Australia added more than 85,000 businesses last financial year, a headline figure the Albanese government can point to as evidence of economic resilience.
Look underneath that number, however, and the employment story is dramatically different.
Almost the entire net increase came from businesses that employ nobody.
New Australian Bureau of Statistics figures show the total number of actively trading businesses increased by 85,130 in 2025–26.
Non-employing businesses accounted for 83,105 of that increase.
The number of employing businesses increased by just 2,025.
And among businesses that survived throughout the year, there was a net movement of 28,127 from employing workers to employing nobody.
The figures arrive as Treasurer Jim Chalmers promotes what he has repeatedly called Australia’s “most ambitious tax reform package for a quarter of a century”, while employers simultaneously confront higher wage rates, new superannuation requirements and persistent operating-cost pressures.
They do not prove that government policy or artificial intelligence is destroying entry-level employment.
But they expose a striking feature of Australia’s current business expansion: the country is creating substantially more businesses than it is creating businesses with employees.
Australia gained 85,130 businesses in a year
The ABS released its latest Counts of Australian Businesses data on August 18.
At June 30, 2026, Australia had 2,814,778 actively trading businesses.
That represented annual growth of 3.1 per cent.
There were 460,461 business entries during the financial year, producing an entry rate of 16.9 per cent.
At the same time, there were 375,331 exits, an exit rate of 13.8 per cent.
The difference produced the net increase of 85,130.
Those headline numbers provide legitimate ammunition for both sides of the economic debate.
Supporters of the government’s economic record can point out that substantially more businesses entered the economy than exited it.
Critics can point to extraordinary churn and ask what kind of businesses are actually being created.
The answer is overwhelmingly businesses without employees
This is where the latest ABS release becomes particularly interesting.
The number of non-employing businesses increased by 4.8 per cent.
In absolute terms, that was an increase of 83,105.
Employing businesses increased by only 0.2 per cent, or 2,025.
Put another way, approximately 97.6 per cent of Australia’s net increase in actively trading businesses came from the non-employing category.
Only about 2.4 per cent of the increase came from businesses with employees.
That does not mean 97.6 per cent of every newly registered business hired nobody.
The figures measure the net movement in the stock of businesses over the year.
But they do demonstrate where virtually all of the country’s net business growth occurred.
There is another number that deserves attention: 28,127
The ABS also tracked businesses that survived from one year to the next but changed employment categories.
In 2025–26, there was a net movement of 28,127 surviving businesses from employing to non-employing.
That is significant.
It means thousands more continuing businesses moved out of the employer category than moved into it.
But it needs to be interpreted carefully.
The statistic does not establish that 28,127 businesses each fired their entire workforce because they could no longer afford employees.
A business can become non-employing for several reasons.
Employees may leave and not be replaced.
An owner may restructure.
A business may reduce operations.
Contractors may replace employees in some circumstances.
Or an enterprise may simply change as it moves through its lifecycle.
The ABS release measures the transition.
It does not assign a motive to every business making it.
The 375,331 figure is real — but ‘closures’ needs qualification
The other dramatic number in the data is 375,331.
That is how many business exits the ABS recorded during 2025–26.
It is tempting to translate that directly into 375,331 businesses going broke or shutting their doors.
That would go further than the statistics support.
ABS business exits are not equivalent to insolvencies.
The dataset counts businesses entering and exiting the population of actively trading businesses.
An exit can occur for multiple reasons and should not automatically be described as a business failure caused by economic distress.
The more accurate picture is still striking: 460,461 entered while 375,331 exited, creating unusually high levels of business turnover.
Business survival rates provide another warning sign
Recent analysis of the ABS figures has nevertheless identified genuine concerns about business durability.
Only 49.2 per cent of businesses entering the market survived for four years or more, according to analysis of the latest ABS data reported by The Australian.
That is a decade-low survival rate.
Survival differs substantially by industry.
Agriculture recorded one of the strongest four-year survival rates, at about 60.2 per cent.
Healthcare and social assistance followed at about 59.2 per cent.
The composition of new businesses also matters because many entrants are sole operators rather than traditional employers.
Australia’s business economy is increasingly dominated by tiny operators
Even before the latest release, non-employing businesses made up the majority of Australian enterprises.
Australian Small Business and Family Enterprise Ombudsman data based on ABS statistics showed roughly 64 per cent of businesses were self-employed or non-employing.
Another quarter employed between one and four people.
That means the Australian business landscape has long been dominated numerically by very small enterprises.
The latest figures accelerate that pattern rather than create it from nothing.
The economic question is whether this reflects healthy entrepreneurship and flexible self-employment, or increasing fragmentation of traditional employment into one-person businesses.
It may contain elements of both.
Chalmers is simultaneously selling a much bigger reform story
The timing of the ABS release is politically awkward because Chalmers has been promoting an ambitious restructuring of parts of Australia’s tax system.
Introducing Treasury Laws Amendment legislation earlier this year, the Treasurer used unusually expansive language.
