WFE Bull Case Becomes the Base Case: Nearly 40% Consecutive Growth Expected in 2026—2027

Overview

The report reassesses the expansion path of the global wafer fab equipment market and incorporates the previous bull case for 2026—2027 into its base-case forecast. Large-scale pull-in purchasing for DRAM, leading-edge capacity expansion at 2nm and 1.4nm, and advanced packaging demand jointly support strong growth. However, as supply constraints ease rapidly, the key determinant of cycle durability will shift to whether compute demand can absorb the additional capacity.

Core views

Morgan Stanley raises its WFE market forecasts for 2026, 2027, and 2028 from USD 1550/2010/2270 hundred million to USD 1630/2230/2540 hundred million, respectively, revising the corresponding year-over-year growth rates from 32%/31%/13% to 39%/37%/14%. Since 2025-12-01, its 2026 forecast has been raised cumulatively by 26%, while its 2027 forecast has been raised by 54%; this is already the fifth revision this year. The industry's annualized run rate is expected to rise from approximately USD 1250 hundred million in Q4 2025 to approximately USD 2000 hundred million in Q4 2026 in only four quarters, indicating that the previous bull case for 2026/27 has essentially become the base case. The report previously believed that cleanroom space and equipment supply-chain constraints would limit near-term upside but extend the medium-term equipment cycle; its latest assessment is that the industry is overcoming both constraints. The removal of these constraints is allowing capital expenditure and capacity expansion to proceed faster than previously estimated, shifting the research focus from “how fast can WFE continue to grow” to “can compute and end-market demand absorb the additional supply.” Based on its estimate of WFE required per GW of compute, the industry's incremental equipment spending of more than USD 1000 hundred million during 2025—2027 broadly corresponds to equipment demand for approximately 30GW, close to the internet team's forecasts of 29GW/36GW of industry compute deployment in 2027/28. The report does not turn bearish, but notes that growth from a higher base will slow significantly to 14% in 2028; maintaining the 37% growth rate of 2027 would require 2028 WFE to exceed USD 3000 hundred million. Leading-edge logic is the first major source of upward revisions. The 2026/27 logic WFE forecasts are raised from USD 890/1140 hundred million to USD 950/1280 hundred million, with the investment mix shifting markedly from mature nodes toward leading-edge nodes. The report's overall model projects leading-edge logic WFE growth of 72%/50% in 2026/27, while mature-node spending declines 3% and then grows 15%; the subsequent detailed analysis provides corresponding figures of 65%/50% and -2%/+18%. Although the specific measurement bases differ slightly, both sets of data point to the same conclusion: leading-edge growth will substantially outpace mature-node growth. The spending shares of leading-edge and mature-node processes are expected to reverse from 39%/61% in 2025 to 63%/37% in 2028; another long-term comparison shows the ratio shifting from 35%/65% in 2024 to 63%/37% in 2028. By process node, monthly 3nm capacity is expected to rise from 18 ten-thousand wafers in 2025 to 26.5 ten-thousand wafers in 2026, with limited incremental additions during 2027—2029. The detailed wafer model projects monthly 2nm capacity of 18/33.5 ten-thousand wafers in 2027/28 and 1.4nm capacity of 6/13 ten-thousand wafers; elsewhere, the report's cycle overview describes the 2nm capacity path for 2027/28 as 33.5/50 ten-thousand wafers. Despite the different measurement bases, the report consistently concludes that 2nm will exceed the peak level of 3nm in 2027 and continue expanding, with new capacity spanning manufacturers including TSMC, Intel, and Samsung. Leading-edge logic WFE is expected to increase from USD 510 hundred million in 2026, up 65% year over year, to USD 760 hundred million in 2027, up 50%, and then to USD 900 hundred million in 2028, up 17%. Spending growth on 3nm and 2nm will be similar in 2026, driven primarily by TSMC, Samsung, and Intel; approximately USD 500 hundred million of spending will be directed toward 2nm in 2027, while 1.4nm equipment investment will begin. The report expects TSMC to invest USD 433 hundred million and Intel USD 205 hundred million in 2027, making Intel the important second-largest spender after TSMC. For 2028, it assumes spending growth