# Inside one of the world’s fastest supercomputers - URL: https://diginpl.com/jhalak/v/inside-one-of-the-worlds-fastest-supercomputers-bd0059712b534606b91786a90e017a1b - Published: 2026-10-09T15:13:38Z - Length: 6:23 (PT6M23S) - Creator: PADAM SUNDAR KAFLE (@padam.sundar.kafle) - Language: English - Credit: BBC News , Youtube - Views: 0 · Likes: 0 · Comments: 0 - Topics: Inside, worlds, fastest, supercomputers, BBC, #supercomputer, #fastest, #bbc, #news ## Description Inside one of the world’s fastest supercomputers BBC News BBC News ## Chapters - [0:00](https://diginpl.com/jhalak/v/inside-one-of-the-worlds-fastest-supercomputers-bd0059712b534606b91786a90e017a1b?t=0) Air Conditioned Chillers - [1:10](https://diginpl.com/jhalak/v/inside-one-of-the-worlds-fastest-supercomputers-bd0059712b534606b91786a90e017a1b?t=70) Water - [3:00](https://diginpl.com/jhalak/v/inside-one-of-the-worlds-fastest-supercomputers-bd0059712b534606b91786a90e017a1b?t=180) Blades - [4:10](https://diginpl.com/jhalak/v/inside-one-of-the-worlds-fastest-supercomputers-bd0059712b534606b91786a90e017a1b?t=250) Millions - [5:40](https://diginpl.com/jhalak/v/inside-one-of-the-worlds-fastest-supercomputers-bd0059712b534606b91786a90e017a1b?t=340) Supercomputers Stay Cool ## Transcript _Automatic speech recognition; may contain mistakes._ [0:00] Stunning Lakeside views and postcard perfect scenery. [0:08] Lugano is tucked into Switzerland's southernmost canton. [0:14] But aside from its beautiful landscape, [0:17] there's one other jewel in the crown of this Swiss [0:20] Mediterranean city. [0:23] Computation of power. [0:27] This is the home of Alps, not the mountains, but a different kind of giant. [0:33] One of the world's most advanced supercomputers. [0:36] It's the latest generation of high performance computing. [0:40] Imagine you have your laptop. [0:42] But now imagine that you have 10,000 or 20,000 of those laptops. [0:48] So here we look at extremely capable processors [0:51] that are more capable than your laptop processor [0:53] and we're looking at nearly 11,000 of those. [0:56] That's a supercomputing center. Wow. [0:58] But such power comes at a cost, hate and a lot of it. [1:04] This isn't your average data center full of noisy fans and air conditioned chillers. [1:10] Instead, what's impressive about this [1:12] place isn't just its brain power, [1:15] it's how it's [1:16] rum, built not only for speed but also sustainability. [1:21] Alps draws from the same source that has [1:23] sustained Switzerland for millennia, glacial [1:26] water. [1:27] "When you think of supercomputers, you don't necessarily think of pipes and pumps." [1:31] What is happening in here? [1:33] So supercomputers burn tens of megawatts. [1:38] And so that turns into heat that you have to somehow remove. [1:43] And so the setup we have here is we take water from the lake. [1:47] And this is the big pipe coming in. [1:50] And then this water goes into a heat exchanger [1:55] where the water stays in the center. [1:59] The centre typically cools around 10 to 15 [2:02] megawatts, but has the capacity to go up to 25. [2:07] Water from Lake Logano is drawn in, where it's typically 7 degrees Celsius. [2:13] Once used to cool the vast array of technical infrastructure, [2:16] it goes back to the lake around 15 to 20 degrees Celsius. [2:20] The inherent problems that we have is we need water for cooling. [2:25] and we need electricity to power the machines. [2:28] The two don't go well if something happens. [2:31] And this is why we are on the raised floor [2:34] and so should something happen, [2:37] you know, some pipe breaks or so and water comes out, [2:41] it will be collected here [2:43] and then this is really separated into different sectors [2:48] so that the water damage over here doesn't move. [2:51] on the other side. [2:53] One level above, the pipes may be smaller, [2:56] but water is still very much at the heart of the supercomputer centre. [3:00] Each of the 2,688 blades has a blue and red pipe, [3:06] colour-coded to avoid confusion during planned maintenance. [3:11] So this is the inside of one of these blades, [3:14] one of these many blades in this system. [3:16] What are we looking at here? OK. [3:19] Here you have a grays and a hopper [3:23] processor, so a multicore CPU and a GPU, [3:27] H100 GPU. [3:28] And they are of course consuming a lot of power, [3:31] so that's why the liquid cooling goes directly on the processor. [3:35] So that's what these copper pipes here are. [3:40] Where the cold water comes in here, is distributed, and then comes in... [3:46] the warm water goes out back into the data center. [3:53] By crunching vast amounts of environmental data, [3:56] Alps helps forecast sea level rise, glacial retreat [4:00] and extreme weather events on a global scale. [4:04] AI is basically you're building a predictive model based on data. [4:10] So you could, for example, increase computationally [4:14] the level of CO2 in the atmosphere. [4:16] What happens then? [4:17] What you can now do with those simulations [4:19] is you can take your little understanding [4:21] of reflective properties, [4:22] but you can now scale it up to the whole planet. [4:25] So these are all things that are not easy to understand, [4:28] but you can run it in massive simulations, [4:30] like on this machine. [4:31] Not far from Alps is the Condra Dam. [4:34] It's the fourth tallest dam in Switzerland, standing at 220 metres high. [4:40] It's source, the Verzasca River, [4:43] which is heavily influenced by rain and snowfall in the [4:46] surrounding mountainous region. [4:49] This valley is famous for its extreme precipitation [4:54] events, so it holds some of the world records of [4:57] precipitation. [4:59] What we are doing, we are using the supercomputer [5:03] ops in Switzerland and on global climate [5:06] simulations. [5:07] One of the models aims to predict the impacts of rising CO2 levels on the planet. [5:13] They do this by performing a series of tests [5:16] replicated on a worldwide scale to see how the [5:19] atmosphere reacts. [5:21] If you sort of think of the Earth, we have millions of... [5:25] We have grid points or locations where we [5:28] actually sample the environmental conditions [5:30] like temperature, wind and so on. [5:32] And then we have to go up in altitude, 150 levels. [5:36] And then we have millions, millions of computations. [5:40] It's basically impossible to run on a normal computer. [5:46] In a world facing both data overload and environmental crisis, Alps [5:52] is a chance for technology and nature to work together, [5:56] setting a standard for how supercomputers stay cool [5:59] without heating the planet.