According to a recent research report from Wall Street financial giant Citigroup, against the macro backdrop of a global wave of AI computing infrastructure construction and a "memory chip super cycle," semiconductor equipment manufacturers—especially the three major giants ASML Holding NV (ASML.US), Lam Research (LRCX.US), and Applied Materials (AMAT.US)—are poised to be the largest beneficiaries of the massive expansion in AI chip (AI GPU/AI ASIC) and DRAM/NAND memory chip capacity. Citigroup's report predicts the global semiconductor equipment sector will enter a "Phase 2 bull market cycle," suggesting a new upward trajectory following the super bull market of 2024-25.
An analyst team at Citigroup, led by renowned Wall Street semiconductor analyst Atif Malik, forecasts that as demand for AI chips and memory chips continues to surge dramatically, the world's three largest chipmakers—Taiwan Semiconductor Manufacturing (TSM), Samsung Electronics, and Intel—along with memory chip maker SK Hynix, are expected to significantly raise their capital expenditure (capex) guidance for 2026 and beyond in upcoming earnings disclosures. Consequently, they predict global wafer fab equipment (WFE) spending in 2026 is more likely to align with their "most optimistic forecast scenario."
Regarding stock performance, the U.S. semiconductor equipment sector has shown remarkable strength since the start of the year. The U.S. ADR price of ASML Holding NV has already hit a record high in early 2026, with a single-day surge of over 8% on January 2nd and a year-to-date gain of approximately 16%, bringing its market capitalization close to $500 billion. The share price of Lam Research in the U.S. market has repeatedly set new historical highs since the second half of 2025, boasting a year-to-date increase of about 20% in 2026. Similarly, Applied Materials' stock price has frequently reached new highs at the start of 2026, with a year-to-date rise of around 15%.
The Citigroup analyst team specifically highlighted that Micron Technology (MU), the primary memory chip competitor to SK Hynix and Samsung Electronics, had already raised its fiscal 2026 (ending August 2026) capital expenditure guidance during its December 2025 earnings call from a previous $18 billion to $20 billion, implying a substantial 45% year-over-year increase, with capital spending for chip fabrication plant construction nearly doubling. Micron also indicated that capital expenditures would continue to grow in fiscal 2027. As the most direct competitor to Samsung and SK Hynix in the memory market, Micron's significant capacity expansion moves may compel these two South Korean memory chip giants to undertake corresponding capital expenditure expansions to maintain their market positions.
Notably, strong market expectations for TSMC's capacity expansion are not only focused on the enormous data center AI chip orders from leading AI chip developers like NVIDIA, AMD, and Broadcom, plus the substantial annual consumer electronics chip orders from Apple, but also on the data center enterprise-grade high-performance SSD segment (falling under NAND end applications). High-performance NVMe (especially PCIe Gen5/Gen6) SSD controller chips are critically dependent on TSMC's advanced process capacity—this implies that TSMC's current capacity is inevitably far from sufficient to meet the "endless orders" driven by AI computing and storage demands, making large-scale capacity expansion extremely urgent.
Following Google's major launch of the Gemini 3 AI application ecosystem in late November, this cutting-edge AI software rapidly gained global popularity, instantly driving a surge in Google's AI computing needs. The Gemini 3 series products immediately generated an immense volume of AI token processing upon release, forcing Google to significantly reduce free access to Gemini 3 Pro and Nano Banana Pro and impose temporary restrictions even on Pro subscribers. Coupled with recent South Korean trade export data showing persistently strong demand for SK Hynix and Samsung Electronics' HBM memory systems and enterprise SSDs, this further validates Wall Street's assertion that the "AI boom is still in the early construction phase where computing infrastructure supply cannot meet demand."
As the global hyperscale AI data center construction drive, led by tech giants like Microsoft, Google, and Meta, intensifies, it comprehensively accelerates capacity expansion for advanced 3nm and below process AI chips, CoWoS/3D advanced packaging capacity, and DRAM/NAND memory chip capacity among chipmaking giants. The long-term bull thesis for the semiconductor equipment sector appears increasingly robust. In the view of Wall Street giants like Morgan Stanley, Citigroup, Loop Capital, and Wedbush, the global AI infrastructure investment wave centered on AI computing hardware is far from over; it is merely at the beginning. Driven by an unprecedented "storm of AI inference-side computing demand," this overall AI infrastructure investment cycle lasting until 2030 could reach a scale of $3 to $4 trillion.
2026: The Wave of Capacity Expansion from Chip Manufacturing Giants is Coming Citigroup's report indicates that the primary investment strategy for chip stocks in 2026 is definitely not a "broad bullish view on semiconductors," but rather a clear focus on the leading players in the semiconductor equipment segment within the equity market (i.e., WFE-related). Citigroup's semiconductor investment strategy locks onto the value transmission chain from "surging capex by major chip manufacturers" to "expansion of the total WFE market size," and then to "order/revenue/profit expansion for semiconductor equipment leaders," betting on a continued upturn in semiconductor equipment prosperity during 2026.
Citigroup's latest calculations project the 2026 global WFE market size model at approximately $115 billion (implying a year-over-year increase of +10%, significantly higher than the average growth rate of the past decade) and note that the three major chip manufacturers (TSMC, Samsung, and Intel) collectively account for about 59% of this. They also emphasize that observed growth trends suggest overall 2026 WFE will "move closer" to their bullish forecast scenario of $126 billion.
