although computing power demand is certain in the long run, short-term fluctuations are inevitable. Investors should remain rational and choose suitable products for positioning based on their own risk tolerance.
