About the episode

Australia is fast becoming prime real estate for the growing AI data centre rollout. But why now?

As multi-billion-dollar projects expand, we're grappling with how to balance the economic potential with the pressure on our energy grids and water supplies.

Associate Professor Kristle Romero Cortes is Senior Deputy Director of the UNSW Institute for Climate Risk and Response. She explains the economic drivers, environmental risks, and how this data centre expansion could be carried out sustainably.

The Business Of podcast is brought to you by UNSW Business School, produced with Deadset Studios. It’s hosted by Professor of Practice Fahim Khondaker.

Want to know more? 

Want to follow the latest research and news from UNSW Business School and AGSM @ UNSW Business School? Subscribe to our industry stories at BusinessThink and follow UNSW Business School and AGSM on LinkedIn. Or listen to more episodes of The Business Of and subscribe to our UNSW Business School and AGSM channels on YouTube.

Edited transcript

Associate Professor Kristle Romero Cortes  00:00
So, the technology is new, and we are learning as we go. But there are so many lessons learned from people working in this space, in climate, in finance, and across a broad range of social sciences that appreciate that now is not the time to say that we can't be creative about solutions.

Professor Fahim Khondaker  00:23
If you look at today's headlines, one thing is clear: as AI use accelerates globally, the data centres needed to support it are growing too. It's easy to think of AI as something intangible, a bit like the cloud. But in reality, it's powered by very physical, industrial-scale data centres. And increasingly, Australia is where some of them are being built. Our vast land, natural resources, and political stability make us an attractive choice for tech companies. So, is this a win for the economy, or a potential climate risk? And who's actually footing the bill? This is The Business Of: AI Explained, a podcast by UNSW Business School. I'm Fahim Khondaker, Partner in Data and AI at BDO and Professor of Practice at UNSW. Joining me is Associate Professor Kristle Romero Cortes, Senior Deputy Director of the UNSW Institute for Climate Risk and Response, and Co-Director of the RISE Finance Lab. Welcome, Kristle.

Associate Professor Kristle Romero Cortes  01:22
Hi. Thank you.

Professor Fahim Khondaker  01:23
Let's start easy: what's a data centre, and what do they need to run?

Associate Professor Kristle Romero Cortes  01:27
Sure, let's demystify data centres. So effectively, all it is is a warehouse, with rows and shelving of computers, in terms of what they could be running, so the servers typically. And inside all of those are either CPUs or GPUs, right? And historically, they have been doing data comps for email websites, you know, very standard practice things, and now we're asking more of them. So, this is why we're moving into the GPU space because those chips can process a lot in parallel. And that's what you want when you want to be training or using an AI model, because that translates to getting a lot done all at one time. But what that means is we're now putting a lot more stress on the environment and on data centres in particular, because those chips generate a lot of heat.

Professor Fahim Khondaker  02:32
So you're saying we had data centres before?

Associate Professor Kristle Romero Cortes  02:34
Yes.

Professor Fahim Khondaker  02:34
We just have more powerful ones now?

Associate Professor Kristle Romero Cortes  02:36
We have ones that are computing more at once.

Professor Fahim Khondaker  02:40
And do we need more data centres, or are we just getting more powerful versions of the same ones we had?

Associate Professor Kristle Romero Cortes  02:44
They are being repurposed to some degree, and they are also being expanded because demand for AI exceeds what we would have seen before in terms of the need to actually compute, since standard email and maybe website practices were quite small by comparison.

Professor Fahim Khondaker  03:06
But we already have AI in Australia. I used ChatGPT this morning. So a lot of us will ask: what's going to be different about a data centre in Australia versus the ChatGPT I used this morning?

