The WFE Bull Case for 2026—2027 Has Become the Base Case, With the Growth Focus Shifting to End-Market Compute Demand

Overview

The report raises its forecasts for the global wafer fab equipment market across 2026—2028 and converts its previous bull case for 2026—2027 into the base case. The upgrades are driven mainly by accelerated DRAM investment, broader leading-edge logic spending, and 2nm capacity expansion, while revisions to NAND are relatively limited. As the industry rapidly overcomes cleanroom and supply-chain constraints, the research focus is shifting toward whether end-market compute demand can support the additional capacity.

Core views

Morgan Stanley raises its 2026 WFE forecast from US$155 billion, representing 32% year-over-year growth, to US$163 billion, representing 39% growth; its 2027 forecast from US$201 billion, representing 31% growth, to US$223 billion, representing 37% growth; and its 2028 forecast from US$227 billion, representing 13% growth, to US$254 billion, representing 14% growth. Since December 1, 2025, the 2026 and 2027 forecasts have been raised by a cumulative 26% and 54%, respectively, marking the fifth revision this year. On a quarterly basis, the industry's annualized run rate is expected to rise from approximately US$125 billion in 4Q25 to approximately US$200 billion in 4Q26 over four quarters. The report has not turned bearish, but it expects a clear deceleration in 2028 from a higher base; to maintain the 2027 growth rate, 2028 WFE would need to exceed US$300 billion. The previous view that the WFE cycle would be durable was based on two types of constraints: cleanroom space on the demand side and the equipment supply chain on the supply side would limit near-term upside but extend investment into the medium term. The report now believes that the industry may overcome both constraints simultaneously, indicating that it had previously underestimated both the strength of demand and the speed of capacity additions. The debate will therefore shift from how fast WFE can grow to whether end-market compute demand is sufficient to absorb the additional supply. The report cross-checks WFE demand against each GW of compute deployment and estimates that the more than US$100 billion increase in industry spending from 2025—2027 corresponds to approximately 30GW, broadly matching the internet team's forecasts of approximately 29GW and 36GW of industry compute deployment in 2027 and 2028, respectively. The overview section expects 2nm capacity to reach 335 thousand wafers per month and 500 thousand wafers per month in 2027 and 2028; the subsequent detailed wafer model lists 180 thousand wafers per month in 2027 and 335 thousand wafers per month in 2028. Both sets of figures point to rapid expansion in 2nm capacity. Logic equipment is one of the main sources of the upgrade. The overall logic WFE forecasts for 2026 and 2027 are raised from US$89 billion and US$114 billion to US$95 billion and US$128 billion. The model summary expects leading-edge logic WFE to grow by 72% and 50% in 2026 and 2027, respectively, while mature-node logic declines by 3% and grows by 15%, respectively. The leading-edge/mature-node mix is expected to reverse from 39%/61% in 2025 to 63%/37% in 2028. The thematic analysis section separately estimates leading-edge logic WFE at US$51 billion in 2026, up 65%; US$76 billion in 2027, up 50%; and US$90 billion in 2028, up 17%, with corresponding mature-node growth rates of -2% and +18%. The largest driver in 2027 is 2nm, while equipment investment for 1.4nm also begins to emerge. The report expects approximately US$50 billion of 2nm spending that year, with TSMC spending US$43.3 billion and Intel spending US$20.5 billion; Intel is expected to become an important second-largest leading-edge logic investor. For 2028, the model conservatively assumes slower spending growth at TSMC and Intel, a modest recovery in Samsung investment, and excludes a potential US$2 billion contribution from SpaceX. From a process-node capacity perspective, monthly 3nm capacity is expected to increase from 180 thousand wafers in 2025 to 265 thousand wafers in 2026, with only modest increases during 2027—2029. The detailed model expects monthly 2nm capacity to reach 180 thousand wafers and 335 thousand wafers in 2027 and 2028, respectively, through joint expansion by TSMC, Intel, Samsung, and other manufacturers. Monthly 1.4nm capacity is expected to reach 60 thousand wafers and 130 thousand wafers in 2027 and 2028, driven mainly by TSMC and Intel. The view on mature nodes is notably more cautious: the report estimates that investment will