The Super Cycle Has Arrived: AI, Memory, and the Race for Supply-Chain Resilience
An independent analysis and synthesis based on the 21st annual KPMG Global Semiconductor Industry Outlook, conducted in partnership with the Global Semiconductor Alliance (GSA).
Executive Summary
The global semiconductor industry has entered a genuine super cycle. After years of executives wondering whether the AI-driven boom would prove to be a bubble, the 2026 KPMG Global Semiconductor Industry Outlook concludes decisively that the super cycle has arrived — and that growth is broad-based, demand-led, and structural rather than speculative.
The 21st edition of the survey, fielded in Q4 2025 among 151 senior semiconductor executives (more than half from companies with annual revenues exceeding $1 billion), captures an industry in the midst of a profound transformation. The key developments are unmistakable:
- AI has become the industry’s dominant revenue engine. Fully 73% of executives now cite AI as their single largest revenue growth driver, up from 67% the prior year, surpassing cloud/data center (61%).
- Memory has risen to co-lead growth. 67% of executives identify memory (HBM, DRAM, NAND) as a top growth opportunity — up sharply from 49% — placing it for the first time alongside, and slightly ahead of, microprocessors (66%).
- Supply chain agility has become the top strategic priority. 45% rank improving supply-chain responsiveness and flexibility as their number-one strategic focus, overtaking talent cultivation for the first time in three years.
- AI is augmenting, not replacing, the workforce. 66% plan to use AI to raise employee productivity, and 65% expect to increase headcount in the coming year.
- Confidence is at a 21-year high point. The industry confidence index climbed from 59 to 63, the third-highest reading in the survey’s history, with a majority planning higher capital expenditure, headcount, and IT spending.
Yet the outlook is not without risk. Geopolitics, trade friction, and physical resource constraints — particularly energy and supply — loom over the industry. Success in 2026, the report argues, will belong to companies that combine flexibility, resilience, and AI-enabled talent and operations.
1. Introduction: A Tale of “Ice and Fire”
The 2026 outlook frames the industry against a sharply divided global economy — a world of “ice and fire.” On one side, an AI boom is channeling hundreds of billions of dollars into compute infrastructure. On the other, traditional manufacturing, real estate, and public budgets remain weak amid trade tariffs and sluggish demand.
The central question: How long can AI-driven spending last? The report marshals compelling evidence that demand is real, not a bubble:
- Data center utilization reached 92.4% in Q2 2025 — a record high — indicating that GPUs are genuinely being used, not sitting idle.
- 2025 saw more than 100 data center deals, representing $61.0 billion in compute infrastructure investment (up from $60.8 billion in 2024). Cumulative investment since 2019 reached $160 billion in the US/Canada, $40 billion in Asia-Pacific, and $24.2 billion in Europe across M&A, asset sales, and equity.
- Endpoint AI (phones, cars, factories) is forecast to generate $438.5 billion in AI semiconductor revenue by 2029, a compound annual growth rate (CAGR) of 25.9% (Gartner).
- Wafer fab equipment (WFE) spending reached $116 billion in 2025 (up 12% from $104 billion) and is forecast to hit $126 billion in 2026 (up 9%) — signaling a multi-year investment cycle.
The importance of AI as a revenue driver also varies by region: 81% of US executives, 70% of European executives, and 67% of Asia-Pacific executives rate it as a primary growth engine.
2. Key Finding 1: AI Is the Primary Revenue Growth Driver
AI has firmly displaced cloud and data center as the industry’s leading source of growth. In 2026, 73% of executives cite AI as their largest revenue driver (up from 67% in 2025), while cloud/data center/network slipped from 63% to 61%.
