JPMorgan: Semiconductor Industry Fundamentals Strengthen, Cloud Capital Expenditure Reaches $954 Billion

Written by: Rita

In its Fall 2026 U.S. Semiconductor and Semiconductor Equipment Series Report released on September 18, 2026, J.P. Morgan noted that semiconductor industry fundamentals are strengthening and that recent sector corrections are constructive. The bank’s models project 118% growth in the semiconductor industry in 2026 and 35% growth in 2027. Wafer manufacturing equipment is expected to grow by 31% in 2026 and 38% in 2027. J.P. Morgan considers sustainability of AI infrastructure spending, uncertainty around frontier lab investments, elevated valuations, and geopolitical risks as key volatility factors, but underlying demand and supply fundamentals remain robust.

Analyst Harlan Sur of J.P. Morgan noted in his report that over the past decade, the semiconductor cycle has been demand-driven, with supply-side expansion exhibiting greater discipline. The last four downturns were all demand-driven, and supply chains have been more restrained in expanding capacity. The firm believes the structural difference in the current cycle lies in supply discipline and diversified end markets, which have reduced cyclicality and shifted industry focus toward profitability and free cash flow expansion.

JPMorgan’s cloud capital expenditure dataset shows that cloud capex is expected to reach $954 billion in 2026, $1.41 trillion in 2027, and $1.54 trillion in 2028. Year-to-date, cloud capex expectations have been significantly raised. The bank noted that early AI monetization and ROI signals are becoming more visible, with accelerated growth in cloud and AI revenues, expanding backlogs, increased long-term contract capacity, and improved profitability in cloud and AI businesses.

If these trends persist, cloud service providers will have strong economic incentives to maintain high levels of AI infrastructure spending. J.P. Morgan believes that improved AI economics support a higher and more sustained spending environment, as well as stronger semiconductor fundamentals. The continued upward revision of cloud capital expenditures is a core pillar supporting visibility into semiconductor demand.

JPMorgan estimates the custom AI ASIC market will reach $60 billion to $70 billion by 2026, with a compound annual growth rate of 40% to 50% over the coming years. As AI workloads become more specialized, hyperscalers and leading model builders are increasingly seeking workload-optimized chips to enhance performance, reduce power consumption, lower cost per token, and ensure supply.

Custom AI ASICs require deep expertise in multi-chip SoC design, high-speed SerDes, HBM memory interfaces, advanced packaging, and network and system-level productization. J.P. Morgan believes that most customers cannot internally replicate these capabilities at scale. Broadcom and Marvell dominate the high-end ASIC market, with estimated market shares of 80% to 85% and 10% to 12%, respectively.

Broadcom's AI revenue is expected to grow from $58 billion in FY26 to $135 billion in FY27 and $245 billion in FY28, driven by projects from Meta, Google, Anthropic, OpenAI, and SoftBank/ARM. Marvell's data center revenue is projected to increase from $10 billion in CY26 to $16 billion in CY27 and $28 billion in CY28, fueled by optical DSP, Amazon Trainium 3+4, Microsoft Maia, and Google XPU initiatives.

J.P. Morgan expects DRAM mixed pricing to rise approximately 250% in 2026 and approximately 30% in 2027. NAND pricing is expected to rise approximately 250% in 2026 and approximately 25% in 2027. DRAM bit demand is projected to grow 29% in 2026 and 32% in 2027. NAND bit demand is projected to grow 24% in 2026 and 29% in 2027.

The long-term strategic customer agreement has become a key cyclical catalyst, locking in pricing and supply for DRAM and NAND, and improving visibility. Capital expenditures remain disciplined, with capital intensity below historical averages, limiting the risk of supply oversupply. The primary demand-side risk is BOM inflation-driven disruption, particularly in the client and consumer markets. J.P. Morgan maintains an Overweight rating on Micron.

The semiconductor recovery is expanding into a more sustained upward cycle. Texas Instruments, Analog Devices, and Microchip Technology all noted improved industrial demand, accelerated exposure to AI and data centers, stronger aerospace and defense sectors, and resilient growth in automotive content. Forward-looking demand signals have improved, channel inventories are healthier, order activity is stronger, and lead times have improved.

Pricing, delivery times, and profit margins are improving. Texas Instruments, Analog Devices, and Microchip Technology all reported strengthening demand, extended delivery times, and more supportive pricing, with higher utilization and product mix expected to benefit profit margins. JPMorgan maintains an Overweight rating on Analog Devices with a $500 price target.

JPMorgan raised its 2026 wafer manufacturing equipment growth forecast to 31% and its 2027 forecast to 38%. Rising AI demand has accelerated tool procurement ahead of multi-year greenfield/brownfield expansions. Foundry/logic spending is supported by TSMC’s N2/N3 capacity builds, with Intel providing additional upside. Asian research indicates TSMC’s roadmap is more aggressive, with multiple fab shells set to break ground in the first half of 2026 and equipment orders rising starting in Q2 2026.

Wafer equipment spending is expanding, led by increased DRAM capacity, while NAND spending is more driven by conversion/migration until potential upside from greenfield projects in the second half of 2028. J.P. Morgan believes KLA offers the optimal risk-reward profile, supported by its exposure to foundry/logic wafer equipment in 2027, as Intel’s increased involvement in yield/cost improvements and higher capital expenditures drive upward revisions to CY27 consensus estimates. J.P. Morgan maintains Overweight ratings on KLA, Applied Materials, Lam Research, and MKS Instruments.

JPMorgan’s large-cap favorites include Broadcom, Analog Devices, Marvell, Micron, and KLA, with additional optimism toward NVIDIA, Applied Materials, and Lam Research. Mid- and small-cap favorites include Astera Labs, Amkor, and MKS Instruments. Broadcom is collaborating with four of its six AI ASIC customers on next-generation chip stacking projects via its 3D SOIC reference platform. Marvell achieves 33% more HBM stacking, 70% lower interface power consumption, and 25% more XPU silicon area with its custom HBM architecture.

$954 billion in cloud capital expenditures supports visibility into semiconductor demand, with AI ASICs and memory being the two main drivers of this upcycle.

Disclaimer

This article is a compilation and interpretation by Chaoxiang Research of a third-party brokerage research report (JPMorgan, September 18, 2026), combined with publicly available market information. The ratings, price targets, earnings forecasts, and related judgments cited herein reflect the views of the brokerage’s analysts and represent only the position of their respective institution; they do not reflect the views of Chaoxiang Research nor constitute any investment advice.

The market carries risks; make decisions independently. This article should not be used as a basis for buying or selling any securities.