For those interested in the mechanics of technological leaps, 1987 stands out as a pivotal year for the semiconductor industry. The so-called “semiconductor boom” wasn’t merely a matter of rising demand for chips; it was the result of massive capital investment, new manufacturing processes, and a reshuffling of the competitive landscape. Yet, for economists and policymakers hoping to measure and interpret this surge, the challenge lies in how to sift the story from the statistics—most notably through ISIC 2610, the code for the manufacture of electronic components.
ISIC 2610 is broad by design. It encompasses not only semiconductor foundries but also producers of resistors, capacitors, and other passive or active electronic parts. To focus on the semiconductor story, analysts need to filter this registry, often beginning with firm-level information from trade publications, corporate annual reports, and industry association directories. Foundries—the high-capital, silicon-wafer-fabricating heart of the industry—are usually visible in news coverage or government filings, as their expansions or new plant openings rarely went unnoticed. Major players in 1987 included not just US giants like Intel and Texas Instruments, but also rising Japanese, Korean, and European manufacturers.
Once the population of relevant firms is defined, attention shifts to capital investment. During the boom, announcements of new fabs (fabrication plants) or major facility upgrades were common, often accompanied by eye-catching dollar figures. Press releases, company reports, and government incentives data provide a stream of capital expenditure numbers—occasionally broken down by location, project, or even line of manufacturing (memory, logic, specialty chips). For some countries, industrial policy or investment tracking offices kept detailed ledgers of both foreign and domestic investment flows into the sector.
To link investment data to changes in production capacity, analysts can examine plant-level reports: how many wafer starts per month a new fab could deliver, or what additional output was promised from a major retrofit. Trade journals and analyst notes from the era sometimes recorded projected vs. actual capacity figures, and international industry associations like SEMI occasionally published aggregated capacity estimates by region or technology node.
Matching investment timelines with capacity jumps is rarely seamless. New plants often faced ramp-up delays, technology glitches, or market slumps that postponed full utilization. Likewise, currency swings, trade policy shifts (notably the US-Japan Semiconductor Agreement), or global economic headwinds could distort the expected correlation. Yet, by layering plant opening dates, reported investments, and subsequent increases in production figures (number of wafers processed, chips shipped), patterns begin to emerge.
In some cases, analysts will want to map these trends against employment, export volumes, or the emergence of contract manufacturing (foundries producing chips for fabless companies). These secondary signals—factory headcounts, customs data for chip exports, or customer lists—can provide additional validation or highlight divergences between investment and output.
Documentation is crucial throughout. Every assumption—about firm classification, investment estimation, capacity conversion, or production reporting—should be recorded, given the notorious opacity and volatility of the sector. Company names changed, plants were sold or repurposed, and reporting conventions evolved.
By bringing together ISIC 2610 firm registries, capital investment records, and granular production capacity data, analysts can sketch a data-driven portrait of the 1987 semiconductor boom. The specifics are inevitably fuzzy: capacity sometimes lagged behind the spending, and the payoffs of the boom were unevenly distributed across the globe. But in the broad sweep of the numbers, the intensity and direction of the technological leap—how money became machines, and machines became chips—becomes unmistakably clear.