at TSMC and Intel will moderate, Samsung will recover slightly, and a potential substantive contribution of USD 20 hundred million from SpaceX is not included. Unlike leading-edge processes, mature-node spending is expected to remain broadly flat during 2023—2026 and recover moderately by approximately 18% only in 2027. DRAM is the largest and most consistently upgraded segment. The 2026/27 DRAM WFE forecasts are raised from USD 490/620 hundred million to USD 530/700 hundred million, totaling more than USD 1200 hundred million over two years and representing year-over-year growth of 70% and 33%, respectively. The 2026 forecast is 47% above its level at the beginning of the year, primarily because manufacturers are overcoming constraints through accelerated equipment deliveries, the release of existing cleanroom space, and capacity reconfiguration. Micron's acquisition of PSMC's Tongluo P5 fab is seen as accelerating equipment installation and adding approximately USD 40 to 50 hundred million of equipment spending in 2026; Micron's migration to 1Gamma also drives 48% growth in DRAM brownfield WFE in 2026; more Samsung P4 cleanroom space has been allocated to DRAM, with monthly capacity expected to reach 8.5 ten-thousand wafers by the end of 2026. The forecast for new DRAM capacity in 2026/27 is raised from 26.8/39.5 ten-thousand wafers per month to 28.3/49.3 ten-thousand wafers per month, with the 2027 upgrade driven primarily by Samsung and SK hynix. Average annual DRAM WFE during 2026—2028 is expected to be USD 670 hundred million, significantly above USD 270 hundred million during 2023—2025; average annual new capacity over the same period is expected to be 42.7 ten-thousand wafers per month, above 17 ten-thousand wafers in the prior three years. Monthly DRAM wafer starts are expected to increase from 216.9 ten-thousand wafers in Q3 2026 to more than 300 ten-thousand wafers in Q2 2028. The additional wafer capacity raises the 2026/27 DRAM bit-supply growth forecasts from 29%/33% to 32%/38%; HBM bit supply is expected to grow 55%/57%, while conventional DRAM grows 30%/36%. However, beginning in 2027, incremental non-HBM wafers are expected to exceed incremental HBM wafers. The report still sees upside potential for DRAM in 2027. Equipment suppliers almost unanimously expect DRAM WFE growth to exceed overall WFE growth, implying a 2027 market size above USD 750 hundred million and growth of approximately 40%, higher than Morgan Stanley's USD 700 hundred million forecast. Its USD 700 hundred million forecast already produces 38% bit-supply growth, making it difficult for the report to derive higher spending from its own capacity model. Nevertheless, given the strong consensus among suppliers, risks remain skewed to the upside. The report believes its forecast of 49.3 ten-thousand wafers per month of new capacity is already high and that brownfield investment is more likely to be underestimated: because node migrations have been pulled forward to 2026, brownfield WFE is assumed to decline from USD 250 hundred million in 2026 to USD 155 hundred million in 2027. If the decline is smaller than expected, 2027 DRAM WFE could be raised further. NAND revisions are relatively limited. The 2026/27 NAND WFE forecasts are adjusted slightly from USD 156/240 hundred million to USD 137/240 hundred million, primarily reflecting lower 2026 brownfield investment, while the new-capacity forecast remains broadly unchanged. As a result, the 2026/27 NAND bit-supply growth forecasts are revised from 24%/31% to 24%/26% and are expected to stabilize around the mid-20% range; beginning in 2027, new wafer capacity will contribute more supply growth. YMTC is expected to continue accounting for approximately half of new NAND capacity in 2026/27. Enterprise SSDs' share of the NAND demand mix is expected to rise from 29% in 2025 to 65% in 2028. Advanced packaging represents another major growth theme. AMAT, LAM, and KLA have all guided to advanced packaging business growth of more than 70% in 2026. The report believes CoWoS capacity expansion will be the primary driver in 2026, while HBM and SoIC will take over in 2027. Equipment demand related to CoWoS, HBM, and SoIC is expected to grow 59%/56% in 2026/27, and advanced packaging growth could still exceed overall WFE growth in 2027. Front-end-related advanced packaging revenue is expected to continue outpacing back-end revenue, with advanced packaging revenue growth for AMAT, LAM, KLA, and TEL all modeled at approximately 70%.

