Market research firm TechInsights expects the global semiconductor market to pass $2.6 trillion next year, with memory taking about 62% of that revenue, according to The Elec’s report on 30 October of TechInsights’ 2027 outlook (our translation from the Korean).
TechInsights forecasts memory revenue above $1.6 trillion in 2027, up from $740 billion this year, which The Elec describes as more than doubling. By our arithmetic, $1.6 trillion is about 2.2 times $740 billion.
Short explainer · Memory to Be 62% of a $2.6T Chip Market in 2027
A vertical news Short narrated from this piece’s headline and opening line.
Business Pill · OPERATING LEVERAGE
A short explainer on operating leverage, why profits swing more than revenue when costs are largely fixed. It teaches the general idea only and says nothing about this forecast.
The key insight: TechInsights forecasts memory at about 62% of a chip market above $2.6 trillion in 2027, with memory revenue more than doubling from $740 billion, about 2.2 times by our arithmetic.
The Elec reported that TechInsights published its 2027 outlook report on 29 October local time. It said the forecast rests on AI infrastructure investment pushing up prices for high-bandwidth memory (HBM), server DRAM and NAND flash, which are in short supply.
The Elec quoted TechInsights as saying it expects NAND to become the next memory bottleneck, as AI drives fast growth in demand for solid-state drive storage. It also quoted the firm as saying that TSMC’s CoWoS and SoIC advanced packaging capacity will grow 20 to 25% a year, and that demand will nonetheless stay strong.
David McQueen, a director at TechInsights, was quoted as saying that next year companies will have to think not only about which technologies to develop but about which memory they can secure, and that this will affect everything from chip design and capital spending to the consumer products that reach the market.
Memory revenue $740 billion in 2026 vs more than $1.6 trillion forecast for 2027, and a 2027 chip market of more than $2.6 trillion with memory about 62%. Source: TechInsights 2027 outlook as reported by The Elec, 30 Oct 2026 (our translation). 1.6/0.74 = about 2.2x, our arithmetic.
TechInsights’ own framing is that securing memory becomes a planning question alongside choosing technology. When one input is forecast to be most of an industry’s revenue, the price of that input is set by the suppliers who have it.
The Structural Read
As we read The Elec’s report, TechInsights ties the forecast to AI infrastructure demand meeting short supply of HBM, server DRAM and NAND, which pushes prices up.
The Elec quotes TechInsights as expecting NAND to become the next memory bottleneck as AI lifts demand for solid-state storage.
As we read it, TechInsights’ point that TSMC’s CoWoS and SoIC capacity grows 20 to 25% a year while demand stays strong places the packaging step alongside memory as a constraint.
David McQueen, TechInsights, as quoted by The Elec (Korean original)
FOR CHIP BUYERS TechInsights’ McQueen says companies will have to think about which memory they can secure, not only which technologies to develop, according to The Elec.
FOR MEMORY MAKERS TechInsights forecasts memory revenue above $1.6 trillion in 2027, up from $740 billion this year, according to The Elec.
FOR STORAGE The Elec quotes TechInsights as expecting NAND to be the next memory bottleneck as AI drives SSD demand.
The framework says “Memory is becoming the new compute. HBM growth is outpacing accelerator demand, making memory the rate-limiting factor in the AI scaling equation.”
As we read it, a forecast that memory becomes most of chip revenue is the scale version of that thesis. The framework predates TechInsights’ outlook and does not discuss it.
What Is Not Established
We read the forecast through The Elec’s report, not TechInsights’ report itself, and the translations are ours. The Elec does not give TechInsights’ 2026 total market figure or memory’s 2026 share, so this piece does not compute a change in share. The figures are forecasts, and the report’s methodology is not described in the article.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
TechInsights forecasts a chip market above $2.6 trillion in 2027 with memory at about 62% of revenue, and memory revenue above $1.6 trillion, up from $740 billion this year, according to The Elec. By our arithmetic, that is about 2.2 times this year’s memory revenue.
A note on sourcing. TechInsights’ forecast comes from The Elec’s report of 30 October 2026 in Korean; the English is our translation. We did not read TechInsights’ report itself or contact the firm. Figures labelled our arithmetic are our own.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
