In The Arena by TechArena - Scaleway's Albane Bruyas on Europe's Sovereign AI Cloud Era

Episode Date: June 25, 2026

Allyson Klein and Solidigm's Ace Stryker sit down with Albane Bruyas, CCO & COO of Scaleway, for a thoughtful conversation on what it takes to build a sovereign European cloud designed for the AI era.

Welcome to Tech Arena, featuring authentic discussions between tech's leading innovators and our host, Alison Klein. Now, let's step into the arena. Welcome in the arena. My name is Allison Klein. We're coming to you from OCP Summit in Barcelona, and this is a Data Insights episode, which means I'm with a striker. Welcome, Ace. Thank you, Allison. It's great to be here. Welcome back in the arena. I'm very excited to have a great guest with us today. as well. We have Alban Bluis from Scaleway. She is chief operating officer. Welcome. Thank you very much. Happy to be there. So Alban, this is your first time on the arena.

Why don't you just go ahead and introduce yourself in your role at Scaleway? I am at Scaleway, which means that I need all the teams that manage the infrastructure on which relies our cloud. This means all the hardware part and the data center part. And then I'm also in charge of international expansion because of course international expansion is first about data center and hardware and finally need the customer services. All right. I had a chance to catch your colleague in the panel this morning to kick off the show and hear a little bit about what Scaleway does. And what I'm hearing is the company describes itself as both a sovereign European cloud provider and

neocloud built for AI infrastructure. What does that combination mean? How do you think about what that means in practice and why does it matter right now? I think it's really important for the clients because we bring, let's say, the good parts of the two worlds. I think you cannot just do training. You have, of course, a lot of needs in the type of storage and type of inferences. So what Scaleware offers is really to have this 25 years of experience in terms of cloud. So very mature public cloud, sovereign public cloud. And at the same time, you have this capacity of training that is very useful. for all the AI growth that we have in all the companies

that we see that in Europe, we have more and more companies that are going to AI and using these training capacities. When you have organizations come to you for services, obviously they may be looking at other providers. Why do they come to you? What are they specifically seeking from Scaleway? And where do you see some of the large multinational hypers

falling short? None of questions. So why do they come to us? And they comes to us because we have now a catalog, which is a pure, mature public catalog, that we have a unique capacity in Europe to be, I will just let on the side the sovereign term, because it's really dangerous, depends on what, not dangerous, but at least there is very different definitions. So what we sell more and what is very interesting for our clients is that we sell a full autonomy. So this means that we have control on all the value chain.

And this is what makes the pure difference for our clients. It's not just, I will serve in our European, is that we guarantee openness on software, on hardware. We guarantee transparent price because we have a full control net. We guarantee the transit because we are part of a huge tackle group. So we guarantee the software because we are not dependent on any other company in terms of software. So it's really a full transparency and a full control that is quite reassuring for European and even other than companies than only European companies. And this control is also a capacity of control on the price.

And this is really interesting because we have this capacity. And finally, what we propose and what has a lot of interest for our clients is all the sustainably part, which is the fact that again when you control and when you build, yourself, you have much more capacity to optimize. And so we have optimized the data centers in France. We have optimized our code. We are optimizing our hardware. And we have the capacity, thanks to OCP,

and thanks to all the team's work to give a good visibility on the carbon footprint. And a good carbon footprint because actually we have made all the efforts. Yeah, awesome. The company's commitment to sustainability, to transparency that you're talking about, clearly has a lot of merit and is driving success for the company. As I understand it, there are unique challenges as well to scaling AI compute in Europe. Can you talk to us about how you think through some of those, whether it's power infrastructure, deployment, timeline, supply chain, where does the process most often stall?

And how do you think about it maybe differently than competitors? I think the chance that we have is that we are French and in France, we have a lot of capacity in terms of power. So I wouldn't say, I don't think it's too much about power infrastructure. It's more interesting to understand that it's not that producing the power, which is important. It's like to bring the power where the data centers are. And this is the real topic. So the first topic is here.

And then what we lobby for, of course, because this is, again, being in control for us and for our clients. To make sure that this power, which is available, goes to data centers detained by European. funds, companies, but anyways, that European have the capacity and found the capacity to build the data centers and have the power. So this is the first thing. Then when it comes to, when it comes to hardware and all the supply topics that we have at that moment, again, it's all the work from our teams working on the OCP, working on visibility. So there is really two criteria. First, if you have the most open hardware as you can, then you have the capacity to buy from different suppliers. And if you have different

capacity to buy from different suppliers, you can have better price, and you can have more capacity because you can go different places. So Scaleway is very president at CPMEA this week, and you're presenting on a number of topics, but one that caught my eye is how you're integrating MHS Open Rec, ORV-3, and Open Facilities for Acceleration of AI Infrastructure. Can you talk a little bit about how you're utilizing these industry standard techniques to help address client requirements. Again, it's the best way when you standardize

and when you are in the OCP community, this is the best way for us to gain information and learn from our peers and at the same time share what our ideas and to have critics that enables us to grow much faster. So it's always the same is that you save time because you think as a group. So you save time because you can concentrate

on your core business because a group has validated it and all that. And finally, you save time because you have the chance to have more providers. Nice. I'm curious about one of the challenges that we face at Solidime sometimes where we make high-capacity SSDs is that procurement teams for so many years have been used to just buying hard drives as a default choice. It's an easy

decision to make. And you're obviously up against the hyperscalers of the world where a lot of procurement teams probably have a similar thought, right? That's the default choice. That's the safe choice. It doesn't even really take a strong argument to go in that direction. So how do you make your case to those procurement teams to consider you as an alternative? What's the kind of differentiating value that you bring?

