Industry hub · Semiconductors data provider

Semiconductors Data Providers

Datadory covers semiconductors data provider with 8 datasets spanning Eurostat Data Browser — Apparel, Textiles & Trade (PRODCOM, COMEXT) , Data.gov Catalog - Telecommunications & Carrier Datasets and U.S. Census Bureau Data API - CBP / Economic Census / Wholesale Trade . Delivered daily, weekly, or hourly — your call.

8 datasets 3 publishers Delivery: API, files, or your warehouse. Daily, weekly, or hourly.

The manifest

Every catch in this slice, ranked by our quality rubric — depth of documentation, freshness, breadth. Pick one, sample it, ship it.

#1 Eurostat (European Commission)

Eurostat Data Browser — Apparel, Textiles & Trade (PRODCOM, COMEXT)

Coverage
EU-27, euro area
#2

Data.gov Catalog - Telecommunications & Carrier Datasets

Coverage
United States - federal publishers plus state, county…
#3

U.S. Census Bureau Data API - CBP / Economic Census / Wholesale Trade

Coverage
United States: national, states, counties, metropolitan…
#4 United Nations Statistics Division

UN Comtrade Plus TradeFlow - Electrical Machinery & Instruments (HS 85/90)

Coverage
Global
#5 World Semiconductor Trade Statistics (WSTS)

WSTS World Semiconductor Trade Statistics - Historical Billings & Forecasts

Coverage
Global - Americas
#6

OECD.Stat — Retail, Industry & Trade Indicators

Coverage
OECD members plus partner economies and aggregates

Plus 2 more in this slice — each with its field dictionary and sample rows on its own page.

The lay of the water

Where semiconductors data provider data actually comes from.

Semiconductors runs as a proxy-built slice in Datadory's catalog: the industry contributes zero primary records of its own and pools eight datasets tagged in from adjacent industries.

That sounds thin until you read what the proxies carry - an industry-native billings series, three independent views of chip trade, a national-competitiveness panel, distributor firmographics to the ZIP code, and a million-strong corpus of training sets.

The semiconductor-native anchor is the WSTS World Semiconductor Trade Statistics Historical Billings & Forecasts record.

Two sheets of roughly 250 rows each hold about 8,000 numeric observations spanning 1986 through June 2026: monthly billings and three-month moving averages for Americas, Europe, Japan, Asia Pacific and Worldwide, stated in thousands of US dollars.

Worldwide billings ran roughly $152 billion in June 2026, and forecast releases extend several years forward, which makes the series both the longest and the most forward-looking demand input in this pool.

Chip trade arrives three ways.

UN Comtrade Plus TradeFlow covers HS chapter 85 at six-digit depth - 8541 diodes and transistors and 8542 integrated circuits - reported bilateral, reporter by partner, from 1962 onward, with values, net weight, quantity and CIF/FOB splits per flow; the HS 85 and 90 chapters alone span millions of records.

Eurostat's Comext files add EU-level trade from January 1988 to the latest reference month (verified to June 2026), including dataset DS-059322, EU trade since 2002 by CN8, alongside Europroms production series.

OECD.Stat contributes balanced bilateral trade at HS 2-digit, 4-digit and 6-digit depth plus STAN industry accounts covering multiple decades.

National competitiveness has a ready stand-in: the World Bank's TX.VAL.TECH.MF.ZS series measures high-technology exports as a share of manufactured exports across 200-plus economies from 2007 to 2024, with a companion current-dollar series, TX.VAL.TECH.CD.

The channel layer is American and granular.

County Business Patterns inside the U.S.

Around those sit the Data.gov Catalog's 552,271 federal records (including FCC broadband deployment and equipment-authorization material) and the Hugging Face Datasets Hub, whose 1,012,353 hosted datasets supply ML corpora and exogenous features.

Seven of the eight records arrive as structured feeds Datadory normalizes into one shape; only the WSTS workbook ships as a plain spreadsheet.

Either way, what you receive is rows - see the full semiconductors data guide for worked examples.

Pick your catch

Three to start with. The rest of the board is above.

Semiconductors data provider EU-27, euro area · 1988-01/2026-06

Eurostat Data Browser — Apparel, Textiles & Trade (PRODCOM, COMEXT)

geo · time · nace_r2 …+4 more

Semiconductors data provider United States - federal publishers… · Historical series reaching back to…

Data.gov Catalog - Telecommunications & Carrier Datasets

dcat.title · dcat.description · dcat.publisher.name …+11 more

Semiconductors data provider United States: national, states… · County Business Patterns annually…

U.S. Census Bureau Data API - CBP / Economic Census / Wholesale Trade

GEO_ID · NAME · YEAR …+21 more

Semiconductors data provider Global · Annual series from 1962 onward…

UN Comtrade Plus TradeFlow - Electrical Machinery & Instruments (HS 85/90)

typeCode · freqCode · period …+13 more

Semiconductors data provider Global - Americas · Monthly and quarterly billings 1986…

WSTS World Semiconductor Trade Statistics - Historical Billings & Forecasts

Year · Month (January-December) · Q1-Q4 …+7 more

Semiconductors data provider OECD members plus partner economies… · Annual industry accounts spanning…

OECD.Stat — Retail, Industry & Trade Indicators

REF_AREA · ACTIVITY · MEASURE …+5 more

Semiconductors data provider Global · Continuously growing repository

Hugging Face Datasets Hub - Fashion/Apparel Datasets

id · downloads · lastModified …+5 more

Semiconductors data provider 200+ World Bank member economies and… · 2007-2024 (published year selector)

World Bank High-Technology Exports Indicator

indicator.id · indicator.value · country.id …+11 more

Who fishes here

Market researchers size the US distribution channel before sizing the market: County Business Patterns cells under NAICS 423430 give establishment and employment counts by county and ZIP, and the 2022 Economic Census subject series supplies the margin structure those territories have to clear.