remain broadly flat from 2023—2026 and recover by only approximately 18% in 2027. This means logic equipment growth is increasingly concentrated in leading-edge nodes rather than reflecting synchronized expansion across the entire logic market. DRAM is another core source of the upgrade. The 2026 and 2027 DRAM WFE forecasts are raised from US$49 billion and US$62 billion to US$53 billion and US$70 billion, representing year-over-year growth of 70% and 33%, respectively, and more than US$120 billion in aggregate over the two years. The 2026 forecast is 47% above the estimate at the start of the year, due to factors including Micron accelerating equipment installation after acquiring PSMC's Tongluo P5 fab, contributing approximately US$4 billion to US$5 billion of incremental spending; Micron advancing its migration to the 1Gamma node, driving 48% growth in brownfield DRAM WFE; and Samsung allocating more P4 cleanroom space to DRAM, with wafer starts expected to reach 85 thousand per month by the end of 2026. The report believes manufacturers have overcome the original capacity-expansion constraints by releasing cleanroom space and accelerating equipment deliveries. Forecast incremental monthly greenfield DRAM capacity for 2026 and 2027 is raised from 268 thousand wafers and 395 thousand wafers to 283 thousand wafers and 493 thousand wafers. Consequently, DRAM bit-supply growth forecasts are raised from 29% and 33% to 32% and 38%. Annual DRAM WFE is expected to average US$67 billion during 2026—2028, significantly above US$27 billion during 2023—2025; annual greenfield capacity additions over the same period are expected to average 427 thousand wafers per month, compared with 170 thousand wafers per month over the preceding three years. HBM bit supply is expected to grow by 55% and 57% in 2026 and 2027, while conventional DRAM grows by 30% and 36%. However, beginning in 2027, incremental non-HBM wafer capacity is expected to exceed incremental HBM capacity. DRAM wafer starts are expected to rise from 2,169 thousand wafers per month in 3Q26 to more than 3,000 thousand wafers per month in 2Q28. The report still sees upside potential for DRAM in 2027. Equipment suppliers are nearly unanimous that DRAM WFE growth will exceed overall WFE growth, implying 2027 spending of more than US$75 billion and year-over-year growth of approximately 40%, above Morgan Stanley's US$70 billion forecast. Its US$70 billion base case already includes 493 thousand wafers per month of greenfield expansion, nearly twice the 283 thousand wafers per month in 2026 and three times the 165 thousand wafers per month in 2025. The report therefore believes further upside to greenfield capacity beyond the forecast is unlikely; brownfield investment is more likely to be underestimated. The model assumes that a substantial portion of node migration has been brought forward to 2026 and that brownfield spending will decline from US$25 billion in 2026 to US$15.5 billion in 2027. This deceleration assumption is the principal source of potential forecast error. Revisions to NAND are smaller. The 2026 and 2027 NAND WFE forecasts are adjusted from US$15.6 billion and US$24 billion to US$13.7 billion and US$24 billion, mainly due to a modest reduction in 2026 brownfield investment, while the greenfield forecast remains broadly unchanged. The corresponding bit-supply growth forecasts are revised from 24% and 31% to 24% and 26%, and are expected to remain in the mid-20% range. New wafer capacity is expected to make a greater contribution to bit growth beginning in 2027. YMTC is expected to continue accounting for approximately half of incremental greenfield capacity during 2026—2027, while enterprise SSDs' share of the NAND demand mix is expected to rise from 29% in 2025 to 65% in 2028. Despite the slightly lower base, the report still expects NAND to be the fastest-growing WFE category in 2027. Advanced packaging maintains strong growth. Applied Materials, Lam Research, and KLA all guide to advanced packaging business growth of more than 70% in 2026. The report believes that CoWoS capacity expansion will be the primary driver in 2026, with the growth driver shifting to HBM and SoIC in 2027. Demand for CoWoS-, HBM-, and SoIC-related equipment is expected to grow by 59% and 56% in 2026 and 2027, respectively, and advanced packaging growth should continue to outpace overall WFE. Front-end-related advanced packaging revenue is also expected to continue outperforming back-end revenue. Advanced packaging revenue at Applied Materials, Lam Research, KLA, and Tokyo Electron is expected to grow by approximately 70%.

