Scott Farquhar, Co-Founder Atlassian, Chair Tech Council of Australia
Thank you to the National Press Club for the invitation.
I’d like to begin by acknowledging the Traditional Custodians of the land we meet on today, and pay my respects to Elders past and present. I extend that respect to any First Nations people joining us today.
I’m also personally delighted to be joined today by my wife Kim, my parents, my Atlassian co-founder Mike, dear friends and colleagues. Your support means a lot to me and I wouldn’t be here without it.
It’s a genuine honour to stand at this podium, a place where so many of Australia’s most important debates have been aired and some of our greatest ideas first shared.
TECHNO OPTIMISM AND THE TIME TO BUILD
Whilst it’s an honour, to be honest I’m here a little reluctantly.
Over my career, I’ve always preferred doing rather than talking. Building things, solving problems, creating new ventures, not giving speeches about them.
But I’m here today because I believe Australia is standing at the edge of the next great industrial revolution. One powered not by steam or electricity – but by artificial intelligence.
And like every revolution before it, the choices we make now will shape not just our economy, but the kind of country we become.
I’ve always been a techno optimist. Let me take you back to 1991, when Daryl Braithwaite was topping the charts. I was living in a lower middle class family, attending Castle Hill Public School under the watchful eye of my teacher Mrs Crawford. Our class was given an assignment to imagine the future. I had a barely functioning computer, had never heard of the nascent internet, nor seen what was then a brick-sized mobile phone.
Let me read you a little excerpt from 12-year-old Scott…
“In the future adults will no longer get up at all hours of the morning to go to work. Instead, they will wake up and walk over to the home computer. The computer will be linked to a main computer at their work. Anything that they would need is on the computer.
In the future there will be no supermarkets. Instead, people will be able to write out a shopping list and order over a machine. You will even be able to find out how much the shopping bill will cost before you order.
There will be factories run by one person and thousands of machines. They will make computer parts, cars, televisions, even other robots!”
Back then it seemed obvious to me that we would have online shopping and education, online work and entertainment. But back in 1991, we didn’t know which companies and which nations would win. The scoreboard was blank.
Now, not everything I predicted came true. I also predicted “sinoleums,” giant 10 storey zero-gravity entertainment places where you could bounce around for fun. Still waiting on those… but in a fit of optimism I did register sinoleum.com
And even though I had just come from the rite of passage that is the Year 6 pilgrimage to Questacon and Parliament house, there was no mention of the role that government would play.
Roll forward 20 years and the last time I gave a major speech was the Bradfield Lecture. I argued then that we were in a golden age of software, which would transform how we live and work. But even then, back in 2014, most of the winners of the software industry were obvious. We had Netflix and Uber and Facebook. We had Workday and Salesforce and Atlassian. Though some companies have emerged since then, the scoreboard had been filled in.
When it came to the government, there weren’t many big requests. Sure, I wanted to make it easier to hire and retain staff (and to be honest we still need those things), but the best thing the government could do in 2014 for software companies was to leave the industry alone.
I’ve thought a lot about JJ Bradfield since giving a speech in his name. Bradfield wasn’t just the visionary engineer behind the Harbour Bridge, he championed electrifying Sydney’s trains and building the underground city circle, a full decade ahead of New York’s subway.
He argued that electrified trains would boost productivity, cut pollution, and open up the suburbs. But at the time, electricity was new and unsettling – invisible compared to steam’s smoke and noise. It required massive investment in power stations, wires, and trains, and it threatened jobs tied to steam locomotives. Yet the government of the day took the bet and Sydney is better for it.
I’m here today to talk about artificial intelligence, and we hear many of the same arguments: it’s mysterious and dangerous; it requires new infrastructure; we will need different jobs. So many of the same arguments that were made against electrification of our train lines 100 years ago.
When I spoke 10 years ago at that lecture, what we needed was for government to get out of the way. Today we need to partner with government to pave the way. Because the scale of the opportunity and the risks of missing out demand a new kind of partnership, one that moves at the speed of technology, not at the speed of bureaucracy.
MY BACKGROUND
So should I be the one standing here today to talk about this?
I’m currently Chair of the Tech Council of Australia – so I speak not only on behalf of the Australian tech companies that do exist, but also those that have not yet been created, but will be if we play our cards right.