He called it “the most ambitious tax reform package for a quarter of a century”.
The government says its reforms are intended to cut taxes for workers, improve housing affordability, reduce distortions between different forms of income and lift economic productivity.
Chalmers has repeated that broader reform narrative since the government’s Economic Reform Roundtable.
Labor argues it is tackling structural economic problems that governments avoided for years.
The opposition and business critics are asking a different question: what is happening to employers while Canberra celebrates reform?
Businesses received a major wage-cost increase on July 1
One part of that concern can be established clearly.
Australian employers paying award wages faced a substantial increase from July 1.
Modern award minimum rates increased by 4.75 per cent.
The lowest ongoing award rate rose to at least $26.44 an hour, or $1,004.90 a week.
Special treatment applied to some entry-level classifications, with the Fair Work system setting minimum floors for those rates as well.
The increases delivered substantial additional income to millions of workers.
For employees dealing with rent, mortgages, groceries and energy bills, that matters.
For employers, particularly labour-intensive small businesses, it is also an additional recurring cost.
Both propositions can be true simultaneously.
Payday Super arrived at the same time
Employers also entered a new superannuation regime from July 1.
Under Payday Super, compulsory superannuation contributions are now required to be paid alongside employees’ ordinary pay cycles rather than under the previous quarterly timetable.
The reform is intended to ensure workers receive super sooner and reduce the amount of unpaid superannuation.
From an employee’s perspective, there is a clear benefit.
Money reaches retirement accounts earlier and missing payments can become apparent more quickly.
For businesses, particularly those managing tight cash flow, the change requires super liabilities to be funded more frequently.
That is a cash-flow adjustment even where the ultimate superannuation liability itself would have existed anyway.
Higher wages are not the only cost pressure
Businesses have also spent recent years absorbing increases in rents, insurance, energy, finance and supplier costs.
CreditorWatch chief economist Ivan Colhoun has identified wages, rents and elevated interest costs among the pressures confronting Australian operators.
CreditorWatch’s July figures also showed trade-payment defaults increasing to their highest level since September 2025.
Hospitality remained particularly exposed.
That context helps explain why business groups focus intensely on payroll costs.
A wage increase that appears manageable in isolation can become much more difficult when combined with rent increases, electricity bills, insurance premiums and weaker consumer spending.
Does that explain the rise of no-employee businesses?
Not by itself.
The ABS figures establish that non-employing businesses grew much faster than employing businesses.
They also establish a net shift among surviving businesses from employing to non-employing status.
They do not establish that the minimum wage increase caused either outcome.
Indeed, the latest annual business-count period ends on June 30, while the new 4.75 per cent award increase took effect from July 1.
That timing is crucial.
The 2025–26 ABS business-count figures therefore cannot be used as evidence that the July 2026 wage increase itself caused businesses to stop employing people during the preceding financial year.
Businesses may have anticipated higher costs, but proving such an effect would require different evidence.
The same caution applies to Payday Super
Payday Super also began on July 1, after the period measured by the latest ABS annual business counts.
It therefore cannot directly explain the 2025–26 movement captured in those statistics.
It can legitimately be included in a discussion of the cost environment employers face now.
But it should not retrospectively be assigned as the cause of business changes that had already occurred.
This distinction matters because the numbers are strong enough without attaching causes they cannot prove.
Then there is artificial intelligence
AI introduces another dimension to the employment debate.
Generative AI tools can now perform tasks that only a few years ago required significant human labour.
Routine drafting, data extraction, customer support, basic coding, document processing, scheduling and administrative tasks can increasingly be automated or accelerated.
For a small business, that can change the economics of hiring.
An owner who previously needed an additional administrative employee may be able to handle more work personally with software assistance.
A larger employer may allow a vacancy to remain unfilled because existing employees using AI tools can absorb part of the workload.
That does not necessarily appear in statistics as a redundancy.
The employee simply leaves and the position disappears.
That is the ‘no backfill’ problem
This phenomenon matters because technological disruption does not always arrive through dramatic mass layoffs.
Suppose a company has 100 employees and five resign over the course of a year.
If management decides technology allows the remaining 95 employees to perform the same work, those vacancies may never be advertised.
No five-person redundancy announcement occurs.
There is no factory closure.
The employer simply hires fewer people.
At an economy-wide level, this can appear as slower hiring rather than conventional job destruction.
That is one reason economists increasingly watch vacancies, hiring rates and workforce entry points alongside headline unemployment.
But the latest ABS business-count figures do not prove AI caused it
This is another distinction worth maintaining.
It is plausible that AI is reducing demand for some categories of labour.
There is considerable international evidence that employers are experimenting with AI precisely to improve productivity and reduce the amount of routine human work required.
But Australia’s latest business-count release does not ask businesses why they changed employment categories.
It therefore cannot tell us how many of those 28,127 net transitions from employing to non-employing involved AI.
Nor can it establish how many entry-level vacancies were deliberately left unfilled because of automation.