Analysis framework

The report first updates the total global WFE market and its growth rate from a top-down perspective, then breaks it down into leading-edge logic, mature-node logic, DRAM, NAND, and advanced packaging. For each segment, it develops bottom-up forecasts using monthly wafer capacity, greenfield and brownfield investment, node migration, bit-supply growth, and capital spending by major manufacturers, and validates them against equipment-supplier guidance. Finally, the report converts equipment investment into the corresponding compute deployment demand in GW to assess whether rapid capacity expansion aligns with end-market compute infrastructure buildout, while comparing the bull case, base case, and potential sources of forecast error.

Methodology notes

• Industry/Sector Analysis FrameworkSupply-demand framework

  Cleanroom and supply-chain constraints and end-market compute demand

  The report treats cleanroom space and the equipment supply chain as near-term supply constraints and uses compute deployment in GW as a proxy for end-market demand to assess the durability of WFE expansion and whether incremental capacity can be absorbed.

• Industry/Sector Analysis FrameworkVolume-price decomposition

  Decomposition of WFE spending, wafer capacity, and bit supply

  The report further decomposes changes in equipment spending into greenfield and brownfield investment, monthly wafer capacity, process-node migration, and bit-supply growth to explain how capital expenditure translates into actual output.

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

  Transmission from compute deployment to chip capacity and equipment demand

  Beginning with compute deployment in GW, the report connects demand for leading-edge logic, DRAM, NAND, and advanced packaging, then maps it to wafer fab equipment investment to validate the end-market basis for more than US$100 billion of incremental industry spending.

• (Out-of-Vocabulary Method)

  Bottom-up leading-edge logic wafer model

  The report develops forecasts based on capacity and spending at the 3nm, 2nm, and 1.4nm nodes and by manufacturers such as TSMC, Intel, and Samsung to determine the growth trajectory of leading-edge logic WFE.

• (Out-of-Vocabulary Method)

  Bull-case and base-case comparison

  The report tracks conditions from its previous bull case, including Intel spending, greenfield DRAM expansion, and the resumption of greenfield NAND investment. As these conditions have gradually materialized, it converts the bull case into the base forecast.

Asset mapping & comparison

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

• Global Semiconductor Capital Equipment (WFE)

  Market forecasts for 2026—2028 are raised across the board, with DRAM and leading-edge logic representing the largest sources of incremental growth.

  Strengths

  High growth of 39% and 37% is expected in 2026 and 2027, respectively, while supply-chain and cleanroom constraints are being overcome.

  Weaknesses

  Growth is expected to slow to 14% in 2028 from a high base.

  Comparison

  The industry's annualized run rate is expected to rise from approximately US$125 billion to approximately US$200 billion within four quarters.

  Risks

  The 2027 forecast could be raised further if DRAM spending exceeds expectations.

• Leading-Edge Logic WFE

  Driven by 2nm capacity expansion, accelerated TSMC investment, and Intel becoming the second-largest spender.

  Strengths

  The thematic model expects growth of 65% and 50% in 2026 and 2027, respectively, with approximately US$50 billion of 2nm spending in 2027.

  Weaknesses

  Spending growth at TSMC and Intel is assumed to slow in 2028.

  Comparison

  Leading-edge nodes' share of logic WFE is expected to rise to 63% by 2028, exceeding mature nodes' 37%.

  Risks

  The report does not include a potential US$2 billion contribution from SpaceX in its 2028 model.

• DRAM WFE

  Accelerated equipment investment, greenfield expansion, and node migration jointly drive 2026—2027 spending above US$120 billion.

  Strengths

  Incremental monthly greenfield capacity is expected to reach 493 thousand wafers in 2027, with bit-supply growth of 38%.

  Weaknesses

  Substantial investment was brought forward to 2026, leaving 2027 brownfield spending with a high comparison base.

  Comparison

  Supplier guidance implies 2027 spending of more than US$75 billion, above Morgan Stanley's US$70 billion forecast.

  Risks

  Brownfield investment in 2027 could exceed the model assumption, prompting further upgrades to DRAM WFE.

• NAND WFE

  Brownfield investment declines slightly in 2026, but greenfield investment and new wafer capacity still support growth in 2027.

  Strengths

  The 2027 WFE forecast remains at US$24 billion, with YMTC accounting for approximately half of incremental greenfield capacity.

  Weaknesses

  The 2026 forecast is lowered to US$13.7 billion, while bit-supply growth remains in the mid-20% range.

  Comparison

  The report still expects NAND to become the fastest-growing WFE category in 2027 from a lower base.