I also co-founded a company called Atlassian – you might have heard of it…
My friend Mike and I went from the two of us 25 years ago with a dream and a bit of Aussie optimism, to 13K staff globally, 300,000 customers and millions of people use our products on a daily basis. Atlassian does about AUD $8B annual revenue and sells to almost every country on the planet.
But when I think about Atlassian, it’s not just the revenue impact. Atlassian is the keystone species of the entire Australian tech ecosystem:
- We provide talent – former Atlassians now hold key roles at companies like CultureAmp, Canva and SafetyCulture. More than 140 startups have been founded by Atlassian alumni.
- We provide capital – Atlassian’s success has allowed us to reinvest in many Australian startups
- We provide mentoring – Mike and I have supported or advised nearly every major Australian software company, from Buildkite to Employment Hero.
I think it would be hard to find an Australian tech company that we have not had some influence on.
I also have deep experience with artificial intelligence. Not just through Atlassian where we have thousands of engineers working on AI, but also through many personal investments in AI companies.
THE NEXT INDUSTRIAL REVOLUTION
To understand the scale of the AI opportunity, it helps to look back. The history of human progress is also the history of how we have harnessed energy in new ways to do useful work.
Historically, these activities were physical. It was once hard to move people across a continent. It was once hard to cross an ocean. It was once hard to plough a field.
Now we barely think about them. Because what was once hard becomes easy. I’ll say that again. When we can harness energy in new ways, what was once hard becomes easy.
And when that happens, nations can rise, or fall, based on how they respond. As someone who has spent most of my life at the cutting edge of technology, I know that when new technologies are created small initial advantages can compound over time into unassailable leads.
200 years ago England benefited from accessible coal, which led to railways, which made coal even more accessible. That early lead led them to invent the Watts steam engine and the Bessemer process for creating steel. On the back of coal powered steam and steel they built a navy—and an empire that lasted a century.
Let’s take a more recent example: Singapore. In the 1970s, Singapore recognized container ships as a transformative technology, not just a passing trend. Singapore opened its first container port well ahead of most Southeast Asian rivals and even faster than many European ports. Singapore rose from a swampland to a first world nation.
It wasn’t a foregone conclusion that Singapore would rise. There were 4 other ports in the region that were geographically well positioned. But Singapore acted early and pressed their natural advantages into economic strength.
Today, we’re at another inflection point. We’re entering the next energy revolution, but this time using energy to power knowledge work. Just as we have gotten comfortable using energy to power physical work, we need to get comfortable using energy to power knowledge work.
We have abundant natural advantages: energy, vast land, a strong legal system, world-class engineers, and a strategic location in Southeast Asia. Just like Singapore, we can press those advantages.
DATA CENTRES
Ironically, Singapore had placed a moratorium on new data centres because they were running out of power and land. Things that were fundamentally hard for them to change.
This brings me to the first opportunity I want to talk about today. Australia has the opportunity to host data centres for the entire Southeast Asia region and potentially beyond.
Much like most people don’t think about the source of their water or power on a daily basis, most people don’t think about how integral data centres are to everyday life.
When an app on your phone reroutes you in traffic, getting you to work on time, it uses a little computing power on your phone, and a lot more in a data centre. When my son learns maths from a chatbot, he’s using computing power hosted in a data centre. When you watch Netflix or, like I do, watch Tour de France replays on SBS, it’s only possible because of data centres.
It’s easy to think that data centres exist to help other people. But the truth is that each of us uses a fraction of a data centre every single day. The buildings, the computers, the power usage of data centres are all in service of making our lives either more productive or more enjoyable.
When Netflix or Amazon or Google or Facebook choose where to put a data centre, they consider a lot of factors: how much it costs to build, the price and availability of power and water, the laws of the country, the talent density to build and run it and finally, in this world of tariffs and geopolitics the availability of cutting edge computer chips.
On all of these factors, Australia scores well. We are surprisingly cost competitive building data centres. We have the talent density and attractively priced green power. But there are areas which hold us back from an even bigger opportunity. We should power the region. We should export megawatts as megabytes (and get paid megabucks). This is a multi billion dollar, multi decade opportunity. To do that, we need to systematically remove barriers to growth.
The first barrier, and this one may seem esoteric, but it is actually important – we need changes to our copyright laws. Australia’s copyright laws are out of sync with the rest of the world. Whilst the USA and Europe have fair use or exceptions for text and data mining, and the Australian Law Commission recommended changes in 2014 and the Australian Productivity Commission recommended changes in 2016, we still remain an outlier when it comes to copyright.
