China chip profits surge 2,579% as AI server shortages grip market

China's electronics industry posted a 96.9% profit surge in the first half, with integrated circuit manufacturing profits skyrocketing 2,579%, as AI computing demand triggered severe server and component shortages across the supply chain.

"Memory prices have tripled from last year — a 10,000 yuan module now costs 30,000 yuan and is still hard to find," Liu Shuqi, product manager at Inspur Computer, said at the Guanghe Organization 2026 Intelligent Computing Conference in Zhengzhou.

Computer manufacturing profits jumped 689.3% and peripheral equipment profits rose 305.8%, according to the National Bureau of Statistics. Overall, major industrial firms with annual revenue of at least 20 million yuan posted an 18.7% profit increase in the first half, with the electronics sector contributing 8.5 percentage points of that growth. Server prices surged from 200,000 yuan to 800,000 yuan within months, with contract price quotes valid for only one week, vendors said. On the same day the data was released, memory maker ChangXin Memory Technologies debuted on the STAR Market, closing 465.8% higher at a 3.28 trillion yuan market capitalization, surpassing Industrial and Commercial Bank of China as the largest A-share company.

The supply crunch reflects a structural imbalance between surging AI demand and constrained production capacity. While the profit boom benefits domestic chipmakers and server vendors, the question for investors is whether valuations have run ahead of fundamentals — CXMT's 3.28 trillion yuan market cap implies extreme pricing of future growth in a sector where supply constraints may persist for years.

Supply Shortages Reshape the Server Market

At the Zhengzhou conference, vendors described a market in turmoil. "We signed a deal to sell servers together, but the price changed every day — from 200,000 to 500,000 to 800,000 yuan, and we still couldn't deliver," a GPU vendor attending the conference said. Even basic components are scarce. "Even the circuit boards are in short supply," Liu said. The shortages, which began late last year, show no signs of easing.

The crisis is driving structural changes in server design. Zhang Panyong, general manager of the server products division at Hygon Information Technology, said the ratio of CPUs to GPUs in data center servers has shifted from 1-to-4 or 1-to-8 to roughly 1-to-1, as AI agents require more task orchestration, memory management, and tool processing on the CPU side. Li Bin, senior vice president at Dawning Information Industry, noted that storage components are playing an increasingly important role in long-context inference workloads. Liao Guangshan of Inspur's AI product division compared the complexity gap between consumer PCs and servers to "a toy car versus a rocket," noting that developing a single server system costs tens of millions of yuan.

The Path to Ubiquitous Computing

Despite the current supply squeeze, industry executives see a future where computing power becomes as accessible as electricity. "Computing power will eventually be like water and electricity — available on demand, whether through local deployment or cloud access," Ying Zhiwei, vice president at Hygon, said. Li Bin predicted that large language models will follow the same trajectory as facial recognition technology, where costs fell sharply as the technology matured and became widely accessible.

The shift toward edge computing is already underway. Liao said that while cloud-based inference handles most consumer AI tasks, local deployment is gaining traction for privacy-sensitive applications. A research institution fine-tuning a model like DeepSeek would need four to eight times the computing resources required for basic inference, while full training requires more than 20 times, he said. Yu Weining, an NBS statistician, cautioned that the external environment remains complex, with international commodity prices still subject to fluctuations and industrial enterprises facing challenges including insufficient market demand and pressure on capital turnover.

This article is for informational purposes only and does not constitute investment advice.