The day of the datacentres

Inside Science introduction

0:00This BBC podcast is supported by ads outside the UK.

0:30An award-winning behavioral scientist and author of the best-selling book, How to Change, as she shares true stories from Nobel laureates, authors, athletes, and everyday people about why we do the things we do and how to make better choices to help avoid costly mistakes. Each episode covers the latest research in behavioral science and dives into themes like the power of self-control, shaping your mindset for success, navigating new beginnings, and why starting

1:00over can feel so hard. Listen to choiceology at schwab.com slash podcast, or wherever you listen.

1:11Hello, I'm Tom Whipple, and this week on Inside Science from the BBC World Service, more news you can use, such as what is the real truth about data centers, water-guzzling climate criminals, or noble saviors of the economy? And should the fact that the cleaner ass can spot itself in the mirror have any impact on your eating habits? Oh, and Roland Pease explores the weather impacts of an evolving El Nino. All that, and we have Caroline Steele, who's been looking through the

1:41journals for the other big science. Caroline, what have you got? I've got the death spiral of my favorite telescope, so it's a sad week for me.

The environmental cost of data centers

1:49Fabulous. First though, data centers. Is your XFeed in uproar? Is your street WhatsApp sharpening their pitchforks? Mine certainly is, because the data centers are coming. Hundreds of them, gigawatts of them, appearing overnight in light industrial estates and open fields like mushrooms of the metaverse, the sprouting weeds of the web. And not everyone is happy. In Scotland, campaigners are calling for a moratorium. In the US, the state of New York

2:20has already imposed a one-year pause. According to some campaigners, a single chat GPT request sent to a data center uses half a liter of water. But wait, according to OpenAI, it uses a teaspoon. According to reports this week, just two data centers in the UK will exceed the carbon emissions of ExxonMobil in Britain. But wait, according to AI companies, data centers can take advantage of a rapidly decarbonizing grid and serve up your requests using gentle wind power

2:53from the North Sea. What's the truth? And given that data is no respecter of borders, why can't we just put them all by an Icelandic volcano and run them off free geothermal power and unlimited rain? Here to answer these questions and more is Deborah Andrews, Emeritus Professor at London South Bank University, who has spent much of her career researching these buildings in which all our technology futures reside. Deborah, thanks for joining us. Why are we all suddenly talking about this now? Hello, thanks for inviting me. I think it's really because although data centers have been

3:27around for a long time, there's absolutely exponential growth since the introduction to the public sphere of AI, which is very resource intensive water, energy and materials. And so to power the AI universe, metaverse, whatever you want to call it, we need more data centers. And they are popping up in inverted commas all over the place at the moment. OK, well, let's get straight onto it then. You mentioned water,

4:01energy, materials. Let's start with water because we see wildly conflicting claims about how much water they actually use. Water in data centers is used for three primary things. First of all, obviously, it's used for cooling compute. Secondly, it's used in conjunction with electricity generation to power the compute services. And thirdly, it's embodied in electronic and other components. And also, if you're looking at

4:33the big broader scope, then in building materials themselves. So it really depends on what your interest and bias is as to the sort of scope that you consider when you're estimating or calculating water use in the data center for particular activities. OK, so let's look at just cooling, because that's the one where if you're concerned about much of Britain is in drought, we've got hosepipe bans, and then we've got data centers turning up in places that have hosepipe bans. If your concern is local use,

5:08what are the worries there? And is it always, you hear about closed loop ones or air cooled ones, are the ways where they wouldn't be using local water? Yes, water demand is very high for cooling because most data centers or many data centers, they actually use water specifically or other liquids to cool servers and the air within the compute rooms as well. So data centers were traditionally air cooled, but that uses a lot of electricity. So there's a kind of trade off and water and liquid cooling has

5:46increased because you can cut down your energy consumption, but obviously you increase your water consumption if you're doing that. So most data centers that use water for cooling, it's evaporated off, but you can use closed loop systems. But then somehow you have to dissipate the heat from the water. So you've got your compute running, heat is transferred to the water, and then somehow you need to get rid of that and to stop thinking of this heat as a waste product and to use it in some way

6:23it can be reused. So that's in a closed loop system. Okay, so you've described this sort of trade off between the cheapest way or easiest way to do it is just evaporate off the water. And if you want to do other things, then you've got to think about energy and maybe use more energy. What about energy use itself? I mean, I've seen estimates that we're projecting AI energy use in the UK, which would sort of match the current energy use for everything else. I think if the data center

6:55industry plans are fully implemented, it wouldn't just match energy consumption, it would significantly increase it. So there's a plan at the moment. And again, it depends a little bit on what figures you look at. But let's say, government reports that current energy uses between two and two and a half percent for data centers. And that will increase significantly. And when I say significantly, a fourfold increase by 2030, if all of the centers of planned centers are constructed.