A detailed breakdown of how executives expect different application categories to change as revenue drivers for the next fiscal year underscores AI’s ascendancy. In the survey’s “AI / GenAI / Agentic AI” category, only 15% of executives expect a decrease in importance, 12% expect it to remain flat, and 73% expect it to increase — the strongest “increase” reading across all categories. By contrast, several traditional categories show softer momentum:
| Application category | Decrease | Flat | Increase |
|---|---|---|---|
| AI / GenAI / Agentic AI | 15% | 12% | 73% |
| Clean technology | 26% | 13% | 61% |
| Cloud / data center / network | 27% | 17% | 56% |
| Wired communications | 44% | 25% | 27% |
| Wireless communications (5G) | 41% | 26% | 33% |
| IoT | 46% | 27% | 27% |
| Consumer electronics | 38% | 25% | 38% |
| Automotive | 30% | 44% | 38% |
| Industrial equipment | 29% | 50% | 26% |
| Personal computing | 26% | 21% | 23% |
(Note: figures are as reported in the survey appendix and rounded.)
The long-run trajectory is equally compelling: the AI semiconductor market is forecast to reach $438.5 billion by 2029, growing at a 25.9% CAGR (Gartner), reinforcing AI’s position as the defining demand engine of this cycle.
3. Key Finding 2: Memory and Microprocessors Become Joint Top Growth Opportunities
For the first time in the survey’s 21-year history, memory has risen to co-lead — and narrowly lead — the industry’s growth agenda. 67% of executives now name memory (HBM, DRAM, NAND) as a top growth opportunity over the next 12 months, a dramatic jump from 49% the prior year, while 66% name microprocessors (GPU, MCU, MPU).
This shift is driven directly by AI compute demand and the rise of High Bandwidth Memory (HBM). Hyperscalers — Amazon, Google Cloud, and Microsoft Azure — are shifting their architectures from conventional DRAM toward HBM, and some DRAM manufacturers are retooling capacity toward HBM production, in turn creating scarcity in conventional DRAM.
The full picture of growth opportunities for the next 12 months (share of executives expecting higher growth):
| Product category | Lower | Flat | Higher |
|---|---|---|---|
| Memory (HBM / DRAM / NAND) | — | — | 67% |
| Microprocessors (GPU / MCU / MPU) | — | — | 66% |
| Analog / RF / mixed-signal | 32% | 52% | 66% |
| Optoelectronics | 28% | 36% | 50% |
| Discrete / passive components | 44% | 45% | 45% |
| Other logic devices | 11% | 8% | 36% |
| Sensors / MEMS | 19% | 18% | 13% |
(Figures reflect the survey appendix; memory and microprocessor “higher” readings are the headline 67% and 66%.)
The strategic implication is clear: architecture choices will define market position. Companies that win will align product roadmaps with AI compute and HBM demand, while an emerging energy constraint may prove as disruptive to supply as any raw-material shortage.
4. Key Finding 3: Supply Chain Becomes the Top Agenda
For the first time in three years, supply chain agility has overtaken talent as the industry’s number-one strategic priority. 45% of executives rank “enhancing supply chain response speed and flexibility” as their top strategic focus (excluding revenue growth), compared with 38% for talent supply, cultivation, and retention.
This pivot reflects a sharpening awareness of geopolitical and physical constraints:
- Geopolitical uncertainty is now seen as the third-biggest industry problem (of 11), cited by 37% of executives.
- Energy is an emerging bottleneck. 34% worry about securing enough power for their own fabs, while a striking 58% fear hyperscalers cannot source enough energy to support AI infrastructure.
- Tariffs are the top geopolitical concern, cited by 73% of executives.
In response, executives are diversifying and digitalizing:
- 54% plan to expand regional distribution of their supply chains within the next 12 months.
- 36% will deploy generative AI in procurement and supply chain for predictive analytics and automation.
- The US and EU emphasize domestic production and reshoring (driven by the CHIPS Act and European Chips Act), while Asia-Pacific focuses on building regional redundancy.
The public-private dynamic is increasingly complex. 48% worry about the impact of government subsidies, 54% agree that government shareholding limits market flexibility and innovation, and 54% believe government investment and equity are nonetheless necessary to build the most advanced fabs.
