Guojin Securities: Demand for AI computing power is expected to continue to be strong in 2026, optimistic about AI copper-clad plates/PCBs, core computing power hardware, semiconductor equipment, etc.

The Zhitong Finance App learned that Guojin Securities released a research report stating that the capital expenditure of the four major North American cloud vendors (Microsoft, Google, Meta, and Amazon) supports continued growth and has a positive outlook on future capital expenditure. The agency believes that the high capital expenditure of the four largest CSPs in North America is sustainable, and there is still room for improvement. Nvidia's AI server cabinets are expected to grow dramatically in 2026, and the number of tokens will explode. Guojin Securities has determined that the number of ASICs will explode, and continues to be optimistic about AI copper-clad plates/PCBs and core computing power hardware. In 2026, we are optimistic about AI copper-clad plates/PCBs and the industrial chain benefiting from core computing power hardware, semiconductor equipment, domestic computing power, storage price increases, and AI end-side hardware.

Guojin Securities's main views are as follows:

AI copper-clad plates/PCBs and core computing power hardware continue to rise in volume and price: Nvidia GPUs are growing rapidly, Google and Amazon ASICs are exploding, and demand for copper-clad boards/PCBs is strong. Some Nvidia Rubin PCBs will start using M9 materials in 2026, and the ongoing orthogonal backpanels will demand M9 materials. Google TPU products are also expected to use M9 materials, and other ASIC manufacturers are also expected to use M9. Guojin Securities has determined that demand for M9 materials will explode in the next three years; AI PCBs have two major technologies Iterative trends, that is, the introduction of orthogonal backboards (transfer of copper cable value to PCB) and use of CoWoP packaging technology (transfer of carrier board value to PCB), Guojin Securities believes that the strong trend of AI copper-clad plates/PCBs will continue. Other AI computing power hardware such as AI servers, optical modules, liquid cooling, and power supplies are expected to continue to increase in volume and price.

The storage production expansion cycle is compounded by autonomous and controllable acceleration. I am optimistic about the semiconductor equipment industry chain: semiconductor equipment is the cornerstone of the semiconductor industry chain. Storage expansion and autonomous control resonate, and there is broad scope for domestic replacement. The memory chip architecture has profoundly changed from 2D to 3D. With the introduction of 3D DRAM technology and the evolution of the number of NAND stacked layers to 5xx layers and above, the requirements for high depth to width ratio etching and advanced film deposition in the manufacturing process have increased exponentially, and related equipment manufacturers will benefit deeply from the dividends brought by the increase in process complexity. The domestic storage industry faces a more severe production capacity gap than the world, and demands for autonomy and control are urgent. Long-term storage and Changxin's expansion of production will directly drive the share of domestic equipment in the core process chain.

Domestic computing power has ushered in new opportunities for development: there is still room to increase capital expenditure of domestic cloud vendors. Tencent, Ali, and Baidu 25Q3 capital expenses were 13 billion yuan/31.5 billion yuan/3.4 billion yuan respectively, -24%/+80%/+107% year-on-year, respectively. Domestic cloud vendors are also simultaneously increasing long-term AI investment. In February of this year, Ali announced that it will invest 380 billion yuan over the next three years to build cloud and AI hardware infrastructure, and is optimistic about the direction of AI chips, memory chips, advanced wafer manufacturing, switching chips, and optical chips with domestic computing power.

The AI side is optimistic about Apple hardware, end-side AI innovation, and the AI/AR glasses industry chain: the global AI model call volume is experiencing rapid growth, and the industry has entered a period of large-scale application explosion. Apple's AI strategy is hardware-based, end-side priority, and strong privacy protection. The core competencies include language text, images, cross-application operation, and personal scenario understanding. Its AI is not a single function or large model, but a “personal intelligence system” deeply embedded in the operating system, chip, and application ecosystem to perform operations across apps based on personal scenarios. AI/AR glasses are expected to continue to achieve technological breakthroughs and sales growth, and are optimistic about core aspects such as SOC chips and optics.

Risk warning:

AI capital expenditure fell short of expectations, AI end-side applications fell short of expectations, rising memory chips affected electronic hardware sales, and storage and fab production expansion fell short of expectations.