7:29So we've got a problem. And this is this is coming into time when much of the world is looking to decarbonize is is there a way in which this increased demand for power effectively turns that back? If you're looking at electricity demand, AI can help to manage demands and so forth and reduce overall demand and delivery of service. But it's, I would say, the benefits of smart electricity

7:59management will be comparatively small in relation to the impact of AI and the operational energy required to run those systems. Why are we talking about this at all? Given that data can go across borders, why aren't we just sticking these data centers at the place where we have the cheapest, greenest electricity? Why aren't we putting them in Iceland next to a volcano? I don't think Iceland is the best place because of the eruption of volcanoes.

8:33Most probably the Nordics would be a safer option, Norway and so forth. But I think there are several reasons that data centers are now regarded as critical national infrastructure. And there have been lots of reports and research related to how they will boost the economy and so forth. But we are completely reliant on their services, be it for banking, for, you know, smart car charging systems or whatever, as well as leisure activities. But in terms of there's lots of arguments about security,

9:08cybersecurity, data security, if we rely on services overseas, who's to say that there could be some sort of geopolitical disagreement that could affect supply of services. But as well, there's a challenge with speed of data processing. And ideally, as an end user, you'd want to be as close to the data center that's processing your data to be as close as possible. Debra, what does the end of the life of a data center look like? You know, presumably, at some point,

9:41they're going to be replaced by even more powerful data centers, maybe quantum ones. What's going to happen then? It depends very much on how the equipment is designed. There's a history of just designing all computing equipment from sort of drawing board to end of first life without any consideration of what happens to the electronics and so forth after that. But there are definitely moves to change the design of equipment. And it doesn't matter whether it's computing equipment or the mechanical services,

10:13etc. in the data center, but to change the design to facilitate repair, reuse, upgrade, etc. And finally, when products reach the end of their life, to actually be able to separate components and recycle the materials a lot more easily. So that's, in theory, that's possible. But how willing the industry is to embrace that at the moment is slightly debatable.

El Nino and coral records

10:41Thank you, Debra. Okay, time to check in with Roland Pease. Roland, what should we be aware of this week? You've not been paying attention if you're not aware of the evolving El Nino, already set in and forecast a peak around December with temperatures in the Central and Eastern Pacific towards South America, three or so degrees above normal, with a whole bunch of weather impacts to follow. By coincidence, a study just out in science suggests the scale of the developing situation is in line with a trend towards stronger events driven by global warming.

11:16How global warming might affect El Nino's has been hard to unpick because significant events are rare and paper weather records don't go back very far. A team led by paleoclimatologist Julie Cole has made up for the missing data by examining Pacific corals. Corals that grow on these Pacific islands can live for decades or even centuries. And they lay down their skeleton at a rate of about one or two centimetres per year, the species that we use. And so we're able to sample that millimetre by millimetre to get a monthly set of samples.

11:51We measure the chemistry of the skeleton and the chemistry of the skeleton tells us the temperature at which the skeleton grew. And then we also utilize dead corals, coral boulders. They really just look like big grey boulders that have washed up on the beach. And where did you go to get these particular corals? These corals come from the Galapagos Islands, which really lies in the heart of the eastern Pacific side of El Nino. And it lies in the heart of the place where the strongest El Ninos have their biggest impact. So it's really well located for studying the history of strong El Nino events.

12:24I mean, because the striking claim is that the strongest El Ninos are getting stronger. Have I got that right? What we're able to see is that the range of variability in our data from high to low is much greater in the corals that grew since the 1980s compared to those that grew in the pre-industrial. We see that the modern El Nino is 36 percent stronger than the pre-industrial El Nino. We have a transitional period in between with a little bit of data between 1850 and 1980 or so that shows a tendency for increasing variability.