Top three strategic priorities (excluding revenue growth):
| Priority | Share |
|---|---|
| Supply chain flexibility | 45% |
| Talent supply / cultivation / retention | 38% |
| Transformative M&A | 37% |
| Implement generative AI | 34% |
| Digital transformation of systems/processes | 28% |
| Implement agentic AI | 25% |
| Inflation / tariff price renegotiation | 17% |
5. Key Finding 4: AI Empowers Rather Than Replaces Employees
A defining theme of the 2026 outlook is that AI is being deployed as an enabler of the workforce, not a replacement for it. Two-thirds of executives view AI as a tool to augment employees, and 41% identify talent as the second-biggest challenge over the next three years — underscoring that people remain central even in an AI-driven era.
AI deployment by function shows a clear pattern of early adoption in high-value technical and operational areas:
| Function | Unclear | Implemented | Will implement in 12 months |
|---|---|---|---|
| R&D / Engineering | 22% | 61% | 17% |
| Finance & Accounting | 30% | 53% | 17% |
| Risk & Compliance | 31% | 50% | 19% |
| HR | 23% | 49% | 28% |
| Manufacturing & Operations | 36% | 49% | 15% |
| Customer Support | 32% | 40% | 27% |
| IT | 28% | 28% | 44% |
| Marketing & Sales | 32% | 35% | 33% |
| Procurement & Supply Chain | 32% | 32% | 35% |
Talent remains a top concern — 50% of executives name talent risk (shortage and competition) as the biggest industry problem over the next three years. To secure talent, executives plan a multi-pronged approach:
| Talent measure | Share |
|---|---|
| Use tech / AI / automation to raise productivity | 54% |
| Training and upskilling | 46% |
| Retention and engagement | 32% |
| Succession and leadership development | 31% |
| Competitive compensation / flexible work | 28% |
| Recruit new specialists | 26% |
| Education partnerships | 16% |
Notably, 65% of executives plan to increase headcount in the coming year — a powerful signal that AI investment is expanding, rather than shrinking, the human workforce.
6. Key Finding 5: Optimists Advance Despite Risks
The industry’s confidence index climbed from 59 to 63, the third-highest reading in the survey’s 21-year history. This optimism is grounded in strong market fundamentals:
- The global semiconductor market is forecast to reach approximately $1 trillion in 2026, with WSTS data pointing to $975 billion, led by logic and memory.
- Semiconductor sales have risen roughly 50% since the end of 2022.
- 54% of executives expect revenue growth of 11% or more.
- 66% expect to increase capital expenditure (up from 63% last year).
- More than half plan to hire, upgrade IT systems, and pursue M&A.
At the same time, 58% expect customer demand uncertainty to have a major impact, and executives remain wary of pressure points in rare earth minerals, chemicals, and energy.
Global semiconductor market size (US$ billions) and confidence index:
| Year | Market size ($B) |
|---|---|
| 2013 | 305 |
| 2014 | 335 |
| 2015 | 335 |
| 2016 | 339 |
| 2017 | 412 |
| 2018 | 469 |
| 2019 | 412 |
| 2020 | 440 |
| 2021 | 556 |
| 2022 | 574 |
| 2023 | 527 |
| 2024 | 631 |
| 2025 | 772 |
| 2026 (forecast) | 975 |
(Confidence index: 2024: 59, 2025: 59, 2026: 63.)
7. Key Recommendations
The KPMG report closes with four strategic imperatives for semiconductor leaders navigating this super cycle.
7.1 Capture Memory and Microprocessor Parallel Growth
- Invest in HBM R&D and manufacturing, focusing on speed, efficiency, and thermal performance.
- Partner with hyperscalers (Amazon, Google Cloud, Microsoft Azure) to align roadmaps and secure long-term contracts.
- Differentiate beyond DRAM on performance and sustainability.
