Construction companies by year of registration
The register records a start date for every company. Grouped by year, those dates show how old the currently active population of companies is.
Broken down by the year they were entered on the register.
By year of registration
Companies still on the register today, by the year they were registered.
- Companies with legal personality
- Sole traders
Show the figures
| Period | Companies | Sole traders |
|---|---|---|
| 2007 | 3,911 | 1,423 |
| 2008 | 3,812 | 1,236 |
| 2009 | 3,565 | 1,153 |
| 2010 | 4,271 | 1,331 |
| 2011 | 4,723 | 1,574 |
| 2012 | 4,231 | 1,712 |
| 2013 | 4,254 | 1,877 |
| 2014 | 4,646 | 2,089 |
| 2015 | 5,056 | 2,208 |
| 2016 | 5,724 | 2,448 |
| 2017 | 6,335 | 2,784 |
| 2018 | 7,067 | 3,183 |
| 2019 | 8,868 | 4,049 |
| 2020 | 10,574 | 4,474 |
| 2021 | 12,836 | 5,419 |
| 2022 | 12,948 | 5,409 |
| 2023 | 14,290 | 6,043 |
| 2024 | 15,980 | 7,828 |
| 2025 | 18,026 | 9,455 |
| 2026 | 11,813 | 7,338 |
Common questions
- Why are the older years lower?
- Because all that is left of an older year is the companies still active. The slope you think you are reading is not only a change in how many companies were registered: it is also how many survived.
- Is the current year complete?
- No. It stops at the register read date shown on the page, so it is not comparable to a full year as it stands.
Method and sources
Counted here are the companies whose main activity code falls under NACE divisions 41, 42, 43 — construction — in version 2025 of the classification.
Register read on 13 Aug 2026. The register publishes one full extract per month; these figures change at that pace and no faster.
Any category describing fewer than 10 sole traders is taken out of the breakdown: a count small enough points at one person, and these pages point at nobody. Removed categories are reported under the table they came from.
Province and region are derived from the seat’s postcode: the register publishes a commune and a postcode, not a province. That derivation is ours.
This chart does not count registrations per year: it counts the companies registered in that year and still active today. Everything that has stopped since is missing from it, which hollows out the older years mechanically.
The register publishes a start date and no end date. A true series of yearly registrations would need one extract per year; this is the honest series we can build today.
Source: Crossroads Bank for Enterprises (KBO/BCE), FPS Economy — open data — re-use conditions
These figures describe what the public register said on the day we read it. They are counts, not a ranking: no commune, no trade and no company is presented here as better or worse than another.