13:00So we see the strengthening of El Nino across the last thousand years. And it closely parallels the change, the increase in global temperature over that time. I mean, you do have a chart, a graph in the paper that looks astonishingly like the famous hockey stick graph where there's sort of wiggles of temperatures going back hundreds of years. And then it sort of shoots up right at the modern era. Yes, I'm glad you caught that. I mean, you're certain this is not a coincidence. Well, I think at first we thought, what if it is a coincidence?

13:33Because we kept expecting to see, I don't know, something different. And we just didn't. We saw we got what we got. And we've analyzed everything in my lab, you know, pretty much. And I don't know what else it is that is changing on that time scale. It has taken us a while to get this out. But, yeah, we could see this trend in our data. And as this event has been developing over the last few months, we've been thinking, well, you know, this is actually not a surprise. Within the known records of El Nidia, I mean, we had some very strong ones.

14:06We had obviously 23, 24. 2015, 2016 was the last really strong one. 1997, again, I think it was 1982. Big one people talk about in 1877. Can you see individual events like those and give us a sense of actually how strong they were? Yeah, our records do show individual events. And we have multiple overlapping records covering the last several decades that show 1997, 98 event very clearly recorded as highly unusual.

14:40All of the records we have, however, show that the coral stopped growing in 1982, 83. You may know that that event killed 95% of all the coral in Galapagos. And our records clearly experienced some dieback. However, they regrew and kept going. So we're able to get more recent periods, but not that particular event because it was so strong. So, yes, we can see individual events. And when we look back at 1877, 78, which is often held up as another super El Nino,

15:11very interestingly, we do not see it as a super El Nino in Galapagos. So that's something I'm digging into a little bit deeper now with my group to try to understand where are the coral records that show it and where are the coral records that do not show it. One of the things we know about that event is from the human history records that document famine in many parts of the world during that time. But we should remember that the famine that may have occurred back then might not be proportional to the temperature change in the tropical Pacific.

15:42And so using that, I think, to sort of say that that was a super El Nino, let's just say that we need to dig into the data that actually come out of the tropical Pacific Ocean to really characterize that. We're doing that now. We're also looking at other parts of the tropics to see if that event was really unusual. And what of the future? The concern then is that on top of the sort of general predictions of the impacts of global warming, there will be additional impact because there could be stronger El Ninos in the future that will disrupt systems even more than we know at the moment.

16:17Yeah, that's the big concern. And, you know, when we think about the impacts of global warming, we know, for example, that the temperature change tends to be concentrated in the highest latitudes, that we're seeing strengthening of rainfall or precipitation, I should say, in the highest latitudes, as well as along the equator. Whereas El Ninos tend to have a characteristic pattern of warm and wet across the subtropics and tropics. And so there's sort of a different spatial pattern, and it might mean that in addition to the impacts of global warming concentrated in certain areas,

16:48there will be impacts of El Nino that may be affecting other areas, spreading the damages. Julie Cole of the University of Michigan on the double whammy of climate change and a strengthening El Nino. Thanks, Roland. You're listening to Inside Science from the BBC World Service. Tell us what science you think we should be investigating. Our email address is insidescience at bbc.co.co.co.co. What if you could get more from what you already do?

17:19When you're a Shell Fuel Rewards member, you can. So your weekday evening could be filled with more rewards, more special offers, and definitely more savings when you're fueling up your car. When you join, you'll save 10 cents per gallon on your first fill, 20 cents per gallon on your second, 30 cents per gallon on your third, and more savings on every fill after that. There are no extra steps and no changes to your routine, just more value built into it. You earn rewards when shopping and dining out, and you always save more at the pump.

17:52So wherever your day takes you, you get a little more out of every stop along the way with Shell. Join the Shell Fuel Rewards program in the Shell app and enjoy life with more. Your nearest Shell station is closer than you think. We'll see you next time.

18:35You can invest your way. No matter your goals or appetite for investing, Schwab has everything you need all in one place. Visit schwab.com to learn more.

