Assessing the Semiconductor Cycle: AI's $1 Trillion Boom and the Macro Backdrop

The semiconductor industry is on the cusp of a historic milestone, with global sales projected to reach roughly $1 trillion worldwide in 2026. This surge is not a typical cyclical upswing. It is an AI-fueled acceleration that has propelled the sector to unprecedented growth rates, breaking through the $800 billion threshold in 2025 with sales of $791.7 billion, a 25.6% increase from the prior year. The momentum is set to continue, with a projected 26% growth in 2026 following 23% growth in 2025. This marks three consecutive years of 20%+ growth, a pattern not seen since the early PC boom of the 1990s.

The scale of this expansion is what fundamentally distinguishes it from past cycles. The growth is concentrated in specific, high-value segments. Sales of logic products, which include the GPUs and processors at the heart of AI, surged 39.9% in 2025. Memory, particularly the high-bandwidth DRAM needed for AI workloads, grew 34.8%. This isn't broad-based demand; it's a targeted infrastructure build-out driven by a single, transformative technology. As one analysis notes, the revised forecast was made possible by a single factor: AI.

This sets up a classic macro cycle tension. The AI-driven demand has rewritten the growth trajectory, pushing the trillion-dollar milestone forward by years. Yet, the very concentration that fuels this boom also creates vulnerability. The industry's health is now inextricably linked to the pace of AI adoption and investment. If macro conditions shift-such as a rise in real interest rates that pressures tech capital spending-or if AI demand slows, the sharp correction risk is amplified. The cycle has been reset, but the new engine runs on a different fuel, making its long-term stability a function of technological and economic forces beyond the semiconductor industry's control.

Macro Forces Shaping the Cycle: Real Rates, Dollar, and Growth

The AI-driven boom is a powerful force, but it operates within a broader macroeconomic framework that can either fuel or throttle the cycle. The extreme concentration in the industry's revenue mix-where AI chips drive roughly half of total revenue but represent less than 0.2% of total unit volume-creates a system highly sensitive to shifts in real interest rates and global growth trends. This isn't just about demand for more chips; it's about the capital available to fund the massive, leading-edge capacity build-out required to produce them.

The scale of that investment is staggering. Global semiconductor equipment sales are set to reach a record $133 billion in 2025, with projections of $156 billion by 2027. This spending is laser-focused on the technologies that power AI: leading-edge logic, memory, and advanced packaging. Such a capital-intensive ramp-up is a classic sign of a boom phase, but it also embeds the industry's future profitability in the sustainability of that investment. If macro conditions tighten-specifically if real interest rates rise or growth slows-the cost of financing this build-out increases, and the projected returns on those investments come under pressure.

This sets up a critical vulnerability. The capital required for this infrastructure is concentrated in a few dominant players. The market's trajectory is now mirrored by a dangerous concentration of capital. The five major technology companies now represent approximately 30% of S&P 500 market capitalisation, a level that echoes the dot-com era. This dynamic creates systemic risk. A reversal in AI investment flows, whether due to macroeconomic headwinds or a slowdown in adoption, could trigger a sharp correction not just in tech stocks, but across the broader market. The boom is being financed by capital that is already heavily allocated to a narrow set of winners, leaving less dry powder for diversification if the AI narrative falters.

The bottom line is that the semiconductor cycle has been reset by AI, but its long-term trajectory is now inextricably linked to the macro backdrop. The record equipment sales of the past year are a bet on sustained growth and low real rates. If those conditions change, the high concentration of both revenue and capital means the industry could face a sharper correction than in previous cycles. The current setup offers a powerful growth story, but it also embeds significant systemic risk that will be tested by the next shift in the macroeconomic cycle.

Valuation, Scenarios, and Key Catalysts

The path to a trillion-dollar industry is paved with both promise and peril. The bullish case is clear: AI infrastructure spending is creating a powerful, durable demand wave that is lifting the entire semiconductor stack. Yet this very strength introduces a classic cyclical risk-the potential for a sharp correction if demand slows or capacity outpaces it. The industry's valuation now hinges on navigating this narrow window between sustained growth and overbuild.

The primary risk is a demand correction. The boom is hyper-concentrated, with AI chips driving roughly half of total revenue but representing a tiny fraction of unit volume. This creates a system where a slowdown in AI adoption or a pause in corporate capital expenditure could quickly swing the market from shortage to glut. As one analysis notes, the sector's visibility gets murkier beyond the next year, even with order books looking full. The record equipment sales of the past year are a bet on sustained growth. If that growth falters, the high concentration of both revenue and capital means the industry could face a sharper correction than in previous cycles.

A key element of this setup is the expansion of advanced packaging capacity. This isn't a minor upgrade; it's a massive, multi-year build-out aimed at meeting the integration demands of AI chips. TSMC's CoWoS packaging lines, critical for connecting logic and high-bandwidth memory, are set to expand nearly fourfold from 2024 to 2026. This surge extends across the ecosystem, with foundries like Intel and Samsung and contract assemblers like ASE and Amkor committing tens of billions to scale. This capacity ramp is a direct response to the AI boom, but it also embeds the risk of overcapacity if the underlying demand for AI chips moderates.

The key catalysts to watch are the pace of AI infrastructure deployment and the execution of multi-year investment plans. For now, the forward view is supported by the massive, multi-year equipment investment plans by foundries, with sales of manufacturing equipment projected to surpass $150 billion by 2027. The industry's trajectory is mirrored by a dangerous concentration of capital. The bottom line is that the current setup offers a powerful growth story, but it also embeds significant systemic risk that will be tested by the next shift in the macroeconomic cycle.