Associate Professor Kristle Romero Cortes  03:18
Let's talk about the electricity required to power a data centre. So, we referred to the chips inside a data centre before, right? These GPU chips put off a lot of heat. Because of that, they need to be cooled, either with additional electricity or with water. Okay, and that's the typical process. We are looking at ways to use other coolants, but right now they're emitting a lot of heat and cannot continue to function if overloaded. So we need to cool them. And this has a real environmental impact because it requires more electricity to power these higher-compute data centres. Then you have about the same amount of electricity required to cool them. So, in the areas where they're operating, you're going to have additional demands on the grid and the water system. So, a lot of what's being discussed right now is ways to mitigate some of those potential negative externalities. But that's not the only impact. Data centres themselves: once it's cooled, the heat doesn't go away; it simply leaves the data centre, meaning the areas around the data centres get heated up, and that's great if you're in Finland. But here in Australia, we're already pretty hot. But beyond that, there's noise pollution, air pollution, and a potential fire risk from the installed batteries. We're working on all of those things. We're working to mitigate those problems, but we don't yet have a clear picture. And so, that means those communities around those data centres could be affected in ways they didn't agree to.  And we've even seen internationally that house prices and property values in the areas around data centres are decreasing. So it's not a trivial question, to say the least. It is true. We run our models using ChatGPT, Claude, and Gemini, but we're also participating in a chain of events that have real economic and environmental outcomes.

Professor Fahim Khondaker  05:43
Why all of a sudden is there such a big push for Australia in particular to want data centres?

Associate Professor Kristle Romero Cortes  05:49
Sure. Well, there is a lot of energy behind that question. There's a lot of urgency and excitement around data centres, and it translates into a few things. But the first is potential sovereignty over this computation, right? So, we also want to make sure that if we become reliant on these models, we have the access and control to compute them here at home.

Professor Fahim Khondaker  06:15
You mentioned sovereignty earlier, and can you just explain to us what sovereignty means in this context, and what happens if we don't have it as Australia?

Associate Professor Kristle Romero Cortes  06:23
When you're looking at our internet usage, we are connected via underwater cables, right? And this allows Sydney and Australia to connect with the world. And that means that we have access, connection and expectations like everybody else does. But there is something to be said for our daily reliance on this tool, and we would not want to offshore any of that potential control in case it could be taken away. We are becoming quite reliant on this technology. We're using it in ways that I don't think we could have ever imagined. And so, we wouldn't necessarily want to put that in the hands of a foreign government or body simply because, well, things look okay now. But there is still some sense of what would happen if we weren't to have access to it. But we also have to appreciate that we are very far, right? So there is a level where we actually benefit, even in terms of latency and connectivity, from having those data centres here in New South Wales, because they're closer to those underwater cables, right? And a lot of that is being decided as you talk about the expansion of data centres and Australia's ability to genuinely power and control some of that technology, because we haven't even gotten to the critical minerals required for chip construction and data centres. And a lot of that is currently controlled by China, but there are also areas where Australia could play in that space to provide some of those critical minerals. And we see that it could be a really interesting way for Australia to take part in that discussion. But once again, it's taking things out of our ground and, you know, building things on top of it. So is that the only use of those potential minerals? And, you know, those questions need to be discussed before we, you know, haven't learned the right lessons and then don't have a chance to, maybe, fix those mistakes.

Professor Fahim Khondaker  08:49
What happens at the local level, where the data centre gets implemented, what are some of the key considerations on the ground?

Associate Professor Kristle Romero Cortes  08:55
Well, there are some economies in the U.S., for example, that are making data centres pay for hospitals or roads, right. So data centres ultimately are infrastructure. We have to be very mindful, however, that previous infrastructure projects are like roads, and you consider those relatively riskless, right? They're installed, and we use them. Are data centres as riskless? Probably not like roads. So, that needs to be compensated in some fashion, and that's why we're seeing backlash. But the reality is, until those certainties are in place, we need to really appreciate that problem. And Fahim, while that potential money is coming in and being waved around, the guardrails have not yet been enacted into law. So we're slightly outpacing some of those government guardrails in terms of actual investment. And that's what I think those local communities will be watching very closely.

Professor Fahim Khondaker  09:58
If we go back to the previous conversation about the importance of sovereignty, and how urgent it is that we have sovereignty over this technology. It's about balancing that with these environmental considerations; is that the difficulty our leaders have to face now?

Associate Professor Kristle Romero Cortes  10:12
There is no technology gap when it comes to using renewables or sourcing them in a way that ensures we're not an addition to the grid. There is simply a desire-and-willingness gap, and when we're here saying, "Well, you have the money; you're going to fund it." Then, that is what gets you over that hump to actually say, "Okay, they're being powered with those renewables." And that's where we're seeing in Ireland, and in Singapore, other economies that have looked at this, and potentially even in Ireland's case, have put a moratorium on the data centre expansion until they could get a handle on the environmental requirements. And so, we're coming into a relatively mature market with a lot of lessons learned; we go in with our eyes wide open.

