Thank you for the invitation to join you at AgileAus, on the lands of the Wurundjeri people of the Kulin Nation.
It is a pleasure to be with a room full of people who spend their days trying to make organisations less cumbersome, products more useful, systems a bit less maddening and the future more intelligent.
That last phrase is your conference theme this year – building our intelligent future.
I’m a big admirer of intelligence. I’m an even bigger admirer of intelligence paired with wisdom. Wisdom is knowing which problems are worth solving. For a technologist, that means asking whether the clever thing is useful, safe, adopted and capable of making life better beyond the demo screen.
My theme today is how innovation happens and what governments can do to encourage more innovation and technological diffusion.
The history of innovation is full of myths about solitary genius. Alexander Fleming returns from holiday, notices mould on a petri dish, and penicillin is born. It’s a cute tale, particularly because Fleming considered calling penicillin ‘mould juice’, which should give comfort to anyone who’s ever launched with a bad product name.
The real story is messier and better. Fleming noticed something remarkable. Then, for more than a decade, little happened. It took a team at Oxford to extract, stabilise and test penicillin. It took American chemical engineers to work out how to manufacture it at scale. It took wartime production systems to turn a laboratory curiosity into a medicine that saved millions of lives.
Innovation almost never arrives as a thunderclap. It creeps forward through iteration, teamwork, exchange and patient improvement. Those of you who work in Agile environments know this already. You’ve built careers on it: test, fail, learn, go again.
In writing The Shortest History of Innovation, I came to think of 3 recurring forces behind human progress: tinkering, teams and trade. Tinkering is the willingness to test, fail and improve. Teams are the recognition that complex problems usually need more than one kind of mind. Trade is the exchange of ideas, people and capital across borders and disciplines.
Those 3 forces are also a useful way to read this year’s federal Budget.
In this year’s Budget, we have tried to make Australia easier to innovate in. The Budget’s purpose is less to anoint a handful of favoured firms than to improve the conditions in which more good ideas can be tried, more promising firms can grow, more useful technologies can spread, and more Australians can benefit from them.
The package includes more than $3.5 billion in new measures directed at lowering taxes on business investment. The core is reforms to business losses, venture capital and a $20,000 permanent instant asset write‑off for small business. The Budget also better targets the Research and Development Tax Incentive towards higher impact R&D. The aim is to encourage risk‑taking, support young firms, improve resilience and push innovation beyond the handful of sectors that already have it.
Tinkering
Innovation begins with the permission to try.
Anyone who has built software knows that experimentation has a rhythm. You test, you learn, you refine, you try again. The process is productive precisely because failure is informative.
The problem is our tax system hasn’t seen it that way. Firms pay tax on profits immediately, while losses may sit unused until the firm becomes profitable later. The real value of those losses declines with time. That asymmetry especially disadvantages startups, growing firms, businesses attempting to pivot and established firms making a bold new investment. Large incumbents are more likely to have profits elsewhere in the business against which they can offset losses. Younger firms often lack that luxury.
That is why the Budget reintroduces permanent 2‑year loss carry‑back for companies with turnover up to $1 billion. From income years after 1 July 2026, firms that make a temporary loss will be able to claim a refund of tax paid in the previous 2 years. The measure is expected to benefit up to 85,000 companies each year. Basically: if you’ve been profitable, invested hard, and then had a bad year — the system will cut you some slack.
Some of the most useful policy reforms don’t pick what gets invented. They just improve the conditions under which more things get tried.
If a business hires engineers, buys equipment, develops a new product and finds that the timing is off on the first attempt, that should place it on a fairer footing against an older competitor. A dynamic economy needs room for firms to try things before they have the comfort of steady profits.
The same principle sits behind the Budget’s reforms to the Research and Development Tax Incentive.
Australia’s Research and Development Tax Incentive is our largest program supporting business research and development. It has helped many firms invest in new knowledge. But the evidence has been accumulating about what it’s actually doing.
One finding from research commissioned as part of the Ambitious Australia review is worth underlining. It separated R&D into ‘core’ activities and ‘supporting’ activities. Core R&D is spending on experimental activities to create new knowledge. Supporting R&D includes literature reviews, equipment maintenance and the like.
The research found that for every dollar of tax offset supporting core R&D, firms generated around $1.58 in additional business research. That is what economists call ‘additionality’. In normal English, it means the policy changed behaviour.
But the benefits of supporting R&D activities generated little additional research and development. That’s despite the fact that these supporting R&D claims accounted for 29 per cent of claims for the tax offset.
The same review found that 86 per cent of claimants surveyed relied on tax consultants to navigate it, which tells you something about how complex the Research and Development Tax Incentive has become.
So the Budget is refocusing the program on the work that actually produces something new.
