Cambricon, the Chinese AI chip designer, disclosed operating data on October 23. Revenue for January through September 2026 reached RMB 9.526 billion, up 106.76 percent year on year, while net profit attributable to shareholders rose 131.49 percent to RMB 3.715 billion, exceeding its full-year 2025 figure by 80.38 percent. Net profit in the third quarter alone reached RMB 1.404 billion. Research and development spending grew 45.35 percent to RMB 1.225 billion over the same period.
In the same disclosure wave, memory maker Biwin Storage said DRAM and NAND Flash prices have risen markedly this year on AI infrastructure buildout plus data center demand, and it is expanding to AI edge devices, PC and enterprise storage, and smart vehicles. PCB supplier Shengyi Electronics expects nine-month net profit of RMB 1.869 billion to RMB 1.952 billion, up 68 percent to 75 percent, with full order books and high capacity utilization. Gartner forecasts global AI spending of USD 2.67 trillion in 2026, up 49.5 percent, with over half flowing to infrastructure.
For the domestic electronic components supply chain, AI compute demand is cascading along the path of chips, memory, PCBs and passive components. Memory plus high-end packaging capacity stays tight, while order books and lead times at power, analog and passive suppliers stretch in step. Quotation cycles are lengthening, and the window for passing higher costs through to end customers is narrowing. Supply visibility now weighs as much as price when system builders plan fourth-quarter deliveries, and buyers who secure allocations early hold the stronger position.