Analysis framework

The report first incorporates the latest equipment purchasing, cleanroom utilization, and supplier guidance into a bottom-up WFE forecast, then separately analyzes spending, wafer capacity, and bit supply for leading-edge logic, DRAM, NAND, and advanced packaging. It subsequently uses WFE demand per GW of compute to cross-check equipment investment against the internet team's compute deployment forecasts, and compares supplier-implied DRAM spending with its own model to assess potential forecast upside and the sustainability of growth in 2028.

Methodology notes

• Industry/Sector Analysis FrameworkSupply-demand framework

  Supply-demand matching among WFE investment, wafer capacity, and end-market compute demand

  The report estimates how much logic and memory capacity equipment investment can add while using compute deployment in GW to measure end-market demand, thereby assessing whether capital expenditure growth is sustainable and whether additional supply could expand too rapidly.

• Industry/Sector Analysis FrameworkUpstream-Midstream-Downstream Industry Chain Transmission

  Transmission from equipment spending to wafer capacity, bit supply, and compute deployment

  The report translates upstream WFE purchases into midstream wafer starts and DRAM and NAND bit supply, then maps them to downstream compute demand to explain the actual impact of equipment investment on industry supply.

• Industry/Sector Analysis Framework

  Leading-edge wafer capacity model

  The report builds monthly capacity and equipment spending forecasts by 3nm, 2nm, and 1.4nm nodes and major manufacturers to identify annual growth nodes, manufacturer contributions, and structural changes between leading-edge and mature-node processes.

• Industry/Sector Analysis Framework

  WFE per GW compute-intensity analysis

  The report estimates wafer fab equipment demand corresponding to each additional 1GW of compute deployment and cross-validates approximately 30GW of WFE demand against industry compute deployment forecasts of 29GW/36GW for 2027/28.

• Event-Driven Strategy and Behavioral FinanceExpectation Gap/Expectation Management

  Migration of the bull case into the base case

  The report tracks five consecutive forecast upgrades during the year and supplier guidance, concluding that the conditions for DRAM, leading-edge, and NAND capacity expansion previously confined to an upside scenario have gradually materialized.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• TSMC

  The report expects it to be the largest contributor to leading-edge logic WFE growth in 2027 and to lead additions in 2nm and 1.4nm capacity.

  Strengths

  Expected to invest USD 433 hundred million in 2027, with accelerating 2nm capacity expansion.

  Weaknesses

  The report assumes its spending growth will slow in 2028.

  Comparison

  Its leading-edge spending in 2027 is expected to exceed Intel's and represent the largest source of growth.

  Risks

  If the extent of spending deceleration in 2028 or end-market compute demand differs from forecasts, the industry's growth trajectory may change.

• Intel

  The report expects Intel to become the important second-largest leading-edge logic WFE spender in 2027.

  Strengths

  Expected to invest USD 205 hundred million in 2027 and participate in new 2nm and 1.4nm capacity.

  Weaknesses

  Its 2028 spending growth is conservatively assumed to slow.

  Comparison

  Expected to rank behind TSMC as the second-largest source of leading-edge growth in 2027.

  Risks

  If actual investment progress falls below the report's assumptions, the leading-edge WFE forecast will be affected.

• Samsung

  Samsung is participating in both leading-edge and DRAM capacity expansion, with more Samsung P4 cleanroom space allocated to DRAM.

  Strengths

  P4 DRAM wafer starts are expected to reach 8.5 ten-thousand wafers per month by the end of 2026, and its 2027 new DRAM capacity forecast has been raised substantially.

  Weaknesses

  The report expects only a modest recovery in its spending in 2028, with a lower contribution to leading-edge growth than TSMC and Intel.

  Comparison

  It trails TSMC and Intel in its contribution to leading-edge growth in 2027.

  Risks

  Changes in cleanroom allocation and the pace of capacity expansion could affect DRAM and logic WFE demand.

• Micron

  Through the acquisition of PSMC's Tongluo P5 fab and accelerated migration to the 1Gamma node, Micron is driving earlier realization of DRAM equipment spending in 2026.

  Strengths

  Accelerated equipment installation at the P5 fab is expected to add USD 40 to 50 hundred million in 2026, while node migration supports brownfield equipment investment.

  Weaknesses

  Some node migrations have been pulled forward to 2026, raising the comparison base for 2027.

  Comparison

  It is one of the primary manufacturers cited by the report to explain why 2026 DRAM WFE substantially exceeds expectations at the beginning of the year.

  Risks

  Accelerated investment could cause brownfield spending to slow in 2027.

• SK hynix

  It is an important contributor to the upgrade in new DRAM capacity and benefits from expansion demand for both HBM and conventional DRAM.

  Strengths

  The report makes a modest upward revision to its 2026 capacity and identifies it as one of the main sources of the 2027 upgrade.

  Comparison

  Together with Samsung, it represents the largest source of the upward revision to the 2027 new DRAM capacity forecast.

  Risks

  Rapid DRAM bit-supply growth could make the alignment between capacity expansion and end-market demand more important.

• YMTC

  It is expected to continue contributing approximately half of new NAND capacity in 2026/27.

  Strengths

  It has a relatively high share of incremental NAND capacity.

  Weaknesses

  The report lowers overall 2026 NAND WFE and bit-supply growth.

  Comparison

  Expected to account for approximately 50% of the industry's new NAND capacity.

  Risks

  Changes in NAND investment scale and the pace of new wafer capacity additions would affect the supply-growth trajectory.

• AMAT, LAM, KLA, and TEL

  The report expects these equipment companies to benefit from growth in advanced packaging demand, with related revenue growth of approximately 70%.

  Strengths

  CoWoS, HBM, and SoIC equipment demand is expected to grow 59%/56% in 2026/27, while front-end-related advanced packaging revenue continues to outpace back-end revenue.

  Weaknesses

  The main sources of growth will shift from CoWoS in 2026 to HBM and SoIC in 2027, requiring a smooth transition in product demand.

  Comparison

  The report expects all four companies to achieve approximately 70% growth in advanced packaging revenue.