Investors and quants treat the WSTS billings series as the demand input it is - forty-one years of monthly regional revenue feeding revenue forecasts and cycle-timing models - then join Comtrade's 8542 flows to trace where supply actually crosses borders. Best data for semiconductors teams collects that shortlist.

Data scientists assemble features from both ends: country-year panels such as the World Bank high-technology exports share join cleanly against Parquet-sharded training corpora from the Hugging Face Hub, and every series in this pool lands long-format, one row per observation.

See best data for semiconductors teams for the pipeline-ready picks.

Journalists, academics and students cite the institutional publishers - WSTS, UN Comtrade, Eurostat, OECD, World Bank, Census - because every figure resolves to a named record with stated coverage. Best data for semiconductors teams holds that citation-grade set.

Choosing between them

Three attributes decide it in this pool.

Grain: the WSTS record runs region-by-month; Comtrade and Eurostat run reporter-by-partner-by-product; Census runs industry-by-geography; the World Bank panel runs country-year.

Match the unit of analysis first - a worldwide billings figure and a ZIP-code establishment count are answers to different questions.

Depth: trade series reach furthest back - Comtrade opens in 1962, Eurostat's Comext files in January 1988, Census County Business Patterns in 1986, WSTS billings in 1986.

The competitiveness panel begins in 2007.

Cycle historians and backtesters want the left tail; forecasters want WSTS's forward releases.

Shape: the workbook flatters analysts who live in spreadsheets; the long-format panels flatter warehouses and notebooks; the federal catalog is document-shaped metadata that flatters search.

Datadory normalizes all of it into one feed, so the choice is about content, not plumbing.

Where records overlap, take Eurostat for EU member-state detail, OECD for cross-country balance, and Comtrade for everything bilateral beyond Europe.

Pick any card above, request a sample, and the same fields described here arrive as rows in your inbox - browse the semiconductors datasets to start.

Frequently asked questions

Which dataset tracks the semiconductor billing cycle by region?
WSTS World Semiconductor Trade Statistics - Historical Billings & Forecasts. Two sheets hold monthly billings and three-month moving averages for Americas, Europe, Japan, Asia Pacific and Worldwide from 1986 through June 2026, in thousands of US dollars - roughly $152 billion worldwide that month - with quarterly subtotals, yearly totals and forecast releases extending several years forward.
Where do I find integrated circuit import and export figures between countries?
UN Comtrade Plus TradeFlow - Electrical Machinery & Instruments (HS 85/90) carries HS 8542 integrated circuits and 8541 diodes and transistors bilaterally, reporter by partner, from 1962 onward. Each flow records trade value, net weight, supplementary quantity and CIF/FOB splits, annually or monthly, covering essentially all trading economies.
How do I compare high-technology exports between countries?
Use the World Bank High-Technology Exports Indicator: series TX.VAL.TECH.MF.ZS expresses high-tech exports as a percentage of manufactured exports for 200-plus economies from 2007 to 2024, and companion series TX.VAL.TECH.CD states the same in current US dollars. One observation per economy per year makes it the standard cross-country benchmark.
What European coverage exists for electronics and component trade?
Two institutional warehouses. Eurostat's Comext files run EU trade by CN8 from January 1988 to the latest month, verified current to June 2026, with Europroms production alongside. OECD.Stat balances bilateral trade at HS 2-digit, 4-digit and 6-digit depth and layers STAN industry accounts over multiple decades - roughly 1,546 dataflows across the platform.
How far back does semiconductor industry data go?
Deeper than most hardware categories. UN Comtrade's chip lines start in 1962, Eurostat's Comext files in January 1988, WSTS regional billings and Census County Business Patterns both in 1986, and the World Bank high-technology series in 2007 - so a single merged panel can cover six decades of the industry.
Can these datasets support machine-learning models?
Yes, from two directions. The Hugging Face Datasets Hub hosts 1,012,353 datasets with declared schemas, shard-streamed as Parquet and filterable to permissive per-dataset terms, while the macro side - WSTS billings, Comtrade flows, World Bank shares - joins as exogenous demand features in clean country-year or region-month format.
How current is the data in this pool?
Reference periods run close to the present where it matters: WSTS billings through June 2026, Eurostat series verified to June 2026, World Bank indicators to 2024, County Business Patterns to its 2023 vintage, and Comtrade monthly or annual periods depending on the classification revision you select.

See the rows before you commit

Any catch on this board, sampled against your own question. API, files, or your warehouse. Daily, weekly, or hourly..