The Semiconductor Industry Association reported on 6 August that global semiconductor sales reached $403.3 billion in the second quarter of 2026 — 35.1% above the first quarter, and 123.6% above the same period a year earlier. June alone accounted for $134.5 billion.
Figures that large are easy to skim past, so it is worth restating plainly: the industry roughly doubled in a year. SIA president John Neuffer expects full-year sales to exceed $1.5 trillion. The growth was not evenly spread — year on year, sales into the Americas rose 160.9%, Asia Pacific 124.4%, China 112.8%, Europe 75.2% and Japan 39%.
For anyone specifying components, the interesting part is not the headline but the queue behind it. When a market moves this fast, wafer capacity, substrates, packaging and test all become contested at once. Allocation comes back, lead times stretch unevenly across a bill of materials, and the part that holds up a build is rarely the expensive one. It is usually an ordinary regulator, crystal or connector that nobody thought to check.
The sensible response at design stage is unglamorous. Second-source what can be second-sourced, avoid sole-supplier parts in roles where a functional equivalent exists, and confirm availability before the schematic is frozen rather than after the first build fails to kit.