On Wednesday 25 February 2026, Prof Tony Haymet delivered a keynote address at the Universities Australia Solutions Summit.
Thank you for that kind introduction and the opportunity to join you at this very important event.
I begin by acknowledging the traditional custodians of the lands on which we meet – the Ngunnawal people – and I extend that respect to all Indigenous people here today.
And I am very pleased this conference has a significant focus on First Nations leadership and inclusion in higher education and university life.
On this day, February 25, the first patent was granted – to Thomas Davenport – for an electric motor. The year was 1837.
At the time it was a specialised technology. But over time it became an invisible but essential enabler of innovation. From innovation comes massive increases in productivity. In concert with other technologies, it would expand the boundaries of human discovery and knowledge.
I mention that anniversary because almost 190 years later on this day – February 25 in 2026 – I want to talk about another transformative technology that is also a driver of discovery … and an engine of change. More in a minute.
It’s hard to overstate the value of Australian universities. They are a source of great optimism to me.
They are part of the very bedrock of modern Australia – and for many years, the bedrock of my life.
When I started at uni in the mid-seventies – with my growing beard– I had no idea my university journey would be such a long and rewarding one.
I am forever grateful for having had a free university education in Australia. I studied at the University of Sydney, and later the University of Chicago, before working as an academic at US universities – Harvard and Berkeley – that truly impressed me. My colleagues and students taught me a lot.
Much later on, I realised that from those four fine Universities, I had learnt to appreciate true scholarship.
In my later roles – from the CSIRO to the Scripps Institution in San Diego – my connection to universities has remained strong.
In fact, when working and visiting national labs around the world, I noticed that all the thriving labs had very close very relationships with universities, and were often co-located.
Those labs were humming – day and night! – with graduate students and postdocs.
Those experiences underscored my deep conviction that universities are powerful drivers of discovery and innovation, and their students are arguably essential for breakthroughs.
Today, I want to focus on one driver in particular of breakthrough discovery and innovation in the decade ahead.
In 1837 the innovation was the electric motor.
Today, the innovation I'm talking about High Performance Computing and Data, or HPCD for short.
I particularly want to talk about it in this forum because it is closely tied to universities, research and – fittingly for this summit – solutions.
At its core, high performance computing and data brings together four things:
- powerful computing systems
- large-scale data storage
- fast networks and complex software
- and – critically – highly skilled people.
These systems solve problems that are too big, too complex or too time-consuming for today’s ordinary computers.
In Australia, their main users are concentrated in geology & minerals, and the science and medical research and government sectors – because it's at the intersection of large-scale data and powerful computation that new discoveries take shape.
But that's not just because of the facilities. It's also about institutions like universities – and the people who design, operate and support these digital systems – working together to enable data-intensive research.
Australia’s high-performance computing and data needs are currently being examined by the Prime Minister’s top science advisory body.
The Minister for Science has asked the National Science and Technology Council to provide advice on Australia’s future needs in this area.
As the Executive Officer of the council, we have been consulting widely across five sectors selected by the Government – sectors in which universities play an important role:
- computer science and artificial intelligence
- energy and advanced materials
- environment and weather prediction
- medical and life science
- and the space and astronomy sector.
High performance computing underpins many breakthroughs and improvements that touch Australian lives every day.
It supports better cancer treatments and medical breakthroughs, strengthens energy and water systems, and improves food production.
It bolsters economic growth, enhances weather forecasting and disaster responses, and plays a key role in climate science and the energy transition that goes with it.
It makes discoveries in astronomy, and is also critical to the use of AI technologies that can deliver enormous benefits to Australia.
So, by powering research and innovation in many fields, HPCD technology makes Australia safer, healthier and more competitive. Increasingly, it will underpin all modern research.
***
Australia’s high-performance computing and data landscape supports a broad range of users. In particular – researchers at universities, but also medical institutes, government science agencies, industry, and through international scientific collaborations.
The landscape is commonly described as having three tiers.
Tier‑1 systems are typically the most powerful machines – publicly funded supercomputers operated as flagship national facilities that support broad‑ranging research.
There are two in Australia:
- the Setonix machine in Perth at the Pawsey Super Computing Centre
- and the Gadi machine in Canberra at ANU at its National Computational Infrastructure facility.
These facilities both do tremendous work – but demand for research time far exceeds capacity, just like proposal requests for ARC and NHMRC funds.
As a consequence, some projects are delayed, reduced in scope, or must sometimes find collaborators offshore.
Next come the Tier-2 and Tier-3 facilities. They’re distributed across universities, government departments and agencies, and industry.
They are vital parts of the high-performance computing eco-system.