Offsets for core research and development will rise by 4.5 percentage points. Supporting activities will cease to be eligible. The small and medium enterprise turnover threshold will increase from $20 million to $50 million, giving more firms access to the highest offset rate. Refundability will be better focused on young firms operating for less than 10 years, where cash flow support is most valuable. Our government expects these changes to make each dollar of tax support about 20 per cent more effective, and to increase research and development by young firms by around $400 million a year.
This is what good innovation policy looks like. It asks where public support changes behaviour, then directs resources there. It rewards real experimentation. It reduces subsidy for activity that firms would likely have undertaken anyway. It recognises that young firms often have the greatest ideas and the thinnest cash buffers.
In software terms, a feature’s longevity in the backlog is not proof of its value. You look at the evidence, learn from users, and improve the design.
Teams
If tinkering is the first ingredient, teams are the second.
The myth of the lone inventor is powerful because it makes for a tidy biography. Yet most important innovation is collective. Thomas Edison did not create the practical light bulb alone. His Menlo Park laboratory employed scientists and craftspeople who tested thousands of materials for filaments. Bell Labs brought together engineers, physicists and chemists, and helped produce the transistor, the laser, the photovoltaic cell and the Unix operating system. A lot of important discoveries happen not at the microscope but in the argument afterwards, when someone asks why the result came out wrong.
That’s also why venture capital is more than a cheque. A good VC brings contacts, hard‑won pattern recognition, help with the difficult work of scaling and a willingness to sit in the room for the decisions that don’t have a manual yet.
The Budget notes that young and fast‑growing firms make a disproportionate contribution to productivity growth. Within 5 years of operation, Australia’s highest‑performing young firms achieve productivity around 45 per cent above industry averages. Young firms also create 6 out of every 10 new jobs. Yet those firms often struggle to access conventional finance because their track record is short, their assets are intangible, their products are unproven or their intellectual property is hard to use as collateral.
Australia’s venture capital market has grown strongly. The technology sector backed by venture capital has expanded from $15 billion in 2015 to $360 billion in 2025. Capital committed under Australia’s main venture capital programs has risen from $6 billion to $30 billion over the same period. Yet venture capital investment in Australia remains modest overall, at 0.16 per cent of gross domestic product in 2024.
The Budget modernises the venture capital tax settings so they better fit the scale of today’s firms.
From 1 July 2027, eligibility under venture capital limited partnerships will extend to investments in businesses with assets up to $480 million, up from $250 million. Early‑stage venture capital limited partnership eligibility will extend to investments in businesses with assets up to $80 million, up from $50 million. Investors in those early‑stage partnerships will retain full access to incentives as investee businesses grow their assets to $420 million, up from $250 million. The maximum committed capital for those early‑stage funds will rise to $270 million, up from $200 million.
By increasing these thresholds, we aim to reduce pressure for premature sale. The new regime will allow funds to stay with promising firms for longer. These changes are particularly important for capital‑intensive sectors such as biotechnology, clean energy, advanced manufacturing and artificial intelligence, where the distance from first idea to commercial success can be long.
The Budget also strengthens the superannuation performance test so that it does not create unintended barriers to investment, including in areas such as venture capital. With Australia’s superannuation pool now around $4.5 trillion, even small improvements in the flow of capital can have meaningful consequences for innovation and growth.
Capital is part of the team too. A country that wants more innovation needs talented founders and skilled workers. It also needs investors willing to back uncertain ideas, advisers who can help firms scale and institutions that understand that the most valuable companies often look strange at the beginning. An app that lets you stay in a stranger’s home? A service that lets ordinary drivers turn their cars into taxis? A platform where people upload home videos for the world to watch? They sound obvious now, but many people initially thought these ideas were absurd.
In The Shortest History of Innovation, I note that more than half of venture capital investments lose money, while roughly 1 in 10 deliver a 5‑fold return or better. Even highly successful investors struggle to predict which firms will succeed once the investment has been made. That is a useful caution for governments. The task is less to pick winners than to create an environment in which more experiments can be run, more firms can be tested, more ideas can compete and more good ones can find their way through.
Trade
The third force is trade, by which I mean more than shipping containers and customs forms. Innovation thrives when ideas travel: between people, between firms, between universities and industry, and between countries.
The wheel did not transform every society in the same way. It was more useful where there were draught animals and trade routes. Printing became transformative when books could circulate. Electricity changed factories after managers reorganised production around it. Artificial intelligence will have its greatest economic effect as it moves beyond frontier companies and begins improving ordinary workplaces across the country.
Which brings me to diffusion.
As the Budget papers note, our business tax reforms are intended to ‘promote the diffusion of innovation through the economy’. This is an underappreciated part of productivity policy. A nation can have brilliant researchers and inventive startups, yet still underperform if better tools and better methods remain marooned in a small number of firms.