Professor Fahim Khondaker  11:01
Do we have the resources to do this sustainably?

Associate Professor Kristle Romero Cortes  11:05
All of that technology exists, and in terms of resourcing it, I think you know UNSW is very well placed to have talked about how some of our researchers have worked on solar panels or vanadium redox flow batteries. Or even looking at graphite as a way to appreciate it and reduce potential cooling costs. Right, so it's working in research; it's useful for industry – it's a matter of bringing those two pieces together and building it out with the available resourcing. And if all of that funding is available for expansion, it should be just as available for renewables.

Professor Fahim Khondaker  11:48
And who's funding these expansions? And the people who are funding and investing in this, have they factored in ESG and climate risk into their considerations, I guess?

Associate Professor Kristle Romero Cortes  11:59
Funding is interesting, because you do talk a lot about the companies themselves. You talk a lot about the state and federal governments. There's private credit, and then potentially just the traditional credit markets. So it is somewhat scattered across the different options. But capital is typically not the problem. It's usually coordination. So the idea of whether or not they're considering these ESG, you know, specifications, it's about literally appreciating that climate risk. And you know, in my own research, we've seen that banks, specifically in the U.S., because that's where we have the data, are actively tightening credit in areas that are more exposed to climate risk even before natural disasters hit those areas. So when you translate here to Australia, we have a very concentrated financial market, right? Our big four banks operate in the same areas. Our capital market is relatively small and is actually powered by many of these companies that will be really interested in this conversation. So, because of that, there is an understanding that you can't easily diversify away some of these risks, so it should be paramount that they pay attention to it. And if they're not, please listen to this podcast.

Professor Fahim Khondaker  13:24
And you obviously specialise in climate risk, and in Australia, we've got droughts, floods, bushfires, you name it. If these keep increasing, what's the flow-on impact for the AI industry and the infrastructure that we're talking about? What happens there?

Associate Professor Kristle Romero Cortes  13:37
They will become more common, and that's an excellent point because people think about how AI affects climate. We also need to think about how climate will affect AI. And the data centre is useless when the flood washes it away. Also, if it's generating heat to begin with, and you're in an area that's getting hotter, well, that's not going to be great for computing within the data centre; same thing. Bushfire risk, and at the end of the day, if anyone's going through a drought, the last place they're going to be wanting to share water with is the data centre down the road. It's very topical on people's minds, given what's happening in real life, and we see it in the Asia-Pacific region. This is not a problem for tomorrow. This is a far-gone yesterday's problem. And, as much as we want to be positive about the potential usage of AI, and there are potentially some positives and some benefits from it, we have to be very mindful of those costs.

Professor Fahim Khondaker  14:32
If we bring it back to AI infrastructure and its rollout, have you seen in your research what different countries and regions are doing, and what they are grappling with?

Associate Professor Kristle Romero Cortes  14:52
I think what is important to appreciate is that we are all grappling with this understanding and this learning process. And the learning curve at first was quite steep, right? Because we were excited and we moved forward with things, and we didn't necessarily appreciate what the potential costs could be. Now we've spoken about the electricity grid cost and potential water usage, but there are also additional potential sources of emissions, right? Building these, right? So the emissions forgone from the concrete and the steel, and moving things in. There's also a really important piece about e-waste, right? I mean, how do we recycle if we can recycle some of these new pieces being moved in? And it's, it's really hard to sometimes get pieces back out, and it's not necessarily commercially viable to recycle some of these. So what it's going to be is that you have to replace these at really high timeframes. And so you're going to see that in certain areas like the EU, which has much stricter regulation, because they appreciate and are really feeling the acute effects of climate change in a way that we can maybe not see as acute in some areas of the world. They're rolling these things out at a different pace. That being said, China really has a lot of the potential critical minerals used in some chip manufacturing. So because of that, they're in a different space in terms of what they're doing, because that requires mining in order to take those out. And we can see that's really interesting, because other countries are then stockpiling some of those to anticipate potential forward-looking needs. Right, so everybody feels like we're all running the same race, just at different paces.

