The Semiconductor Business Cycle

Introduction

Cyclicality is the defining characteristic of semiconductor economics and the feature that most differentiates chip company analysis from software or internet company analysis within TMT. While SaaS companies generate predictable recurring revenue and internet platforms grow steadily with user bases, semiconductor companies experience dramatic swings in revenue, margins, and profitability driven by inventory dynamics, demand fluctuations, and the multi-year lead times inherent in chip manufacturing. The 2023 downturn, the seventh major semiconductor cycle since 1990, saw industry sales decline 9.4% to $520 billion, with memory chip sales dropping 31% (approximately $40 billion). The subsequent recovery pushed revenues above $588 billion in 2024. For TMT investment bankers, understanding the semiconductor cycle is not optional: it determines when deals happen, how companies are valued, and what strategic decisions management teams face.

The semiconductor cycle follows a predictable four-phase pattern, though the duration and magnitude of each phase varies with each cycle.

Semiconductor Business Cycle

The recurring pattern of expansion and contraction in semiconductor demand, revenue, and profitability. A complete cycle typically lasts 3-5 years and follows four phases: upcycle acceleration (rising demand, lengthening lead times, capacity constraints, expanding margins), peak (record revenues, historically high margins, aggressive capacity investment), downcycle (declining demand, inventory correction, falling margins, capex cuts), and trough/early recovery (stabilizing demand, inventory depletion, improving bookings, early margin recovery). The cycle is driven by the interaction between demand fluctuations (end-market growth or decline) and supply dynamics (the 18-36 month lag between capacity investment decisions and new fab output).

Phase 1: Upcycle acceleration. Demand for chips exceeds available supply. Lead times (the time between ordering and receiving chips) extend from the normal 8-12 weeks to 20-50+ weeks. Customers respond by placing larger orders and building safety stock, which amplifies apparent demand beyond actual end-market consumption. Semiconductor companies report accelerating revenue growth, expanding margins, and increasing order backlogs. This is when TMT bankers see the most aggressive M&A activity: companies have strong stock prices (useful as acquisition currency), rising cash flows, and the confidence to pursue strategic deals.

Phase 2: Peak. Revenue and margins reach cyclical highs. Companies invest aggressively in new capacity (building fabs, ordering equipment from ASML and Applied Materials) to meet perceived demand. However, the demand signal is partially artificial: customers have been double-ordering and building inventory buffers, and the actual end-market demand may be growing more slowly than order books suggest.

Phase 3: Downcycle. Customers begin consuming their excess inventory instead of placing new orders. Revenue declines sharply as the bullwhip effect reverses. Margins compress because manufacturing costs are largely fixed (fabs must keep running to maintain equipment calibration and employee skills), but revenue is declining. Companies cut capex, implement hiring freezes, and in severe downturns, restructure operations. The 2023 downturn saw industry inventories exceed $60 billion, and memory makers' raw material stockpiles surged to five times their historical average.

Phase 4: Trough and recovery. Inventory levels normalize as customers deplete their excess stock. New orders begin to stabilize, then grow. Lead times return to normal levels, and pricing stabilizes. The recovery phase is when TMT bankers see distressed M&A activity: weaker companies with depleted cash flows become acquisition targets for stronger competitors. The 2024-2025 recovery demonstrated this pattern: after the severe 2023 downturn, semiconductor revenues rebounded sharply, driven initially by AI chip demand (Nvidia's data center revenue surged) and then more broadly as inventory levels across the supply chain returned to normal. Memory chip prices, which had collapsed during the downturn, recovered as DRAM supply tightened and data center demand for high-bandwidth memory (HBM) accelerated.

The practical application of through-cycle valuation involves several steps. First, identify where the company sits in the current cycle by analyzing inventory levels, lead times, book-to-bill ratios (the ratio of new orders to shipments, where above 1.0 indicates growing demand and below 1.0 indicates declining demand), and capacity utilization. Second, calculate normalized revenue by averaging across the cycle or projecting current revenue adjusted for the cyclical position. Third, apply normalized margins (averaging peak and trough margins, weighted by the typical duration of each phase) to arrive at normalized earnings. Fourth, apply appropriate valuation multiples: forward P/E ratios are standard for mature semiconductor companies (8-12x for cyclical commodity players, 15-25x for structural growth companies like Nvidia), while EV/EBITDA provides a capital-structure-neutral comparison.

The rise of AI as a structural demand driver has introduced debate about whether the semiconductor cycle is becoming less pronounced. AI chip demand (data center GPUs, custom AI accelerators) is growing on a multi-year trajectory that may be less sensitive to traditional inventory cycles. However, the cyclical nature of other end markets (smartphones, PCs, automotive, industrial) means that the overall semiconductor industry will continue to exhibit cyclicality, even if AI-exposed companies experience more secular growth patterns. Consensus projections suggest modest growth in 2025-2026, a potential correction in 2027, and recovery in 2028, with wafer shipment growth rates of approximately +5%, +5%, -6%, and +10% respectively.

The semiconductor cycle directly shapes TMT deal flow in predictable ways. During upcycles, companies pursue strategic acquisitions (using elevated stock prices as acquisition currency), IPOs (timing public offerings to coincide with peak revenue and earnings), and growth-oriented capital raises. During downturns, the deal mix shifts toward distressed M&A, restructuring advisory, and defensive transactions. Understanding where the cycle stands helps TMT bankers anticipate which types of mandates will dominate the pipeline and how to structure pitches to semiconductor management teams and boards. The best semiconductor bankers build cycle timing into their coverage strategy, originating sell-side mandates near the peak (when valuations are highest) and buy-side mandates near the trough (when targets are cheapest).