- Secure energy supply through joint planning with utilities and renewable providers, and consider co-owning energy infrastructure.
- Use AI to optimize energy consumption per unit of compute, both in internal fabs and customer solutions.
7.2 Enhance Supply Chain Flexibility Against Geopolitical Turbulence
- Engage governments strategically, negotiating flexibility while accepting subsidies.
- Disperse sourcing and regions, targeting a 50% or greater increase in distribution over three years to reduce Taiwan concentration.
- Digitalize and automate the supply chain using AI predictive analytics.
- Balance reshoring (US/EU) with complementary global sourcing.
- Model tariff impacts — with 73% citing tariffs as the top geopolitical issue, build flexible pricing and operational plans.
7.3 Use AI to Boost Talent Resilience and Productivity
- Deploy AI in IT, R&D, procurement, and supply chain to accelerate decision-making.
- Use AI to eliminate repetitive tasks, improving employee satisfaction and retention.
- Deliver continuous AI literacy training, covering generative AI and agentic AI, cross-functionally and ethically.
- Apply agentic AI in marketing and sales for personalization, forecasting, and shorter time-to-market.
7.4 Sustain Strategic Growth Amid Uncertainty
- Focus capital expenditure on high-growth logic and HBM aligned with AI and data center demand.
- Model multiple demand scenarios and tie hiring, IT, and M&A to triggers rather than fixed targets.
- Invest early in advanced wafer capacity, advanced packaging, and rare earth/chemical agreements (66% plan capex increases).
- Pursue M&A to acquire AI technology, regional manufacturing, and differentiated capabilities.
- Maintain liquidity and control margins, preserving the ability to slow investment if geopolitical or demand conditions deteriorate.
8. Conclusion
The 2026 KPMG Global Semiconductor Industry Outlook delivers an unambiguous verdict: the semiconductor super cycle is real, and it is being driven by genuine AI demand across compute, memory, and networking chips. With 73% of executives identifying AI as their top revenue driver, memory rising to co-lead growth for the first time in history, and confidence at a 21-year high, the industry is positioned for sustained expansion — with the global market approaching $1 trillion.
Yet the path forward demands more than optimism. The industry must navigate supply chain concentration, energy scarcity, geopolitical friction, and an increasingly complex relationship with governments. The winners will be those who convert this moment of abundance into durable advantage — through supply chain agility, an AI-enabled workforce, and disciplined, scenario-driven risk management.
Appendix: Summary of Key Survey Statistics
| Metric | Value |
|---|---|
| Survey respondents | 151 senior semiconductor executives |
| Companies with revenue > $1B | >50% |
| Survey period | Q4 2025 |
| AI as top revenue driver | 73% (up from 67%) |
| Memory as top growth opportunity | 67% (up from 49%) |
| Microprocessors as top growth opportunity | 66% |
| Supply chain flexibility as #1 priority | 45% |
| Plan to use AI to raise productivity | 66% |
| Plan to increase headcount | 65% |
| Plan to increase capex | 66% (up from 63%) |
| Confidence index | 63 (up from 59) |
| Expect revenue growth ≥ 11% | 54% |
| 2026 global market forecast (WSTS) | $975B (~$1 trillion) |
| AI semiconductor market by 2029 (Gartner) | $438.5B (25.9% CAGR) |
| WFE spending 2025 / 2026F | $116B / $126B |
| Data center utilization (Q2 2025) | 92.4% |
| 2025 data center compute investment | $61.0B |
| Executives citing tariffs as top geo issue | 73% |
| Worried hyperscalers can’t secure energy | 58% |
| Talent risk as biggest industry problem | 50% |
Source: KPMG & Global Semiconductor Alliance, “2026 Global Semiconductor Industry Outlook” (21st annual edition). External data referenced from Gartner, WSTS, SEMI, and other cited sources. All percentages rounded; survey fielded Q4 2025.