This is a barrier to many AI companies training or hosting their models in Australia. This is even a barrier to Australian born companies. We are in a perverse situation where copyright holders aren’t seeing any more money, but we also don’t see the economic upside of training and hosting models in Australia.
My first ask today is for the attorney general to urgently amend our copyright act to look at fair use and text data mining exceptions. Fixing this one thing could unlock billions of dollars of foreign investment into Australia.
Another barrier is talent. We talk about the theoretical future job losses due to AI, but ignore the very real jobs that are available today.
Electricians, plumbers, air conditioning technicians, solar and battery installers. These are high paying jobs we need right now, and we don’t have enough. It can take a 4 year apprenticeship to train or retrain in these areas. But many of these jobs only require a subset of the skills. It is also hard for mid-career people to take on a 4 year apprenticeship.
I’m calling on support from unions to help create Tech Trades: digital apprenticeships. Short, practical qualifications that can be completed in 6 to 12 months and get people job-ready for high-demand industries like data centre construction and battery installation. These should be stepping stones to full trades like electrician or plumber, but in the meantime, we get more Australians trained and out there working.
The last barrier I want to talk about today is sovereignty. Just like the Australian Government and Australian businesses like to host our data locally, to be subject to our laws and our protections, other countries have the same desire. If we want their business, we need to solve for this.
I propose a bold and novel plan. Just like we host embassies from around the world on our soil, Australia should host digital embassies — secure, sovereign cloud vaults that host countries’ most important data, operating under the laws of that country.
Why host your foreign data centre in any other country when you could host it in Australia, with cheaper power, faster build times and under the laws of your own country. We could be the provider of choice for every government in the region and for every business that needs a Southeast Asian data centre.
AI AND GOVERNMENT
While data centres are a nation-building opportunity and infrastructure is critical, the real economic transformation happens at the application layer.
If Australian businesses simply adopted existing AI tools, we’d see a major productivity lift. Research by the Tech Council shows this could add up to $115 billion a year to our economy by 2030.
Business leaders regularly ask me – what should I be doing differently with AI? I tell them 3 things:
- Use it yourself — daily. If you’re not hands-on, you won’t understand its potential. (I really mean this)
- Redesign your internal processes — especially in areas like customer service, sales, and admin. For example, this smart ring on my finger has a battery issue. Over breakfast, with one hand occupied and my mouth full, I tapped
“contact support” in the app. A virtual agent offered a replacement, confirmed my size and address, all in seconds. - Reimagine your core offering — what problem do you solve, and how could AI solve it 10x better? Every tech shift gives you a choice: disrupt, or be disrupted.
The best businesses in Australia are methodically doing this – reinventing themselves.
So, given businesses are capable and motivated to adopt AI themselves, what is the role of government? It’s the same:
Firstly, Politicians, ministers, heads of government departments need to be using AI on a daily basis to understand what is possible.
Second, governments need to be redesigning back office systems and how they interact with customers
Governments aren’t exactly known for delivering world-class customer service. But in some cases, private businesses can ease the pain. If you’re a small business, you may have experienced this – many accounting systems already file electronic statements with the government such as your BAS or your tax return. Or at a bus stop you’ve used your favourite app to check when the bus arrives.
This is possible when the government provides an API.
An API is simply a way for two systems to talk to each other. It’s what lets your phone app check your bank balance, or an online store track your delivery in real time. APIs are the invisible plumbing of modern software. Let me give you an example.
Christian Beck, CEO of Leap Legal, tells me they’ve made law firms dramatically more productive by using AI to help lawyers automate workflows, speed up contract review, and complete e-discovery in record time.
But there’s one thing still dragging them backwards, the courts.
Despite all the AI and automation inside a law firm, when lawyers need to interact with the court system, it’s still one step above paper-based. They’re forced to copy and paste information into clunky court systems. Waiting for updates they manually refresh web pages. They re-enter the same data over and over again.
Why? Because the court system doesn’t have an API.
As businesses move at an AI pace, governments will increasingly become the bottleneck.
I call upon all levels of governments to make their services available via an API. Interacting with the courts, renewing a license, applying for a passport, claiming the childcare subsidy should all be possible electronically and embedded in third party applications. And these shouldn’t be 7 year consulting projects. Modern tools mean we can do this in weeks and months not years.