When I talk about control, this is really something that talks to a procurement team. And why I talk so much about control and value chain is because I'm a buyer at the beginning. So I fully understand what they're thinking. and a buyer, a procurement team, wants to make sure that they are safe on their procurements. And what is changing right now, and what I lobby since the beginning, I'm working for Skalway, is that I'm always really surprised to see how in the digital world people are just forgetting that their principal, strategical supplier is a unique supplier. They cannot move out from.

And there is no solution if there is a crash, there is no solution if there is a crash, There is no solution if there is different prices that come out, no solution if there is an external government that asks for data. You have no choice. Whereas I come from a world in the industry or if you have a supplier for a strategical component and you have this, I'm sure. You don't have one factory. You choose at least so because you want to make sure if there is a fire, I still can buy.

And this is something in digital economy we are sometimes forgetting. And so this is what we say to the big clients that will go say, you will never be fired because you choose one of the aposcalo. So it's just like test an alternative. You need to have one and you would be happily surprised and therefore we will go for a long time together. Nozky always gotten a lot of attention for its collaborations with companies like Mistral and Hugging Face in terms of showcasing AI development in Europe. What kind of ecosystem are you trying to build in Europe? and how does that reflect on Europe's broader opportunity with AI development? First, we want to be close to them because it's our business.

They are our clients and we share a lot of clients. Of course, it's important to us. Because again, this is our mission, let's say, of sovereignty for Europe. And so let's not think only about our value chain, but all the value chain. And so there is scale way and then there is a ging face and then there is mistral. And the three together, we can really build something which is reliable of high quality. And altogether, it's an alternative that we give to European, and we need to be good enough. So working together is a bit the same as OCP.

The most we work together, the most we have an offer which is sustainable and which will be good enough to be presented in front of hyperscalilus. I'd like to ask a question about infrastructure. Yeah. And a lot of companies are grappling with this as use cases evolved. We're seeing in the world of AI this shift from simple LLM input-output-type interactions where you're asking for simple advice to these longer, more complex, agentic workflows now. And there's all sorts of implications to that, right, in terms of the hardware and software that's needed to support that.

Now you're dealing with lots of more context memory. Latency becomes a very big concern as you pull in more capable tools to deal with more complex problems. So how is the scale way infrastructure evolving to keep pace with that change on an agentic AI first world? What we think is that we have very different needs. And so in this world where you have the agentic, but you still have the training and all that, our goal is to have all the solutions, main solutions. I think we are working exactly the same way as I was talking about our clients that should choose hyperscalular plus something else. We are looking at all the technologies and we want them to be available for the clients. So we have all types of GPUs.

We are working with those which are developing them right now because we want to have them and we want to make sure they go to the requirements. We are building high level clusters and always taking the last technology, but we also have a full capacity of GPUs in our cloud ecosystem because, yes, there is so many differences that we need to have a bit of everything. And then for inference, you have this very specific GPUs, also some CPUs that could be interesting. And so we really try to have all the whole in scale way history. We have been known because we were the first to do arm servers. And we want to continue to be like really at this level of technology where we can put something in place that nobody has.

And this is really important for us. When you look at a three-year time horizon across infrastructure, ecosystem, and policy, what do you think needs to be in place to ensure that European-based providers are at the top of the market in terms of cloud services and region? You mean in terms of regulation? Just in terms of what you're doing with the value chain to ensure that your infrastructure is differentiated, your ecosystem partnerships, and then policy that is in place to support you? We need to develop the ecosystem.

This is why we work with Mr. Allen-Uggingface, but this is also why we work with all the chip designer in Europe. There is a lot of them. Some are present here at E.C.Bassadarra. So we want to work with all the people that are capable to come in the value chain to make sure that we build this European infrastructure. So it's not only, we don't really think that it's European grants that will help. It's really that you build an ecosystem thanks to particularly.

They saw those. So when we can buy European, we buy European. And when they need some information, we give all the information that helps them really build all the ecosystem. And then we need regulation, but European regulation. And I think this is the most important part. We need to stop this regulation per country and make sure that more and more we go through European regulation, this means that we have a huge market and not too partialize market because so many countries and so many regulations. And then we can really compete because we have a huge market that can address easily, which is not really for the moment. Okay, last question, Alvin, which is if folks want to learn more about Scaleway and particularly your work within the OCP

ecosystem, where should they go? So many ways. Our blog, scaleway.com, you can go to CP, I have members of my team which are leading one project on your new clouds and AI. So really follow that because it's really important. And contact me directly on LinkedIn. I will be happy to tell you more about us. Thank you so much for being on the show today. Tech Arena is a big fan of Scaleway. So we're super happy that you're on the show today

and spent some time with us. Thank you for your experience. Thank you. Thanks for joining Tech Arena. Subscribe and engage at our website, Techorina. All content is copyright by Tech Arena.