Guojin Securities has issued a research report expressing a positive outlook on AI-related copper-clad laminate (CCL)/printed circuit boards (PCB), core computing hardware, semiconductor equipment, and the Apple supply chain.

The firm projects that the number of application-specific integrated circuits (ASICs) for companies like Alphabet (NASDAQ: GOOGL), Amazon, Meta, Open AI, and Microsoft will experience explosive growth in 2026-2027.

Currently, many AI-PCB companies are seeing strong orders, operating at full capacity with all products sold out, and are actively expanding production, with high-performance growth expected to continue.

Demand for AI-grade CCL is also robust. Due to slow capacity expansion for CCL overseas, leading mainland CCL manufacturers are poised to benefit significantly.

Recent earnings beats from multiple companies in the AI supply chain, strong demand for NVIDIA's (NASDAQ: NVDA) next-generation Vera Rubin platform, a sevenfold year-over-year increase in Google's AI token processing volume, explosive revenue growth at Anthropic, accelerated capacity expansion for Taiwan Semiconductor Manufacturing's (NYSE: TSM) advanced nodes used in AI chips, cloud service providers securing long-term storage capacity agreements, and rising GPU rental prices all point to very strong short- and medium-term demand for AI.

Key industry sentiment indicators show: consumer electronics (steadily improving), PCB (accelerating upward), semiconductor chips (steadily improving), semiconductor foundry/equipment/materials/components (steadily improving), displays (bottoming out), passive components (accelerating upward), and packaging & testing (steadily improving).

Guojin Securities' primary viewpoints are as follows:

GPU and ASIC procurement are in sync; focus on areas with potential for above-expectation performance. NVIDIA's Vera Rubin platform has entered a phase of high-volume procurement, with the CCL and PCB supply chain already receiving large orders.

Advanced Micro Devices' (NASDAQ: AMD) latest AI GPU, the MI450, will see high-volume shipments in the second half of the year, and the company plans to gradually increase shipments of its Helios AI racks during the same period.

AMD CEO Lisa Su stated that customer demand has exceeded the company's internal projections for 2027, with core customer forecasts surpassing initial plans and new clients negotiating large-scale deployments.

Google's TPU V8 and Amazon's Trainium3 are also expected to enter high-volume procurement phases.

The firm judges that the quarter-over-quarter performance of core AI computing hardware companies in Q2 and Q3 is expected to accelerate, suggesting attention on directions with potential for earnings surprises.

SK Hynix is aggressively expanding its DRAM capacity. SK Group Chairman Choi Tae-won indicated that memory demand will explode with growing KV cache needs, and the memory shortage is expected to persist until 2030.

SK Hynix aims to double its wafer capacity within five years, planning to increase monthly DRAM wafer capacity from the current 550,000 wafers (including about 200,000 from its Wuxi, China plant) to approximately 1 million by 2030.

This expansion will focus on the Yongin Semiconductor Cluster. The Yongin Phase 1 factory is divided into six cleanrooms, with equipment installation in the first cleanroom (Phase 1) starting in February 2027.

After equipment installation, capacity is planned to increase by 60,000 wafers, with an additional 60,000 wafers added every six months in the next cleanroom, gradually ramping up capacity.

By the first half of 2030, the Yongin Phase 1 factory alone is expected to add 360,000 wafers of monthly DRAM capacity.

Additionally, the Cheongju M15X wafer fab is currently under expansion. M15X will begin production in the second half of this year with a monthly capacity of 40,000 wafers, reaching about 80,000 wafers per month next year.

Combined with the 360,000 wafers from the Yongin fab and the additional 80,000 from the M15X expansion, SK Hynix's monthly DRAM wafer capacity is projected to reach approximately 1 million by 2030 or 2031.

Taiwan Semiconductor Manufacturing Chairman C.C. Wei stated at the shareholders' meeting on June 4th that AI-related demand will remain robust.

Benefiting from strong demand for advanced nodes, specialty technologies, and advanced packaging, and based on demand forecasts from customers and their clients, TSMC is very confident about its growth over the next several years.

The current capital expenditure estimate for 2026 leans towards $56 billion, reflecting strong market demand, with no current market signals indicating a need to reduce capital expenditure.

TSMC will continue investing in advanced process and packaging technologies to solidify its core position in the AI semiconductor supply chain.

As AI servers evolve towards higher computing power while simultaneously needing to reduce power consumption, requirements for power stability and signal integrity have increased significantly.

This drives significantly higher demand for MLCCs with high capacitance, high voltage resistance, and low loss, as well as high-end power inductors.

This not only leads to a multiple-fold increase in usage per unit but also pushes the average selling price (ASP) upward.

Taking NVIDIA's new Vera Rubin platform as an example, the value of passive components required per AI server has increased approximately 4 to 6 times compared to the previous generation platform.

The growth driver has shifted from volume expansion to specification upgrades and increased value per unit.

Following MLCC price increases, inductors are also poised for potential price hikes, particularly those used in AI server VRM modules and automotive DC-DC converters, which typically require a saturation current (Isat) above 40A.

Currently, lead times from major overseas suppliers have extended to over 20 weeks, suggesting attention on the benefiting supply chain.

Recent earnings beats from multiple AI supply chain companies, strong demand for NVIDIA's Vera Rubin platform, a sevenfold YoY increase in Google's AI token processing, explosive revenue growth at Anthropic, accelerated capacity expansion for TSMC's AI chip nodes, rising GPU rental prices, and material shortages and price increases all indicate very strong short- and medium-term AI demand.

Risk warnings include: risks of demand recovery falling short of expectations; risks of AIGC progress lagging expectations; and risks of further escalation in external sanctions.