Alongside these three tiers are cloud-style research computing services – both public sector and commercial.
They can provide a more scalable alternative to traditional on-site computers and storage. But, in many cases, they do not meet all research requirements.
So, in short, that’s the landscape – and it’s one in which future capacity will be a defining consideration for Australian R&D.
Our existing supercomputers are ageing.
Just like our smartphones, new models and faster machines continue to supersede technology that, in its early days, was among the cream of the crop.
At a global level, the benchmark for computing capability continues to rise.
Countries and regions – including Europe, Japan, China, the USA and Singapore – are making large investments in high performance computing.
But this does not mean Australia should simply match their spending.
Like all nations, Australia has finite budget resources. The question is not whether to invest in this technology, but how to make strategic, pragmatic, clear-eyed choices that maximise the benefits to our nation.
Recent discussions in Australia’s scientific community raise legitimate concerns about whether our digital infrastructure in this area can keep pace with existing and projected demand.
There are concerns about infrastructure that is fragmented, ageing, or not structured for emerging needs in areas such as AI and genomics.
These are healthy discussions that help to clarify issues by raising questions about the allocation of resources.
For example, questions like these.
- If we increase our digital infrastructure in this area, do we upgrade existing facilities or build new ones?
- And how much capability do we set aside for demand arising from AI?
- Do we go it alone – or partner with trusted nations? And when is sovereign capability essential?
- What is the role of government in fostering a sovereign HPCD sector in Australia?
- And in terms of logistics, do we bring the computational systems to the data? Or, establish fast networks that can transmit massive amounts of sometimes-sensitive data?
There are no simple answers, but these are the right questions to be raising.
We now live in an era defined by data. And as our computing power grows, our data capability must grow with it.
If our data storage capacity lags behind our computing power, this can create bottlenecks that undermine research.
In a country as big as Australia, we must also carefully consider questions about connectivity and the location of our facilities.
When it comes to the marriage of supercomputing and data facilities, co-locating them can be an important consideration. It can reduce the cost and difficulty of massive data transfers, and can reduce latency in computer processing.
On the other hand, high-speed, high-volume networks can also be deployed to efficiently connect data storage to computing systems.
And in some cases, it’s a matter of bringing the computing system to the data storage location – rather than the other way around.
Not all data-sets are equal. And sovereign governance of that data in Australia – for reasons of privacy and security – is a key consideration.
This can include critical national security information, natural resources data, medical research using anonymous personal health records, or data from Indigenous communities.
When it comes to sovereign capability, data management and governance are not secondary considerations; they are central.
When you supersize data systems and computers, you generally supersize their energy consumption and heat output.
These problems are not insurmountable. The designers of these facilities are looking at innovative ways to power and cool them.
For example, Australia’s Pawsey supercomputer uses geo-thermal groundwater to efficiently manage supercomputer heat – thanks to a system designed in collaboration with the CSIRO.
Meanwhile, both of the tier-1 facilities in Australia use a significant amount of renewable energy.
Australia is a world leader in the development and deployment of renewable energy sources. High performance computing and data systems are part of that – helping to design renewable energy infrastructure that underpins a global energy transition.
The computing technology helps us coordinate power grids and manage water resources. It enhances our climate modelling and weather forecasting, and the development of new low-carbon materials and technologies.
Earlier, I mentioned the option of co-locating technology, but co-locating people can be just as useful.
When research teams are partnered with HPCD professionals – including those embedded within the teams – the outcomes can be far greater than the sum of their individual parts.
I’m talking here about people such as computational scientists, data engineers, and system administrators. Plus the data curators, research software engineers, and other technical and user support team members.
Their expertise not only increases research productivity but also maximises the value and effectiveness of our digital infrastructure.
So today, I’m sending them a big shout-out!
And I’m encouraging people – in schools, universities and the workforce – to consider these exciting roles that strengthen our national infrastructure.
As technology advances, it’s essential to keep pace:
- by training more specialists
- by supporting secure career paths that keep people in the workforce
- and by upskilling researchers so everyone can fully utilise our HPCD resources.
I want to conclude by going back to those electric motors I mentioned – first patented on this date about 190 years ago.
They’re all around us in this venue. Largely unnoticed – but indispensable. Directly and indirectly, the impacts of those machines permeate nearly every aspect of modern life.
High performance computing and data systems will have a similar effect. They will continue to drive discovery and innovation.
Directly, and indirectly, their work will touch so many parts of our lives.
Over the years Australia has developed those systems, and their talented workforce.
And now, in this time of unprecedented digital change, it’s imperative that we discuss and consider how those systems can serve our nation in the years ahead.
Thank you.