  Risks

  If the pace of advanced packaging projects or transitions between technology paths fall short of expectations, related growth could be affected.

Key data

• 2026 WFE market forecastUSD 1630 hundred million, up 39% year over yearPrevious forecast: USD 1550 hundred million, up 32% year over year
• 2027 WFE market forecastUSD 2230 hundred million, up 37% year over yearPrevious forecast: USD 2010 hundred million, up 31% year over year
• 2028 WFE market forecastUSD 2540 hundred million, up 14% year over yearPrevious forecast: USD 2270 hundred million, up 13% year over year
• Forecast upgrades since 2025-12-012026 raised by 26%; 2027 raised by 54%The report states that this is the fifth WFE forecast revision this year
• Industry annualized run rateApproximately USD 1250 hundred million rising to approximately USD 2000 hundred millionOver four quarters, from Q4 2025 to Q4 2026
• 2026/27 logic WFE forecastUSD 950/1280 hundred millionPrevious forecast: USD 890/1140 hundred million
• 2027 leading-edge logic WFEUSD 760 hundred million, up 50% year over year2nm is the largest growth driver, with 1.4nm capacity additions beginning
• Major leading-edge spenders in 2027TSMC USD 433 hundred million; Intel USD 205 hundred millionIntel is expected to become the important second-largest spender
• 2027 2nm equipment spendingApproximately USD 500 hundred millionThe report expects 2nm to become the primary driver of leading-edge processes in 2027
• 2026/27 DRAM WFE forecastUSD 530/700 hundred millionPrevious forecast: USD 490/620 hundred million; more than USD 1200 hundred million in total over two years
• 2026/27 new DRAM capacity28.3/49.3 ten-thousand wafers per monthPrevious forecast: 26.8/39.5 ten-thousand wafers per month
• 2026/27 DRAM bit-supply growth32%/38%Previous forecast: 29%/33%
• 2026/27 HBM bit-supply growth55%/57%Conventional DRAM bit-supply growth during the same period is 30%/36%
• Average annual DRAM WFE during 2026—2028USD 670 hundred millionThe 2023—2025 average was USD 270 hundred million
• 2026/27 NAND WFE forecastUSD 137/240 hundred millionPrevious forecast: USD 156/240 hundred million
• 2026/27 NAND bit-supply growth24%/26%Previous forecast: 24%/31%
• Enterprise SSD mix shareRising from 29% in 2025 to 65% in 2028The report's forecast for the NAND demand mix
• Advanced packaging equipment demand growthGrowth of 59%/56% in 2026/27Includes equipment demand related to CoWoS, HBM, and SoIC

Impact & implications

The report believes cleanroom and supply-chain bottlenecks are no longer sufficient to constrain near-term WFE growth, and that DRAM, 2nm/1.4nm logic, and advanced packaging will drive consecutive high-speed equipment demand expansion in 2026—2027. At the same time, the approximately USD 1000 hundred million two-year industry increase and rapidly emerging memory and logic capacity mean that subsequent analysis must focus on compute and end-market demand; even if the market reaches USD 2540 hundred million in 2028, growth is expected to slow significantly.

Risks

• 2027 DRAM WFE could exceed USD 700 hundred million: equipment suppliers' assessments imply a market size above USD 750 hundred million, and the report believes forecast risks are skewed to the upside.
• The report may underestimate 2027 DRAM brownfield investment because its model assumes brownfield WFE will decline from USD 250 hundred million in 2026 to USD 155 hundred million.
• A large amount of DRAM investment was pulled forward into 2026, significantly raising the 2027 comparison base and potentially causing growth to slow.
• WFE growth is expected to decline from 37% in 2027 to 14% in 2028; maintaining the previous growth rate would require the market to exceed USD 3000 hundred million.
• As cleanroom and supply-chain constraints are overcome, whether additional capacity can be absorbed by compute and end-market demand becomes a key risk to cycle durability.

What to watch

• Track whether actual compute deployment in 2027/28 can approach the 29GW/36GW forecasts cited in the report.
• Monitor whether final 2027 DRAM WFE is closer to Morgan Stanley's USD 700 hundred million forecast or the supplier-implied level above USD 750 hundred million.
• Watch whether DRAM brownfield investment declines from USD 250 hundred million in 2026 to USD 155 hundred million in 2027 as modeled.
• Track TSMC's and Intel's 2nm equipment investment and whether Intel can become the second-largest leading-edge spender in 2027.
• Monitor whether mature-node spending can recover by approximately 18% in 2027 after remaining broadly flat during 2023—2026.
• Watch the contribution of new NAND wafer capacity to bit supply beginning in 2027 and whether YMTC continues to account for approximately half of new capacity.
• Track whether the growth driver for advanced packaging can transition smoothly from CoWoS in 2026 to HBM and SoIC in 2027.