4.77 million Australian ABN entities, by legal form and age

What 4,770,875 actually counts

The Australian Business Register issues the Australian Business Number to any entity dealing with government: companies, sole traders, partnerships, trusts and a range of narrower structures. A snapshot of the register taken 3 September 2026, filtered to active status and to non-individual entity types (excluding sole traders operating under their own name), held 4,770,875 rows.

That filtering choice matters for what the number means. It excludes individual sole traders entirely, so it undercounts "every Australian business" in the broadest sense, while it includes structures such as self-managed super funds and discretionary trusts that are not trading businesses in the way the phrase is usually meant. The number is best read as "active, non-individual entities on the ABN register," not as a company count on its own.

The schema behind this figure is deliberately narrow: each row carries an ABN, an entity-type code, a state, a postcode and a start date, and nothing else. There is no GST-registration field in this extract at all, so a claim about what share of the register is GST-registered cannot be answered from this dataset, and should be treated as unavailable rather than estimated.

Legal form: five buckets, one big caveat

Grouping the register's entity-type codes into five buckets produces the following split, ranked by count.

The "other" bucket needs its own sentence rather than a footnote. Of its 1,365,394 rows, 707,573 are self-managed super funds, a retirement structure rather than an operating business. Any reading of the register that treats the "other" bucket as businesses, without naming that a majority of it is superannuation structures, overstates the addressable market for anything commercial.

Private companies are the largest single bucket by a wide margin, and the group most likely to be the actual target of a B2B outbound campaign, since discretionary trusts and self-managed super funds rarely run day-to-day commercial operations under their own ABN.

Public companies, at 29,648 entities, are by far the smallest of the five buckets, well under 1% of the total register. That scarcity is expected rather than a data quality issue: the public-company structure carries reporting obligations that most Australian businesses, even large ones, have no reason to take on, and it remains the exception across every state the register covers.

Partnerships, at 364,188 entities, sit between the trust and public-company buckets in size. A meaningful share of these are professional-services partnerships, structures common among law firms, accounting practices and similar advisory businesses, though the register itself does not break the partnership bucket down further by profession.

Company age: how old is the ABN register

Age, calculated from each entity's ABN status start date against the 1 September 2026 snapshot date, splits into five bands.

Every one of the 4,770,875 rows carries a valid, 8-digit start date, so there is no "unknown" band to account for separately. The distribution skews toward older entities: nearly a quarter of the register is over 20 years old, and the two oldest bands together, 20-plus years and 10 to 20 years, account for close to half of all active entities. Read alongside the legal-form split, that suggests a register with a large base of long-standing structures rather than one dominated by recent registrations, though "structure" here again includes trusts and super funds alongside genuine long-running companies.

The under-2-years band, 741,750 entities, is still the largest single age band on its own even though the two oldest bands outweigh it combined. That is a useful correction to any assumption that new registrations are a small share of the register: new entities continue to be added at real volume even against a base this large and this old.

State by legal form: where each structure concentrates

Cutting legal form by state shows the same private-company dominance holding in every jurisdiction, with discretionary trusts making up a noticeably larger share in Victoria than in New South Wales.

Victoria's discretionary-trust share, 276,964 against 552,793 private companies, is proportionally the largest of the five states shown, one trust entity for every two private companies, against one for every 4.5 in New South Wales. Neither figure explains why on its own; it is a pattern worth noting rather than a conclusion the register data supports by itself.

Where the entities cluster by city

Mapping postcodes on the register to their labour-market region shows the same concentration the state table implies, down to the city level. Sydney leads with 1,140,498 entities, followed by Melbourne at 980,865, together accounting for close to 45% of the entire 4,770,875-entity register on their own.

The next two entries outside the top eight, Moreton Bay and the Mornington Peninsula, sit at 66,776 and 66,577 respectively, both regional-adjacent areas around Brisbane and Melbourne rather than a state capital in their own right, a reminder that entity density does not stop cleanly at a capital city's boundary.

These city figures are derived by joining each register row's postcode against a labour-market region label, then excluding any result that resolves to the full state name or an "Other Territories" catch-all rather than a genuine city or regional area. That exclusion is why the city figures below the top entries do not sum back to the state totals in the earlier table; the two tables are answering related but distinct questions, one about administrative state boundaries and one about labour-market regions.

Reconciling against the ABS actively-trading figure

The Australian Bureau of Statistics separately reported 2,814,778 actively trading businesses as at 30 June 2026, a figure built from its own business register criteria rather than a raw ABN extract. That number sits well below the 4,770,875 active ABN entities described above, and the gap is a methodology difference rather than a contradiction.

The ABN register counts registration status: an entity is "active" if its ABN has not been cancelled, regardless of whether it currently trades, employs anyone, or generates revenue. The ABS's "actively trading" criteria are narrower and designed specifically to estimate the operating business population, which excludes dormant registrations, non-trading trusts, and structures such as self-managed super funds that were never trading businesses in the first place. Comparing the two without naming this difference produces a false impression that one source is wrong; both are measuring a real but different population.

The gap between the two figures is 1,956,097 entities. That is close to the 2,221,710 entities sitting in the "other" and discretionary-trust buckets combined in the legal-form table above, and the overlap is not a coincidence. This is the kind of cross-check worth running whenever two counts from different sources disagree: does the size of the gap line up with a plausible explanation already visible in the data, or does it need a separate one.

What this means for targeting Australian outbound

For anyone building an outbound list from the ABN register rather than a narrower "trading business" source, the practical filter is legal form first: private and public companies are the buckets worth targeting directly, discretionary trusts and partnerships need a secondary check for whether they run an operating business under that structure, and the self-managed-super-fund-heavy "other" bucket should be excluded outright unless the campaign specifically targets superannuation trustees.

Age is a weaker filter on its own than legal form, since a 20-year-old dormant trust and a 20-year-old actively hiring manufacturer sit in the same age band. Age is more useful layered against a second signal, such as an active trading indicator or a company size estimate, than used by itself to prioritise a list.

Geography adds a third layer worth stacking on top of the first two. A campaign targeting private companies under 5 years old will find a different shaped list in Sydney or Melbourne, where the largest raw counts sit, than in a smaller city, simply because of where the register's overall volume concentrates. None of the three filters, legal form, age or city, works as well alone as the three combined.