Skip to content

Technology · Issue 01

The chip supercycle

Semiconductors did $792 billion in 2025. WSTS’s spring 2026 forecast put 2026 at $1.51 trillion — a 90% leap, almost all memory and AI.

TechnologyUpdated 2026-08-308 min read
04088151,2231,631proj.201920202022202320252026Total semiconductor sales
WSTS spring-2026 forecast for 2026 sales
$1.51 tn
2025 actual sales
$792 bn
2025 growth, strongest since 2021
+25.6%
Memory, 2026 forecast
>$800 bn

This is not the 2017 crypto-and-phone cycle

World semiconductor sales of $792 billion in 2025 were up 25.6% — the strongest year since the 2021 rebound — and the driver was AI, not a broad consumer restock. Nvidia’s growth, HBM allocations at SK Hynix, Samsung and Micron, and foundry utilisation at the leading edge all tell the same story: a few package types, bought by a few customers, at prices the rest of the industry cannot believe. Then WSTS’s spring 2026 revision took the 2026 outlook from a rich-but-plausible ~$975 billion to $1.51 trillion, with memory more than $800 billion and +250% year on year. If that lands, it is the largest nominal cycle in the history of the chip industry.

Memory is the tell

Logic gets the keynotes. Memory gets the scarcity. High-bandwidth memory stacked on advanced packaging is the binding constraint on accelerator shipments. A 250% jump in the memory segment is what a rationing market looks like in a revenue statistic. It is also what a bust looks like if the hyperscalers pause. The 2022–23 downturn is the scar: from $574 billion to $527 billion in a single year when phones and PCs coughed. This cycle is more concentrated, which means the upswing is more violent and the downswing, when it comes, will be too.

Geography is industrial policy

Taiwan still makes the leading-edge logic. Korea still makes the leading-edge memory. The United States still designs. China still spends to close the gap and still buys what it can. Export controls, subsidy races and ‘friend-shoring’ are now part of the cost of goods. A semiconductor thesis that ignores export licences and foundry geography is a 2015 thesis.

How to hold a supercycle without marrying it

The long view is that AI wants more compute per year for a long time — see the power chart. The trading view is that 90% revenue growth is not a multiple; it is a warning that expectations have no slack. Own the bottlenecks (HBM, leading-edge foundry, equipment, advanced packaging, the power that feeds them). Be humble about 2027. WSTS itself pencils another +27% to about $1.9 trillion in 2027. History says someone on that path is wrong.

Investing lens

Horizon 2–8 years · Educational, not advice

Stay with the physical bottlenecks of accelerated compute — HBM, advanced packaging, leading-edge foundry and the equipment that serves them — and size positions as if a mid-cycle pause is the base case, not a tail.

Where the map points

  • HBM producers with allocated 2026–27 capacity
  • Leading-edge foundry and EUV equipment
  • Advanced packaging and substrate suppliers
  • A smaller sleeve in power-management and cooling silicon

What can break it

  • A hyperscaler capex pause would hit this cycle harder than 2022 hit analog
  • Forecast revisions of this size can reverse
  • Export-control and China-demand shocks

CHART does not recommend securities, funds or trades. Figures can be revised by their publishers. Do your own research and consider regulated advice before allocating capital.

Sources

Every headline number traces to a named publisher. Contextual sources inform the essay, not the key stat.

  1. 01PrimaryWorld Semiconductor Trade Statistics2026
    Spring 2026 semiconductor market forecast

    2026 market projected at $1.51 trillion, +90%; memory >$800 billion.

  2. 02CorroboratedSemiconductor Intelligence / WSTS2026-03
    AI drives strong semiconductor market in 2025–2026

    2025 actual sales $792 billion, +25.6% — strongest since 2021.

101 graphics in this issue