JPMorgan Tech Investing 101: Industry Framework and Investment Rules for Semiconductors, Hardware, and Software

Overview

This is a JPMorgan introductory deep report on the information technology industry investment framework, covering three major sub-segments: semiconductors, hardware and equipment, and software and services. Starting from the value chain, product cycles, semiconductor supply-demand cycles, regional capacity distribution, S-curves, EPS revisions, and valuation methods, the report explains why the tech industry has long-term growth potential while also carrying high volatility, high attrition, and high disruption risk.

Core views

The core viewpoints of the report are: first, the technology industry does not have a stable equilibrium, and company valuation rankings in this sector change significantly every 10 years, with two to three new entrants entering the top five in TMT on average; second, semiconductors are the foundation of technology hardware, with long-term revenue and unit growth around 10%, but cyclical growth can swing from contraction to 20%-30%; third, GenAI is becoming a new disruptive force, and at this stage value is flowing more toward semiconductors than software vendors; fourth, in driving hardware and semiconductor stock prices, EPS revision momentum is generally stronger than valuation itself.

Analysis framework

The report uses a combined top-down and bottom-up approach: it first defines three sub-sectors of information technology and representative companies, then breaks down the semiconductor ecosystem into raw wafers, EDA, IP, design services, equipment, foundries, fabless, OSAT, power electronics, DRAM, and NAND; then explains price volatility through product S-curves and semiconductor supply-demand cycles; finally discusses valuation and trading signals using 12-month forward P/E, P/B, reverse DCF, and EPS revision frameworks.

Methodology notes

• Industry ClassificationThree-part Technology Framework

  Semiconductors, Hardware & Equipment, Software & Services

  The report splits information technology into semiconductors, hardware and equipment, and software and services to differentiate capital intensity, revenue model, cyclicality, profitability, and valuation drivers.

• Technological ProgressMoore’s Law

  Doubling performance or maintaining the same performance at lower cost over a two-year period at constant cost

  The report uses Moore’s Law to describe the long-term productivity progress foundation of the semiconductor industry, meaning the same semiconductor performance can be delivered in two years at 20%-25% lower cost, or performance can improve by 25%-30% at the same cost.

• Product CycleS-curve Framework

  Use an S-curve to locate where companies and products are in their growth stages

  The report argues that markets often linearly extrapolate, so mispricing is likely in early and late phases of the product cycle; following the product cycle after confirming the trend is usually more effective than contrarian trading.

• Cyclical AnalysisSemiconductor Cycle Framework

  Demand, inventories, capacity utilization, pricing, and capex expansion jointly drive semiconductor cycles

  The report describes the semiconductor cycle as moving from weak demand, inventory drawdown, and low utilization, to demand recovery, low inventory, and rising prices, then to overordering, capex expansion, weakening supply-demand balance, and inventory correction.

• Valuation Framework12M Forward P/E, P/B, Reverse DCF, EPS Revisions

  Valuation toolset for technology hardware and semiconductors

  The 12-month forward P/E is the most widely used metric; P/E fails at cycle inflection points, while P/B is more suitable as a downside anchor; DCF is highly sensitive to earnings and cash flow volatility, reverse DCF is better for testing market expectations; EPS revision momentum typically trumps static valuation.

Asset mapping & comparison

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

• Semiconductors

  The foundational layer of information technology; all hardware devices depend on chips.

  Strengths

  Long-term demand is broad, and new demand such as AI servers can generate strong incremental upside; Moore’s Law supports long-term productivity gains.

  Weaknesses

  It is capital-intensive, with pronounced supply-demand cyclicality, and inventory and capacity utilization swings can amplify earnings volatility.

  Comparison

  Compared with software, it is more manufacturing and cyclical; compared with hardware brands, it sits lower in the value chain as core technology and supply-chain infrastructure.

  Risks

  Geopolitical risks, capacity mismatch, inventory correction, price declines, and failed technology-node transitions.

• Logic Semiconductors

  Responsible for data processing, including CPUs, SoCs, GPUs, ASICs, and FPGAs.

  Strengths

  Relatively lower volatility than memory, with more common fabless and foundry models, tied to AI, compute, and high-performance chip demand.

  Weaknesses

  Still linked to global GDP and end-demand cycles, with high barriers in advanced-process and design.

  Comparison

  Accounted for 71% of global semiconductor market revenue in 2025, higher than memory semiconductors.

  Risks

  Advanced-process bottlenecks, customer concentration, export controls, and intensifying competition.

• Memory Semiconductors

  Responsible for data storage, including DRAM, NAND Flash, and NOR Flash.