Fish sentience and commercial fishing

18:46Look, tomorrow is Friday. Friday, what does that mean? That means Fish and Chips Day, of course, but not so fast. Each week we are interviewing one of the shortlisted authors for one of the biggest prizes in science writing, the Royal Society Trivedi Book Prize. And this week I spoke to Rose George, who wants to change how you feel about fish, or rather, and this matters, about fishes. My book is about fishes and fishing. So it's about the industry of fishing, but it's also very definitely about the creature, the animal that is fish.

19:23And there's also quite a lot of humans in it. One of the interesting things that you explained in the beginning is you use the plural fishes rather than the plural fish, even though fish means lots of fish and one fish. Explain that. I am very suspicious when it comes to animal welfare of singular plurals. I think it makes it very easy to dismiss these, the animals, as a kind of mass or as not individuals. So I wanted to use the word fishes to remind us that these are animals.

19:54These are sentient beings. Yet in most kind of discourse, we talk of them as tonnage and stocks and as if they're fruit. And in fact, you know, in some languages, they are called fruit. We have in French and German and Italian, they talk of, you know, fruit de mer. And they're not fruit. I mean, they're animals. So I'm just trying to rehabilitate fishes to where they should be, which is as animals worthy of scientific interest. And what have we lost?

20:24If we look around the waters of Britain, most of us can't see under the sea. But if we could see under the sea and see what's changed over the last hundred years, what are the sort of numbers that we're talking about? Well, the numbers are pretty horrific. So the United Nations organisation, UNCTAD, says that 90% of it says fish stocks. I'm going to say populations of fishes. But 90% of fish stocks are in trouble. So they're either overfished, they're under pressure. They're somewhere in trouble and they're definitely not in a state they should be in.

20:57But if we look at our supermarket shelves or the fishmonger or the fish counter, there's no gap, as far as you can tell, because the fishes come from elsewhere, because it's a global economy. And so we don't know that we are missing local populations of fish. So things like herring, for example, obviously herring used to be absolutely so abundant you could walk on it like a carpet. That's gone. Yeah, it's hard to think of a fish that's not really in trouble. Well, maybe the Alaskan pollock, but in general, we have depleted the oceans in a quite horrific manner.

21:36But when I want to make myself feel good about myself, I buy line caught salmon or marine certificate certified fish that tells me it's sustainable. Are these things not? The industry, I think, is very good at kind of verbal theatrics. So you just said line caught salmon. Now, a line caught fish, you probably think it's a bloke on a small boat with a rod and a line. But it's really not. Line caught in technical terms can be a line 100 miles long with thousands of hooks.

22:07But it is very difficult. But yes, you can make better choices. But the problem is that the underlying fundamental reality of the state of life in the ocean is so dismal that you're making a less bad choice. And we can care about this because we care about ecology in general and habitats. But your book tries to make us care about it because we should care about fish and fish sentience. What research did you come across that maybe changed or augmented your view of what it is like to be a fish?

22:40Well, this is one of the absolute most compelling triggers I had for writing the book is realising in the last 20 years or so how much work has been done on the very controversial concept that, number one, fish have feelings and are sentient. And are clever and also that they feel pain. One of the most powerful sets of research has been in looking at the wrasse, which is a very popular aquarium at fish. But the cleaner wrasse in particular has been shown to pass the mirror mark test, which is a very big deal in animal cognition and sentience.

23:20If an animal can recognise itself in the mirror. So if you put a mark on an animal's face, including a fish face, and they then recognise themselves in the mirror and they don't attack that image, then they're known to have what's called a self-face. They know who they are and they know what they look like. So wrasse has passed that test, which is extraordinary. You did a lot of different things for this book. You went diving. You went on fishing boats. You spoke and seemed to get on extremely well with fishermen, despite, I suspect, philosophical differences. You vomited on a lot of boats.

23:52I did, yeah. Of all of these, the experiences you had, what left the biggest impression? The two fishermen who I went fishing with, one is a guy called Rex, who is a small boat fisherman, lives in Filey on the East Yorkshire coast and looks like a Viking. I mean, basically central casting fisherman. He single-handedly pretty much has reduced seabird bycatch by changing his nets in Filey Bay, which is when you catch other species as well as what you're trying to catch.

24:24So he's a bit of a star. And the other thing is, the other boat on which I vomited more, actually, was operated by Captain Tom McClure, who, I mean, you say we probably had philosophical differences. I mean, there were absolutely profound differences, and yet we took them through, and he knows what I think, and I know what he thinks, and I respect him. I don't like his fishing method. I don't like bottom trawling. I would happily see it banned because it's just too destructive. But he would disagree with that, and he's got every right to do that, and we're still friends.