The permanent $20,000 instant asset write‑off for small business is part of this story. It gives smaller firms greater certainty when investing in equipment and technology. Its main contribution will be practical: better software, a 3D printer that saves ordering replacement parts, a barcode scanner that turns stocktake from an afternoon into an hour. Diffusion is unglamorous. It’s also where most of the economic dividend actually lives.
Our government has also asked the Productivity Commission to inquire into business dynamism, with a focus on the regulatory barriers faced by firms seeking to innovate and grow, especially young firms. The final report is due by May 2027. That inquiry is important because policy should extend beyond financing research. It should also ask what makes it hard for firms to enter, expand, compete and adopt better ways of working.
For this audience, the idea of diffusion should feel familiar. Agile began in software, but its influence has travelled far beyond it. Shorter feedback loops, cross‑functional teams, user‑centred design and continuous improvement now shape work in banks, hospitals, retailers, government agencies and charities. A method that stays inside one profession is interesting. A method that spreads through the economy becomes productive.
So when we talk about an intelligent future, let’s resist the image of a single machine, a single lab, a single genius. An intelligent future is messier than that: more people testing ideas, more firms surviving long enough to matter, more capital finding its way to the right places, more methods escaping the organisations that invented them.
That’s the bet this Budget is making.
It is a pro‑experimentation budget and a pro‑scale budget. It is also a pro‑diffusion budget and a pro‑evidence budget.
It recognises that risk‑taking is often productive, and that a tax system should treat young firms fairly while they mature. It recognises that promising firms need capital and time, particularly in sectors where breakthroughs are expensive before they are profitable. It recognises that productivity growth comes both from invention at the frontier and from the wider use of better tools across the economy. And it recognises that good policy should learn, adapt and improve.
The human purpose of innovation
Innovation is a means, not an end.
We seek new technologies because they can help people live longer, work better, learn faster and connect more easily. They can ease drudgery, lower costs, widen access and open opportunities that previous generations could scarcely imagine.
None of that happens automatically.
Every technology enters a society with existing inequalities, existing institutions, existing incentives and existing habits. Artificial intelligence could improve medical diagnosis, speed up scientific discovery, help small businesses do more with less and give workers better tools. It could also be deployed in ways that leave workers with less say over their jobs or that concentrate gains in a narrow set of firms.
The right response combines curiosity with judgement. It is to build institutions that encourage useful innovation, spread its benefits and give people a voice in how technologies are used.
The budget measures I have described are part of a longer innovation agenda. They are a set of reforms built from evidence, designed to improve incentives, and open to further refinement as we learn more. The Ambitious Australia review has already helped shape the first stage of reform, including changes to the Research and Development Tax Incentive and venture capital settings. A National Resilience and Science Council will help guide further work.
The most capable governments, like the most capable teams, keep learning.
A final challenge
Let me close by switching from my role as Assistant Minister for Productivity to my role as Assistant Minister for Charities, and issuing a challenge.
The technology community has built extraordinary things. You have made banking simpler, logistics smarter, medicine more precise and communication almost frictionless. You hold skills that are now essential to almost every organisation in the country.
The charity sector needs those skills.
Many charities are doing life‑changing work with brittle databases, patchy cybersecurity, clumsy websites, manual rostering systems and donor records held together by goodwill. They may need a better case‑management system, a more accessible website, help cleaning their data or advice on the responsible use of artificial intelligence. Browse a volunteer match site like GoVolunteer or SEEK Volunteer, and you’ll find plenty of examples of charities looking for this kind of help.
Lawyers often donate their expertise to charities. Doctors give their expertise in clinics and humanitarian settings. Many technology firms give back to the community through volunteering and philanthropy, but I would like to see even more, so ‘tech pro bono’ becomes as normal a phrase as ‘legal pro bono’.
For some charities, a few days of expert tech help could save hundreds of staff hours. For others, it could mean reaching vulnerable people faster, protecting client data better, strengthening donor systems or freeing volunteers from administration so they can spend more time helping people.
And there’s a symmetry here. The skills that make good technologists – an understanding of technical systems, intolerance of waste, rigorous analysis, the habit of asking what users actually need – are exactly what many stretched charities are asking for.
So build the companies. Build the systems. Build the tools. But consider giving one sprint a year to an organisation that could never afford your day rate.
Some of the most important innovations in the next decade may come from frontier science, venture‑backed firms, artificial intelligence labs or university research teams. Others may be humbler: a food rescue charity that finally gets its logistics right, a family violence service with safer records, a disability organisation whose clients can actually use its website, or a community legal centre that spends less time wrestling with its database and more time helping people.
The test of an intelligent future is not how clever the tools become. It is how widely their benefits are felt.
Acknowledgements
I am grateful to officials from the Australian Treasury for valuable comments on an earlier draft.