APIs are also the building blocks to be able to use artificial intelligence at scale. We need to build on them and go even further. We should create digital agents for every interaction with governments. I’m going to use an example from local government, housing, because it is such a pressing issue for our nation.
As anyone who has built or renovated a house knows, the approval process can be arduous. I have some first hand experience of this myself! You take months to compile hundreds of pages of documentation across all facets of what you’re building, submit it to council (in some cases by printing it out) and then wait. Sometimes weeks, sometimes months. The process seems opaque and arbitrary. You can’t start locking in builders or tradies as you’re unsure of your chance of success, and as we have seen over the past few years, as you wait – the cost of everything goes up. If something needs to be changed, you’re changing it at the end of your process, the worst time to be making changes.
Not only is this frustrating for those who file a development application, but imagine how many applications aren’t filed? How many people decide not to do it?
Instead, what if you trained an AI agent on every application that had ever been filed. All the applications that were successful and all the applications that weren’t. And all the changes people made along the way.
What if that agent worked with the council officer to ensure approvals or rejections happened within a week of filing?
Let’s take it one step further. What if that agent was available to the public? Anytime, day or night, you could send in your application and within minutes get a percentage chance of success? Wouldn’t that change the way you worked with your builder?
Let’s take it further still. A developer would know whether the block of apartments they want to build are likely to be possible, before they even purchased the land. Wouldn’t that dramatically increase not only the number of applications submitted, but also the number of complying and successful applications?
We could dramatically increase the number of dwellings built, without changing a single law.
You can take this concept to any level of approval at every layer of government. Imagine a smart, talented European who is thinking of moving to Australia. They’ve found their dream job on Seek. What if they could find out in minutes what their chances of a visa are.
Let’s create a world where every interaction with governments is assisted by digital agents. How much more productive would we all be? This is not pie in the sky stuff. The technology already exists, companies are already doing it, so why not our governments?
The third thing …. Governments, along with businesses, need to step back and rethink their core service offerings. They should not look at what is possible today, but what will be possible in the near future.
About 30% of the federal budget is spent on health, education and defence. In these areas the government either provides the services, or heavily regulates them, or both. All of these areas are being disrupted by AI. It’s clear the direction many of these areas are headed, and we as a nation need to skate to where the puck is going.
Take health diagnosis for example. A company I have invested in, MetaOptima, has a machine that uses dozens of phone grade cameras to scan for skin cancer, working in a similar way to airport body scanners. Just don’t get the two confused as one of them works best wearing only your underwear. Once scanned, it uses AI to detect the difference between a mole and skin cancer.
Today, skin cancer screening is the domain of dermatologists and GPs, but often with a 3 month wait to see the dermatologist.
Do we want dermatologists spending their time with a magnifying glass searching for skin cancers? Or with a scalpel removing skin cancers?
These technologies are coming. They will save lives. But they will also disrupt the current way of doing things. Governments should move fast to remove regulatory barriers to new ways of working. Not just health, but also education and defence. In both cases it’s clear that the future will be augmented with software and AI. Let’s race towards that future.
History has shown that countries that adopt new technologies earlier gain enduring advantages. Living in the future helps you create it.
I’ve focused this speech on bold, new ideas but there are plenty of basics we still need to fix.
It’s still too hard to migrate talent to Australia. It’s still too hard for start ups to reward them with stock options. And while I believe our current regulations are broadly fit for the AI age, the uncertainty around future legislation is already deterring international investment.
These and other recommendations will be contained in a whitepaper that the Tech Council of Australia will be releasing ahead of the Productivity Roundtable.
I’m also delighted to announce the launch of the National Security Tech Alliance (NSTA) — a new initiative bringing together 27 of Australia’s leading technology companies. The NSTA will act as a strategic partner to the government, to help build partnership between Government and Industry for the future.
CONCLUSION
To close, an AI enabled future is more rosy than my 12 year old self could have ever imagined.
History has shown that in times of change, nations that capitalise on their natural advantages can build compounding and enduring prosperity. A century ago JJ Bradfield showed what can happen when governments build for the future instead of the past.
We have the raw materials – the talent, the energy, the values, the location to lead the world in this next industrial revolution. But leadership is a choice. And we have to choose it now.
Let’s build AI capabilities, not wait for them.
Let’s build skills, not accept shortages.
Let’s build digital infrastructure, not bottlenecks.
The scoreboard of the AI era is blank. The race is still on. And Australia has everything to play for.
Thank you.
