Citigroup raises its expectations for the Chinese AI-PCB market, stating that Google TPU demand may surpass Nvidia by 2028 BlockBeats News: On June 23rd, Citigroup announced that AI servers and high-speed network upgrades are reshaping the growth prospects of China's PCB industry, and has raised the target prices of Shanghai Electric Power Co., Ltd. and Shenghong Technology, citing that AI-PCB demand may continue to expand rapidly in the next three years. In a report released on June 21st, Citigroup raised its forecast for the total potential market size of AI-PCB from 2026 to 2028 to RMB 152 billion, RMB 307 billion, and RMB 562 billion, corresponding to year-on-year growth of 86%, 102%, and 83%. The bank stated that the new model incorporates CPU and optical module PCB requirements and provides the first forecast for 2028. Citigroup predicts that AI-PCB demand will be dominated by ASIC in 2027, accounting for approximately 34%; NVIDIA GPU related requirements account for 24%, CPU accounts for 16%, switches account for 14%, and optical modules account for 12%. In terms of growth rate, the demand for network related products is the strongest, especially for 1.6T optical modules and switches, which may drive a 135% and 178% increase in optical module PCBs in 2026 and 2027, respectively. The report states that Google TPU related PCB procurement may become a key increment in 2028. Citigroup predicts that Google's TPU PCB procurement may reach $16 billion by 2028, exceeding Nvidia's GPU related PCB demand. Divided by the new demand from 2027 to 2028, Google TPU contributes about 30%, Nvidia contributes 25%, and switches and optical modules each contribute 14%. With the rapid expansion of AI-PCB demand, production capacity may become a new constraint. Citigroup expects that the next round of PCB expansion announcements may begin in the second half of 2026 to support demand in 2028. The bank said that it usually takes 13 to 15 months for PCB factories in Chinese Mainland to go from greenfield projects to mass production, and the production expansion cycle of upstream fiberglass and copper clad laminate is longer, which may reach 18 months. This means that the supply of AI-CCL may further tighten. Citigroup believes that with the launch of major AI-PCB projects in the second half of 2026, frontline PCB manufacturers with the ability to lock in raw materials will have increased profitability visibility in 2027. The related expansion announcement may also become an industry catalyst, strengthening investors' confidence in the long-term growth of AI-PCB. In terms of individual stocks, Citigroup maintains a "buy" rating for Shanghai Electric Power Co., Ltd. and Shenghong Technology. The bank raised the target price of Shanghai Electric Power Co., Ltd. from 119 yuan to 189 yuan, citing the company's strong mass production and delivery capabilities in the fields of data center switches and AI server PCBs. Citigroup predicts that Shanghai Electric Power Co., Ltd.'s net profits from 2026 to 2028 will be 6.7 billion yuan, 12.1 billion yuan, and 23.2 billion yuan respectively, and believes that its compound annual growth rate of profits can reach 86%. Shenghong Technology's target price has been raised from 415 yuan to 456 yuan. Citigroup has lowered its net profit forecast for 2026, mainly reflecting that Rubin's related revenue contribution is lower than previously expected, but believes that the profit in 2027 will be basically stable, and has introduced a net profit forecast of 23.2 billion yuan for 2028 for the first time. The bank stated that the long-term relationship between Shenghong Technology and Nvidia, HDI production capacity, and potential opportunities for data center switches and ASICs still support its valuation premium. However, Citigroup also pointed out that the AI-PCB market still faces multiple risks, including lower than expected GenAI related order shares, yield issues, price competition in the automotive supply chain, capital expenditure cuts by cloud vendors, weak macro demand, rising material costs, and geopolitical risks between China and the United States. Overall, Citigroup's assessment is that AI-PCB demand is no longer limited to Nvidia GPUs as a single mainline, but is expanding to ASICs, Google TPUs, high-speed switches, and optical modules. As demand enters the capacity planning period ahead of schedule in 2028, China's PCB leader with high-end processes, customer resources, and material assurance capabilities may continue to be the core beneficiary of AI infrastructure expansion.