During the first-half 2026 earnings season for China’s A-share market, listed companies tied to AI computing infrastructure broadly reported strong growth, underscoring the momentum driven by demand for computing power. Wind data shows that as of August 31, 848 listed companies associated with AI and computing power had released their first-half results, with more than 60% reporting profits.

Li Xuan, a senior technology analyst at Guotai Haitong Securities, told Securities Daily that sectors including optical modules and storage are moving from a period of massive capital expenditure into a new phase in which order deliveries translate into rising volumes and profits. Economies of scale are materializing faster across core components and system manufacturing, with profit growth significantly outpacing revenue growth throughout the supply chain.

AI Servers and System Manufacturing

AI servers and system manufacturing, the core hardware underpinning computing power deployment, delivered a concentrated surge in earnings in the first half of the year, making the segment one of the most resilient links in the supply chain.

Against a backdrop of surging global demand for AI computing power, Foxconn Industrial Internet posted record first-half revenue and net profit. Its first-half 2026 report showed revenue of 557.861 billion yuan, up 54.63% year on year, while net profit attributable to shareholders rose 95.99% to 23.740 billion yuan. The company said revenue from AI servers sold to cloud service providers increased 2.3-fold, while shipments of GPU-based AI racks rose 3.2-fold.

Sharetronic Data Technology and Shenzhen Jwipc Technology reported first-half net profit attributable to shareholders growth of 325.51% and 281.92%, respectively. Sharetronic said it recognizes revenue from its intelligent computing business after computing clusters are delivered and accepted, with deliveries and billing for related clusters still underway. Jwipc said it is developing two AI computing nodes in Shenzhen and Tianjin.

“The most striking feature of first-half financial results from companies in semiconductor equipment, optical module equipment, PCBs and other parts of the computing power supply chain was that profit grew significantly faster than revenue,” Wang Haoyu, CEO of Mairui Asset Management, told Securities Daily. “This reflects the combined impact of economies of scale and the higher technological value added in these segments.”

As large models expand into inference workloads, gigawatt-scale data center construction, edge and device-side computing power, and liquid cooling technology will become key areas of investment for AI server companies.

Optical Modules, Storage and Computing Power Operations

The large-scale construction of computing clusters has benefited more than system manufacturers. It has also acted as an engine for profit growth across the broader electronics sector, including optical modules, storage chips, PCBs and computing power operations.

Among listed companies across the supply chain, Zhongji Innolight generated first-half revenue of 41.778 billion yuan, up 182.49% year on year, while net profit attributable to shareholders surged 241.70% to 13.651 billion yuan. The company said shipments of its 800G and 1.6T products grew rapidly and accounted for a steadily rising share of revenue.

Several storage companies also recorded rapid earnings growth. Shenzhen Longsys Electronics reported first-half net profit attributable to shareholders of about 1.058 billion yuan, up from roughly 15 million yuan a year earlier. Shenzhen Dapu Microelectronics swung from a loss of 354 million yuan to a profit of 1.334 billion yuan.

“Building computing clusters is no longer simply about stacking GPUs. It is becoming a complex systems-engineering undertaking that requires coordination across computing, connectivity and storage,” Li said. Computing clusters supply computing power, high-speed optical interconnects provide connectivity, and enterprise storage forms the data foundation. Together, these three layers have increased the value of each rack severalfold compared with the traditional model. As companies enter peak delivery periods, operational services are also beginning to contribute profits. The industry is shifting from construction-led growth to a model powered by both construction and operations.

Overall, listed companies in AI computing infrastructure are carrying high inventory levels. Inspur Electronic Information Industry held 52.679 billion yuan of inventory at the end of the first half, accounting for nearly half of its total assets. Contract liabilities reached 28.369 billion yuan, up 45.34% from the end of the previous year. The company attributed its sharp earnings growth to capitalizing on the industry upswing. Its YuanNao SD200 supernode server has achieved ultra-fast inference on large models with trillions of parameters. These figures reflect rising order volumes and delivery scale, but also point to mounting pressure on inventory turnover and delivery management.

In the first half of 2026, growth in China’s AI infrastructure supply chain expanded beyond complete server systems into high-speed optical interconnects, enterprise storage and computing power operations. But growth in orders, shipments and profits does not necessarily translate into stable cash flow. Inventory, delivery cycles, customer acceptance and capital turnover remain critical indicators of whether the current wave of AI infrastructure investment can be sustained.