  Strengths

  Rising demand for storage capacity from AI, servers, and endpoint devices can support up-cycle moves.

  Weaknesses

  Stronger commodity characteristics, with prices and supply-demand more prone to sharp volatility.

  Comparison

  Accounted for 29% of global semiconductor market revenue in 2025, with more visible cyclical swings than logic chips.

  Risks

  Oversupply after capacity expansion, declines in DRAM/NAND prices, and inventory correction.

• Hardware & Equipment

  Covers PCs, smartphones, servers, networking equipment, EMS/ODM, and key components.

  Strengths

  Brand, channel, manufacturing efficiency, and supply-chain capability are key competitive advantages; AI servers and power, connectors, optical modules, and other components can benefit in new cycles.

  Weaknesses

  Products become commoditized over time, margins are easily compressed, and growth depends more on replacement cycles.

  Comparison

  Compared with semiconductors, capital intensity is lower, but the sustainability of differentiation is generally weaker than for core chips and software subscription models.

  Risks

  Weak terminal replacement demand, brand competition, declining manufacturing margins, and replacement by new form factors.

• Software & Services

  Covers operating systems, enterprise applications, cloud platforms, SaaS, and IT services.

  Strengths

  Asset-light, high gross margins, and recurring subscription revenue; historically benefited from high valuation due to earnings visibility and predictability.

  Weaknesses

  Currently facing GenAI disruption, with part of the value potentially shifting from software vendors to semiconductors and compute infrastructure.

  Comparison

  Compared with semiconductors and hardware, software has stronger revenue visibility but can also be re-rated during paradigm shifts.

  Risks

  GenAI substitution or compression, client budget reallocation, valuation contraction, and slowing growth.

Key data

• Information technology weight in the S&P 50039%The report states that information technology is the largest sector in the S&P 500.
• Information technology weight in MSCI Asia-ex JP45+%Asian exposure is mainly concentrated in Taiwan and South Korea.
• SOX Index market capitalizationUS$14.8 trillionAs of June 1, 2026.
• Global semiconductor market revenueUS$796 billionData for 2025, up 26% year-over-year, sourced from WSTS as cited in the report.
• Logic chip revenue share71%Global semiconductor revenue composition in 2025.
• Memory chip revenue share29%Memory is more cyclical and has clearer commodity-like characteristics.
• Taiwan share of foundry capacity45+%The report says Taiwan remains a key foundry hub.
• Combined Taiwan and China share of OSAT capacity70+%Outsourced assembly and test capacity is concentrated in Taiwan and China.
• South Korea share of DRAM capacity55%South Korea is a key DRAM market.
• NAND capacity concentration90%South Korea, Japan, and China together account for about 90% of global NAND capacity.
• Long-term semiconductor revenue and unit growtharound 10%The report sees long-term growth as around global GDP plus low single digits, with pronounced cyclical swings in the cycle.

Impact & implications

For investors, the technology sector should not be screened using static valuation alone. Product-cycle stocks may deliver multi-bagger opportunities once the trend is confirmed, but they also exhibit Boom-Bust characteristics; so-called perfect-cycle stocks such as memory, foundry, and power semiconductors are more dependent on supply-demand balance, pricing, and capacity utilization; truly stable-growth companies are scarce and usually have strong cash flow, stable margins, lower capex, and cross-platform technology capabilities. AI servers, GenAI adoption, and semiconductor supply constraints are among the most important variables to monitor now.

Risks

• Technology evolves rapidly, market-cap leadership rotates frequently, and corporate failures can be difficult to recover from.
• Semiconductors and hardware are distinctly cyclical, and demand, inventories, pricing, capacity utilization, and capex pace can amplify earnings volatility.
• Chip supply chains are at the center of US-China tension, with elevated geopolitical risk.
• Hardware products become increasingly commoditized over time, and margins may be compressed.
• Software and services remain asset-light with recurring revenue but are currently under GenAI disruption concerns.
• At cycle inflection points, conventional valuation metrics such as traditional P/E and historical P/B can mislead investment judgment.

What to watch

• Momentum in EPS revisions, especially changes in earnings expectations for representative tech names such as NVDA, SK Hynix, AMD, and TSMC.
• Inventory levels, order and shipment relationships, capacity utilization, and new capex plans.
• DRAM prices, NAND prices, wafer prices, and power semiconductor prices.
• AI server demand, GenAI adoption speed, and token growth.
• Capacity changes in Taiwan, South Korea, China, Japan, Europe, and the United States across foundry, OSAT, DRAM, NAND, and IDM.
• The gap between 12-month forward P/E, cyclical-bottom P/B, reverse DCF implied expectations, and actual EPS revisions.