24:57So I think it was, I wanted to make these people human. I wanted to get their stories, I wanted to convey that even though I may disagree with the scale of industrial fishing and what it has done, I respect these people. I respect that they are out on the sea, that they understand their environment, probably like no one else. So if that came across, that I warmed to them, I'm glad because, and also Tom was really kind when I vomited four times in a day on his butt. I mean, I'm kind of a terrible person to write a book about fish and fishing because I am horribly seasick, and I haven't touched fish as a food since I was five years old.

25:36So I've no idea what it tastes like, really. Thank you very much indeed, Rose. Caroline, talk to us about science.

The final days of the Swift telescope

25:42What have you looked at? So, Tom, have you been following this attempt to save the Swift telescope? Yes, I thought it was doomed. It is. Well, I'll save the punchline for later. But so the Swift telescope was launched in 2004 and it spent 20 years studying gamma ray bursts, which are super interesting because they're released when massive stars die and go on to form black holes. So incredibly important sort of astrophysics information. And Swift is special because it reacts really quickly, so it can detect and turn towards an explosion in about a minute, whereas if you compare it to Hubble, Hubble moves one degree per minute.

26:20So Swift is really sort of nimble and it can detect gamma rays and x-rays, which other telescopes like Hubble can't. So it's very valuable. But atmospheric drag basically means it's been slowly spiralling towards Earth and it's getting very close to burning up in our atmosphere. The NASA decided to try and save it. They gave Catalyst, a private company, $30 million and nine months to design a spacecraft to try and save Swift. And in July, a spacecraft called Link was launched to kind of chase after Swift, grab it and pull it into orbit.

26:58But sadly, Link spiralled out of control. It lost its fuel. It's not going to be able to save Swift. And Swift is going to burn up in our atmosphere at the end of the year, which I'm really sad about because, you know, there's nothing to replace Swift. And we're just not going to have this information about gamma ray bursts that we used to. What a way to, I mean, it's such a lovely arc. It was going to be a sort of Project Hail Mary style thing. They turn up, they save it. Yeah, no, it's very, and just to clarify, because I think it's fascinating.

27:28We talk about it burning up in our atmosphere, but also atmospheric drag. There isn't really a place where I used to think the atmosphere would stop suddenly and then, but what's happening? Why is it slowing? Yes. So there isn't a sort of concrete line. Like I sort of used to picture the atmosphere as something that wraps around the Earth. Instead, it just gets sort of less dense as you move away from the Earth. And even though, you know, Swift is almost existing in a vacuum, it's not quite. And it basically gradually slows it down. It starts getting a little bit closer to Earth until eventually.

28:01But it's a gradually then suddenly thing, isn't it? Yes. And it's going to be. There's enough drag and then it's just whoosh. And it's going to be suddenly at the end of this year. So time is up for Swift, which is sad. But it also makes me think about how amazing NASA is at building things that last so much longer than anticipated. So this is sad. But Swift was launched in 2004. It was meant to last two years. It still works now. It's only atmospheric drag that's kind of the problem. It could hypothetically keep going. Hubble launched in 1990, was only meant to last to 2005. It's still going.

28:31The Voyagers. Yeah, the Rovers. But are they just under-promising and over-delivering? I always wonder whether this is a PR thing. It's a great PR thing. I'm blown away. It just seems like so many of their spacecraft just last way, way beyond what we're expecting. Well, look, we can't guarantee that you will live to 110 thanks to our episode. All we can really guarantee after all that is that we are all, like space probes everywhere, 30 minutes closer to burning up in our own fiery death.

29:01We hope it was a wise use of those 30 minutes, and Roland Pease will be with you next time when we will all have shuffled a week further along the mortal coil. Until then, it is goodbye from Caroline. Goodbye. And a cheery goodbye from me. We face the greatest challenges of our time. Challenges that test our limits. There is a place where we can find answers. Where journeys break new ground and connections unlock opportunities.

29:34Where innovation can spark real change and our actions can push our world forward. Singapore. Where business events can create lasting impact. Beyond a great point.

29:54You You