Australian companies by industry and state: 1.58 million sector-tagged entities

Two datasets, two different questions

"How many Australian companies are in industry X" and "how many Australian companies are in state Y" turn out to be answered by two different underlying datasets, not one. A company database of 3,992,217 rows carries an industry sector tag on a subset of its records. A separate Australian Business Register extract of 4,770,875 entities carries a complete state field but no industry field at all. Neither one answers both questions at once.

Presenting both on one page, with the coverage gap in each stated plainly, is more useful than picking whichever number looks bigger and calling it "Australian companies by industry and state" without the caveat.

The reason a single combined dataset does not exist yet comes down to how each was built. The register extract is a direct copy of a government dataset, filtered but not reclassified; the company database is built and maintained through a separate process that tags industry sector as records are processed, and that tagging has reached a subset of the total rows rather than all of them so far.

Industry: what "sector-tagged" means here

Of 3,992,217 total company rows in this database, 1,578,291 carry a known industry sector, meaning the sector field is populated and not marked "uncategorised." The remaining 2,413,926 rows, just over 60% of the total, have no usable sector tag. Any industry breakdown drawn from this dataset is therefore a breakdown of the 1,578,291 tagged rows, not of the full company population, and should be labelled that way rather than presented as covering every row.

3,710,066 of the 3,992,217 total rows carry an active status, separate from whether a sector is known. Sector coverage and active status are two different fields; a row can be active with no sector tag, or inactive with one, and the counts below apply to sector-tagged rows regardless of active status.

Two further fields are worth naming as absent rather than assumed present: this database carries no city field and no employee-count or revenue field for any of its 3,992,217 Australian rows, all confirmed empty on direct inspection. A county-level field is populated for 203,872 rows, but city-, headcount- and revenue-level breakdowns, available for this vendor's Lithuanian and Polish coverage, do not currently exist for Australia.

Top 10 industries by company count

Finance & Insurance and Professional Services together account for more than half of every sector-tagged row, 845,906 of the 1,578,291 total. That concentration is worth noting before drawing any conclusion about which industries are "biggest" in Australia overall: it may reflect how the sector-tagging process itself classifies ambiguous company names, rather than the true industry composition of the full 3,992,217-row population.

Below the top two, the drop-off is sharp rather than gradual. Construction & Trades, the third-largest tagged sector at 158,476, is less than a third the size of Professional Services and less than a third of Finance & Insurance. From there down to Wholesale & Distribution at 34,586, the remaining eight sectors in the top 10 sit within a comparatively narrow band of each other, none dominating the way the top two do.

State: the register view

The Australian Business Register extract used for the state breakdown is a separate 4,770,875-row snapshot, taken 3 September 2026, filtered to active, non-individual entities. It carries a state field on every row but no industry classification at all, which is why the industry and state breakdowns on this page come from two different sources rather than one combined cut.

Combining a state figure from this register with an industry figure from the company database, to estimate for example "construction companies in Queensland," would mean multiplying counts from two populations that are not the same size or built the same way. Either figure is useful cited on its own; combined without a stated method, the result would not be a real number.

A genuinely combined industry-by-state cut for Australia, built from one consistent source rather than two, is not something this page claims to offer. Stating that gap directly is preferable to quietly estimating one and presenting it as measured.

State breakdown table

New South Wales and Victoria together account for close to 61% of every entity on the register, with 18,075 further rows, 0.38% of the total, carrying no recorded state at all. The concentration broadly mirrors Australia's population distribution, though this table measures registered entities, not population, and the two do not move in exact proportion.

The remaining three jurisdictions, the Australian Capital Territory, Tasmania and the Northern Territory, together hold 162,702 entities, 3.41% of the register combined, less than a tenth of what Queensland alone holds. A state-by-state outbound plan built purely from raw counts will spend most of its effort in New South Wales, Victoria and Queensland almost by default, and that is worth deciding on purpose rather than discovering after a list is already built.

Why the two totals do not match

3,992,217 company rows against 4,770,875 register entities is not a data error; it is two different collection methods measuring an overlapping but non-identical population. The register extract is a direct snapshot of every active, non-individual ABN holder, including discretionary trusts and self-managed super funds that are not operating businesses in the ordinary sense. The company database is built and maintained separately, with its own inclusion criteria and update cadence, and does not attempt to mirror the register row for row.

Neither figure should be quoted as "the" number of Australian companies without naming which dataset it came from. A reader combining the two, for example by trying to compute an industry share of the full register, would be dividing numbers from different populations, and the result would not mean what it appeared to mean.

This is a common trap in any market with more than one public or first-party dataset available, not one specific to Australia. The fix is always the same: name the source next to the figure, state its snapshot date, and resist the pull to quietly reconcile two numbers that were never measuring the identical thing in the first place.

What this data can and cannot tell a buyer

Both datasets can support real targeting work: the industry table narrows a list to a sector, the state table narrows it to a geography, and the two can be applied to a list separately even though they come from different sources. What neither dataset contains, and what this page does not claim, is any job-posting or hiring-signal coverage for Australia.

Two separate vacancy and hiring-signal databases used elsewhere in this outbound method hold zero Australian rows between them. One, a 33,367-row vacancy database spanning postings from 2015 to August 2026, covers Germany (14,267 rows), Poland (8,146), the United Kingdom (4,154), France (3,877), the Netherlands (1,634), Switzerland (1,247), Lithuania (34) and Estonia (8); Australia does not appear at all. A second, a 15,985-row Adzuna-sourced job-posting database covering postings tracked since mid-2026, holds records for Poland (10,586) and Germany (5,399) only; Australia again does not appear.

Neither absence is a gap waiting to be filled by this page; it is a fact stated plainly so a reader comparing datasets knows exactly what each one does and does not cover. Any claim of Australian hiring-signal coverage, from any vendor, is worth checking directly rather than assumed from the strength of its company data alone.

A strong company database and a working hiring-signal feed are two different products, built from two different kinds of source. A vendor can have one without the other, and this page is a direct example of exactly that: real, sourced Australian company data by industry and by state, with no hiring-signal claim attached to it.

A buyer evaluating vendor claims on this specific point has a straightforward test available: ask for one example of an Australian job posting or headcount-growth signal the vendor's own system surfaced, with a date. A vendor with genuine coverage can produce one immediately; a vendor without it cannot, no matter how the rest of its pitch is worded.