Bank of China International: Computing power infrastructure drives AI “from 0 to 1” main line “cloud resonance” dominates storage and terminal innovation opportunities

The Zhitong Finance App learned that Bank of China International released a research report saying that looking ahead to 2026, investment in computing power infrastructure is expected to become the main logical line driving AI growth, and investment opportunities are expected to emerge “from 0 to 1” in the industrial chain. At the same time, storage price increases may continue throughout the 2026 timeline, and technologies such as 4F2+CBA will become the direction of development; consumer electronics cost pressure and innovation opportunities coexist, and innovative end-side AI products face a trend where “danger” and “opportunity” coexist.

Bank of China International's main views are as follows:

Computing power infrastructure drives the main line of AI upgrades, and the industry focuses on the “0→1” transformation node

The rapid growth in demand for tokens on the big model inference side pushes CSP vendors to actively increase capital expenditure and invest in computing power infrastructure construction. While pursuing higher computing power density, faster interconnection speed, and lower interconnection losses, AIInfra places higher demands on PCBs, electronic cloth, copper foil, resin, and optical interconnection architectures. The bank anticipates that Nvidia Rubin and RubinUltra's midplane/orthogonal backpanels will be equipped with an M9 grade PCB/CCL solution with high-frequency high-speed resin+HVLP copper foil+Q cloth, while the optical interconnection level will also evolve towards the trend of OCS/CPO. Domestic AI chips are also actively improving computing power and interconnection cluster performance, while advanced advanced manufacturing processes and advanced packaging are also actively cooperating to expand production.

Bit output growth was limited or caused storage price increases throughout 2026. Storage vendors switched to technology such as process upgrades, high-rise digital stacks, and 4F2+ CBA

According to Frost & Sullivan's estimates, the global storage product market will grow from US$263.3 billion to US$407.1 billion from 2025 to 2029, with a CAGR of about 11.5%. Among them, cloud-side and end-side storage are the two fastest growing segments. According to TrendForce estimates, from 2025 to 2026, the DRAM industry's capital expenditure will increase from US$53.7 billion to US$61.3 billion, YoY +14%; Nand industry capital expenditure will increase from US$21.1 billion to US$22.2 billion, YoY +5%. Storage bit output growth was limited in 2026, and the upward trend in storage prices may continue throughout 2026. At the same time, storage vendors are shifting from pure capacity expansion to new technologies such as process upgrades and high-rise digital stacks. Chinese domestic storage vendors are also actively developing 4F2+CBA technical architectures to meet the technical competition of leading global manufacturers. The structural changes of 4F2+CBA are expected to bring incremental changes to the supply chain.

Cost pressures and innovation opportunities coexist, and consumer electronics focus on structural opportunities

TrendForce expects that the increase in storage prices in 2025 and 2026 will drive the overall cost of smartphones to rise 8-10%/5-7%. Higher storage prices have led to increased consumer electronics BOM cost pressure, and cost conduction price increases may affect terminal product sales. At the same time, structured investment opportunities still exist. On the one hand, Apple's pricing strategy shows a trend of friendliness, and products such as affordable Macbooks, folding machines, and smart glasses are expected to boost the growth momentum of the fruit chain. On the other hand, new end-side AI hardware features such as Doubao AI phones, Quark AI glasses, and Apple Intelligence are gradually being implemented, which is expected to bring innovative experiences to users.

Key risks

Overheating demand for AI has sparked an industry bubble. AI technology is progressing faster than expected. Long-term supply-side overcapacity has led to a decline in prices. Technological changes have led to the elimination of existing products.