biannual industry forecast

The biannual industry forecast in the semiconductor world is the WSTS consensus, published through SIA twice a year. Each release states expected market size for three forward calendar years, so at any moment three overlapping forecast vintages are live.

One cataloged record anchors the category: the SIA Semiconductor Market Data & Factbook, described as carrying 'biannual forecasts from 2000 covering three forward calendar years'. That single line fixes the two facts buyers need - cadence and horizon. It also implies an archive structure: because each year contributes two releases, a forecast-history file grows by two vintages annually, and comparing them is how analysts measure how wrong the industry was and when. The Factbook pairs the forecasts with actuals, which is what makes vintage comparison possible; the related sia-factbook and wsts-market-data entries cover those companion series.

Forecast vintages age fast, and using the wrong one quietly changes conclusions. A capacity plan built on the spring release can be invalidated by the autumn revision of the same year; citing 'the forecast' without its month invites exactly that confusion.

Three buyer consequences follow. First, horizon discipline: each release covers three forward calendar years, so year-three projections carry compounding uncertainty that year-one figures do not - treat them as different products despite sharing a document. Second, revision tracking is the real value: with releases twice a year since 2000, the interesting dataset is the sequence of revisions, not the latest headline. Third, sourcing clarity matters because WSTS produces and SIA publishes - a team quoting 'SIA forecasts' and one quoting 'WSTS forecasts' are usually citing the same numbers, but only if both trace to the same release month.

1. Pin the release month, not just the year. With two releases annually since 2000, a citation missing its month cannot be verified against the archive.

1. Respect the three-year horizon boundary. The record specifies three forward calendar years per release; any projection past that edge is extrapolation, not sourced forecast.

1. Compare vintages deliberately. Pulling the same target year from successive releases shows how the forecast moved - the core analytic use of a biannual archive.

1. Pair forecasts with actuals before scoring accuracy. The Factbook supplies both sides; scoring a vintage against later WSTS/SIA actuals is the only fair test.

Entries covering the companion semiconductor market series:

• sia-factbook
• wsts-market-data