Professor Fahim Khondaker  16:47
We often hear of really innovative stories from data centres around the world, with people doing really cool things with data. Can you tell us a few stories that you've come across?

Associate Professor Kristle Romero Cortes  16:57
Sure. So I'm aware of one in the UK where they're using the excess heat from a data centre to warm the swimming pool next door. And then, you've looked at places like Louisiana or Michigan, where they've rolled out some of the abilities to get these data centres in these local areas. But in exchange, they're asking for building roads or hospitals, and to that end they're using the carrot of bringing them to their locality, but then also saying, "Okay, but in return, we want you to invest in this local area."

Professor Fahim Khondaker  17:37
And then we often hear things in the media that are a little bit on the wild side. So, like we've heard about data servers in space?

Associate Professor Kristle Romero Cortes  17:45
Yes.

Professor Fahim Khondaker  17:46
Data servers underwater in China as well. Is that viable?

Associate Professor Kristle Romero Cortes  17:50
They're realistic in the sense that they exist, right? So there is a real data centre underwater, and SpaceX is attempting to get permits to put satellites into space that use solar panels, which will have great access to the sun and won't need the same type of evaporative water use because they will be in space. But I think it comes back to being just how mindful that financing decision must be. How high those costs must be for people to take such drastic steps to avoid them. Because it's not just one data centre, or maybe the second data centre. It's on those flow-on costs.

Professor Fahim Khondaker  18:36
And we are obviously talking pretty early days at the moment, and there are a few assumptions being made as we go along this journey. What could go wrong?

Associate Professor Kristle Romero Cortes  18:45
Look, you can make your data centres as carbon neutral as possible, but if the models inside are simply powering the expansion and exploration of future fossil fuels, then it's game over. So that is the real stake here. But the stark reality is that this technology could be used for good or bad. And the good could be where we see actual applications. Like the UN is using it to detect methane leaks, and they're able to detect them earlier than humans. And with that, you're effectively seeing the equivalent of taking millions of cars off the road if you address those leaks early. But on the other side, that use of the current companies that drill and dig, being able to be more efficient in that process and pumping out more of that pollution, could be somewhere to the order of three to thirteen times the amount of the emissions coming from what would then be considered trivial data centres. So the technology is new, and we are learning as we go. But there are so many lessons learned from people working in this space, in climate, in finance, and across a broad range of social sciences, that appreciate that now is not the time to say that we can't be creative about solutions. And I have a bit of a thought experiment for you. So, we know what the Industrial Revolution looked like, and we've seen those photos, and you know, the smokestacks going into the air and seeing the black skies and what came of that: acid rain. And I want you to flip that script and imagine a pristine blue sky because it was powered by solar panels, and wind turbines, and geothermal, and all of this technology exists, and you know, flow batteries that have durations of 20 to 30 years. Things that we humans have already created and simply need to put into practice. And that comes back to being aware of those decisions and financing those decisions with the willingness to stop the flow of what could just be path dependency. And say, we're going to do this, but we're going to do this the right way this time.

Professor Fahim Khondaker  21:16
What are some of the key questions? If we were going to be specific, to give people who are involved in this space, what are some of the key questions that we should all be asking?

Associate Professor Kristle Romero Cortes  21:25
Personally, I would want them to do more. I would want them to say, we have this incredible new technology, and we seem to have funding for days. So, what is the best we could do with this? What could we genuinely create with a product now? We could be solving these global challenges, and I think climate change is right up there with them.

Professor Fahim Khondaker  21:54
Kristle, that's been really insightful. Thank you so much for joining us.

Associate Professor Kristle Romero Cortes  21:56
Thank you for having me.

Professor Fahim Khondaker  21:58
Associate Professor Kristle Romero Cortes is Senior Deputy Director of the UNSW Institute for Climate Risk and Response, and Co-Director of the RISE Finance Lab. If you enjoyed this conversation, hit subscribe wherever you're listening or watching, and leave us a rating or review. It really helps more people find the show. The Business Of is brought to you by the University of New South Wales Business School and produced with Deadset Studios.