Particularly for TSMC, the global foundry leader, Citigroup expects TSMC's 2026 capex guidance range to be around $46–48 billion and judges a significant upward revision for the full year is possible (with communication to institutional investors suggesting it "could reach $50 billion"). In its October earnings call last year, TSMC had already raised the lower end of its 2025 capital expenditure range from the previous $38-42 billion to $40-42 billion.
In Citigroup's view, the "computing-storage-advanced chip manufacturing" chain driven by the AI infrastructure frenzy dictates that the stickiness of semiconductor equipment capex is stronger than in any previous cycle: massive computing demands based on AI training/inference not only boost demand for advanced-node logic chips but also significantly increase demand intensity for high-end memory chips (especially HBM/enterprise SSD related). Against the backdrop of rising chip manufacturing process complexity, the number of "leading-edge advanced process steps" per wafer increases, making it easier for the equipment side to experience sustained demand and improved order visibility.
Citigroup's WFE model breakdown explicitly assumes NAND +30%, DRAM +12%, and Foundry/Logic +10% (according to their model), indicating the institution believes 2026's growth will not be solely supported by the AI computing segment but will involve more balanced expansion of advanced chip processes, which typically benefits semiconductor equipment leaders with broad coverage and complete product portfolios (e.g., those focused on deposition, etching, cleaning, metrology, advanced packaging, and other multiple segments).
Recent semiconductor industry outlook data from the World Semiconductor Trade Statistics (WSTS) organization shows that the global chip demand expansion trend is expected to continue strongly in 2026, and chips like MCUs and analog chips, which have seen persistently weak demand since late 2022, are also expected to enter a strong recovery curve. WSTS expects that after a strong rebound in 2024, the global semiconductor market will grow 22.5% in 2025, reaching a total value of $772.2 billion, higher than its spring outlook; the total semiconductor market value in 2026 is then expected to expand significantly to $975.5 billion on the basis of 2025's strong growth, approaching SEMI's projected 2030 market size target of $1 trillion, implying a potential year-over-year surge of 26%.
WSTS stated that this trend of strong growth for two consecutive years will be primarily driven by sustained robust momentum in the logic chip sector led by AI GPUs and the memory sector dominated by HBM memory systems, DDR5 RDIMM, and enterprise data center SSDs, with both sectors expected to achieve very strong double-digit growth, benefiting from持续强劲扩张需求 in areas like AI inference systems and cloud computing infrastructure.
Lithography, Etching, Thin-Film Deposition, and Advanced Packaging Equipment Enter a Cycle of Surging Demand The Citigroup analyst team stated that the "Phase 2 upturn cycle" signifies a shift in the valuation anchor from "valuation bottoming and recovery" to "sustained earnings revisions upwards": when the total WFE pie shifts from the base scenario towards the bull scenario, the earnings elasticity of leading companies in the semiconductor equipment field could even exceed revenue elasticity (due to economies of scale, rising capacity utilization, and an increasing proportion of high-end processes). Therefore, Citigroup chooses to express the "upslope" prospect through a semiconductor equipment portfolio comprising ASML Holding NV, Lam Research, and Applied Materials.
These three semiconductor equipment giants essentially cover lithography, etching, thin-film deposition, and advanced packaging equipment—precisely the semiconductor equipment sub-sectors best positioned to benefit from the AI infrastructure frenzy and the memory super cycle.
The Extreme Ultraviolet (EUV) lithography machines launched by Netherlands-based lithography giant ASML Holding NV (ASML Holding NV) are essential equipment for major chipmakers like TSMC and Samsung Electronics to produce the AI chips that power the world's most advanced AI applications like ChatGPT and Claude amidst the unprecedented global AI boom since 2023. They are also indispensable machine systems for memory giants like SK Hynix and Micron Technology to build core storage equipment such as HBM memory systems and data center enterprise SSDs/DDR against the macro backdrop of the current memory super cycle potentially lasting until 2027.
Performance leaps for AI GPU/AI ASIC accelerators highly depend on advanced logic nodes (3nm to 2nm, or even more advanced 1.8nm, 1.6nm), and the critical layers of these nodes must use EUV or even High-NA EUV to achieve smaller linewidths and higher yields. ASML's EUV/High-NA EUV equipment is clearly targeted at the mass production needs of 3nm and sub-2nm logic and globally leading-performance DRAM, representing "bottleneck capital goods" for advanced process capacity expansion.
Simultaneously, AI training/inference has ignited the "memory side": HBM memory systems paired with AI GPUs/ASICs require further miniaturization of DRAM manufacturing nodes and significantly increased manufacturing steps and equipment density for etching, thin-film deposition, CMP processes, and most critically, stacking packaging and interconnect steps (TSV/hybrid bonding advanced packaging, etc.). Applied Materials pointed out in its latest technical interpretation that the HBM manufacturing flow adds approximately 19 additional materials engineering steps compared to traditional DRAM and claimed its most advanced semiconductor equipment covers about 75% of these steps. The company also significantly launched bonding systems for advanced packaging/memory chip stacking, positioning HBM and advanced packaging manufacturing equipment as strong medium-to-long-term growth vectors, while new chip manufacturing node equipment like GAA (Gate-All-Around)/backside power delivery (BPD) will be core drivers for its next wave of strong growth.
Compared to Applied Materials, Lam Research's (Lam Research) advantages are comprehensively concentrated in the high-aspect-ratio (HAR) etching/deposition and related process capabilities required for advanced HBM memory, and both 3D NAND/advanced DRAM structures and interconnects also heavily rely on Lam's proprietary HAR processes.