Those are important research questions, not conclusions contained in this dataset.
Entry-level workers have particular reason to watch what happens next
The potential impact of AI is nevertheless particularly significant for young Australians.
Many traditional first jobs in white-collar workplaces involve precisely the tasks generative AI handles most easily.
Junior administration.
Basic research.
Routine customer inquiries.
Simple document preparation.
Entry-level coding.
Data processing.
Those jobs have historically served another purpose beyond completing the work itself.
They are training positions.
A graduate begins with routine assignments, learns how an organisation works, develops professional judgment and gradually progresses towards higher-value work.
If technology removes the bottom rung, the consequences can extend far beyond the number of individual jobs automated.
The productivity benefit creates a genuine dilemma
It would also be misleading to portray every business using AI instead of hiring as behaving badly.
Businesses compete.
If technology allows a small operator to perform administrative work in one hour instead of five, using it may be the difference between survival and closure.
Productivity improvements are ultimately one of the main ways economies raise living standards.
The policy challenge is what happens during the transition.
If established workers become more productive while fewer opportunities exist for inexperienced people to enter the workforce, aggregate productivity can improve while a generation of young workers finds career entry increasingly difficult.
That is a very different problem from traditional technological unemployment.
One-person businesses are not necessarily evidence of economic weakness
The extraordinary growth in non-employing businesses also has benign explanations.
Technology makes it easier than ever for one person to operate a viable enterprise.
Online marketplaces can find customers.
Cloud accounting can manage finances.
Digital advertising can replace a marketing department.
AI can assist with writing, analysis and customer communication.
Contractors can provide specialised skills when needed.
A successful modern business therefore does not necessarily need a large permanent payroll.
An increase in one-person businesses can represent entrepreneurship rather than distress.
The ABS data alone cannot distinguish between an ambitious new founder building a scalable company and a former employee creating a sole-trader operation because conventional work is unavailable.
For the government, the total number still matters
Labor can legitimately point to the fact that Australia finished the financial year with more actively trading businesses than it began with.
Despite 375,331 exits, there were 460,461 entries.
The net business count increased by 3.1 per cent.
Even employing businesses increased slightly — by 2,025 — marking the first net increase in the total number of employing businesses since 2021–22.
That final detail complicates claims of outright employer collapse.
The number did not fall.
It grew.
It simply grew at an extraordinarily weak rate compared with non-employing businesses.
That may be the more important story
The strongest criticism of Australia’s latest business figures does not require claiming that every exit was a bankruptcy or every disappearing employee was replaced by artificial intelligence.
The official numbers themselves present a legitimate economic question.
Australia added 85,130 actively trading businesses.
Only 2,025 of that net increase came from employing businesses.
Another 83,105 came from businesses without employees.
And among surviving businesses, there was a net movement of 28,127 from employing to non-employing status.
Those numbers do not tell policymakers why.
They do tell them what needs explaining.
The next financial year will provide a much cleaner test
This is particularly important because several major employment-cost changes only began on July 1, 2026.
The 4.75 per cent modern award increase is now in force.
The National Minimum Wage has risen.
Payday Super has commenced.
Businesses continue adopting increasingly capable AI systems.
And households remain sensitive to living costs and interest rates.
The 2026–27 data will therefore provide a much better test of how employers respond to this new combination of wage, regulatory, technological and demand pressures.
If employing-business growth accelerates, the current figures may look like a transitional low point.
If the non-employing sector continues expanding while employer numbers stagnate, concerns about Australia’s capacity to generate traditional small-business jobs will become harder to dismiss.
Chalmers has made the government’s ambition clear
The Treasurer is not hiding the scale of what Labor wants to achieve.
He has described the government’s tax agenda as the most ambitious in a quarter-century.
He says it will improve opportunity, help workers and first-home buyers and make Australia’s economy more productive.
Those are substantial promises.
The latest ABS figures provide another benchmark against which that economic program will eventually be judged.
Not simply by how many Australian businesses exist.
But by how many can grow.
How many survive.
How many can afford to employ people.
And whether young Australians can still find the entry-level jobs that allow them to begin building careers.
The headline is positive. The employment breakdown is much harder to celebrate.
Australia did not lose businesses on a net basis in 2025–26.
It gained 85,130 of them.
That is a fact.
But 83,105 of the net increase were businesses employing nobody.
That is also a fact.
The number of employing businesses increased by just 2,025.
And a net 28,127 surviving businesses moved from employing to non-employing status.
None of those figures proves Labor’s policies caused employers to abandon hiring.
None proves AI is eliminating Australia’s entry-level workforce.
And 375,331 business exits should not automatically be described as 375,331 bankruptcies.
But as Chalmers promotes a quarter-century-scale reform agenda, the numbers identify a challenge Canberra cannot measure merely by counting ABNs.
A growing economy needs entrepreneurs.
It also needs businesses willing and able to employ people.
On the latest ABS figures, almost all of Australia’s net business growth is happening without them.