• Advanced Packaging Equipment

  Driven by CoWoS capacity expansion in 2026, with the growth driver shifting to HBM and SoIC in 2027.

  Strengths

  Related equipment demand is expected to grow by 59% and 56% in 2026 and 2027, respectively, continuing to outpace overall WFE.

  Comparison

  Advanced packaging revenue at Applied Materials, Lam Research, KLA, and Tokyo Electron is expected to grow by approximately 70%.

Key data

• 2026 WFE ForecastUS$163 billion, +39% YoYPreviously US$155 billion, +32%
• 2027 WFE ForecastUS$223 billion, +37% YoYPreviously US$201 billion, +31%
• 2028 WFE ForecastUS$254 billion, +14% YoYPreviously US$227 billion, +13%
• Cumulative Forecast Upgrade2026 +26%; 2027 +54%Compared with the forecasts as of December 1, 2025
• Industry Annualized Run RateApproximately US$125 billion rising to approximately US$200 billionFrom 4Q25 to 4Q26
• 2026—2027 Logic WFE ForecastUS$95 billion/US$128 billionPreviously US$89 billion/US$114 billion
• 2027 Leading-Edge Logic SpendingUS$76 billion, +50% YoYIncluding approximately US$50 billion of 2nm spending
• Major Leading-Edge Logic Investors in 2027TSMC US$43.3 billion; Intel US$20.5 billionIntel is expected to become an important second-largest investor
• 2026—2027 DRAM WFEUS$53 billion/US$70 billionPreviously US$49 billion/US$62 billion; more than US$120 billion in aggregate over two years
• Incremental Monthly Greenfield DRAM Capacity283 thousand wafers in 2026; 493 thousand wafers in 2027Previously forecast at 268 thousand wafers and 395 thousand wafers
• DRAM Bit-Supply Growth32% in 2026; 38% in 2027Previously 29% and 33%
• 2026—2027 NAND WFEUS$13.7 billion/US$24 billionPreviously US$15.6 billion/US$24 billion
• NAND Bit-Supply Growth24% in 2026; 26% in 2027Previously 24% and 31%
• Enterprise SSD MixRising from 29% to 65%From 2025 to 2028
• Advanced Packaging Equipment Demand Growth59% in 2026; 56% in 2027Covering CoWoS-, HBM-, and SoIC-related equipment

Impact & implications

The report believes that the upside in WFE during 2026—2027 is no longer merely an extension of the cycle caused by constrained equipment deliveries, but a higher base created by simultaneous expansion in leading-edge logic, DRAM, and advanced packaging. The clearest beneficiaries are demand for 2nm, DRAM, and HBM/SoIC-related equipment, while mature nodes and NAND in 2026 are relatively moderate. At the same time, the industry's rise from an annualized run rate of approximately US$125 billion to approximately US$200 billion within four quarters significantly raises the growth hurdle for 2028. Subsequent assessments will need to rely more heavily on end-market compute demand, absorption of bit supply, and execution of manufacturers' capital-spending plans rather than solely on the WFE growth rate itself.

Risks

• Equipment-supplier guidance implies 2027 DRAM WFE of more than US$75 billion, above the report's US$70 billion base-case forecast, leaving forecast risks skewed to the upside.
• The report may underestimate 2027 brownfield DRAM investment because the model assumes brownfield spending declines significantly from US$25 billion in 2026 to US$15.5 billion.

What to watch

• Monitor whether approximately 29GW and 36GW of compute-industry deployment in 2027 and 2028 can support rapidly growing WFE and chip supply.
• Track whether 2027 DRAM WFE moves toward the more than US$75 billion level implied by suppliers and whether brownfield investment exceeds US$15.5 billion.
• Watch whether TSMC's 2nm expansion and Intel's more than US$20 billion of leading-edge logic spending proceed as planned.
• Track whether incremental greenfield DRAM capacity reaches 493 thousand wafers per month in 2027 and whether bit-supply growth rises to 38%.
• Monitor the resumption of greenfield NAND investment, YMTC's approximately 50% share of incremental capacity, and the rising enterprise SSD mix.
• Watch whether the growth driver for advanced packaging shifts from CoWoS to HBM and SoIC in 2027 as the report expects.
• Monitor whether 2028 WFE can exceed US$300 billion, the scale required to maintain the industry's current growth rate.