CUDO Compute

Everyone wants to talk about GPUs. But the real bottlenecks in AI infrastructure? Power, delivery timelines and whether sites can actually support what’s being deployed. Great discussion at @Engineering_SES's roundtable with our CBDO & Co Founder, @CudoPete on what the UK data centre sector is now up against as AI demand accelerates. Because the challenge is no longer just getting access to compute. It’s getting infrastructure live, operational and scalable in the real world.

For the past few years, one question has dominated the AI industry: how quickly can we build and deploy it? That focus has driven extraordinary innovation. But as organisations move from experimentation towards deployment at scale, a different set of questions is becoming harder to ignore. Can we access enough power? Is the surrounding infrastructure ready? What happens when demand for compute grows faster than the systems designed to support it? AI does not operate in a vacuum. Every model depends on physical infrastructure, viable economics and the operational capacity to support it. The challenge is no longer solely what the technology can do, but whether the foundations required to deliver it can keep pace. Read our CEO, @HawkinsTech, latest piece on why AI's biggest challenge isn't technology: itbusinessnet.com/2026/07/why-ai…

Recorded live at Datacloud Global Congress 2026 in Cannes, the Inside Data Centre podcast highlights how GPU-driven infrastructure is reshaping the market. Hyperscalers, traditional cloud providers and emerging AI infrastructure companies are often positioned as competitors. In reality, their roles may be far more complementary. “Neoclouds are disruptive as they represent a new layer, a level of demand in the market that complements hyperscalers” - @CudoPete, CBDO and Co-Founder of CUDO Compute. What actually defines a neocloud and how it differs from an AI factory? Find out more here: open.spotify.com/episode/5R2Hot…

Our takeaways: Companies now have a clearer understanding of their position, and it's evident that go-to-market strategies across the industry have matured. These conversations matter because building the next generation of AI infrastructure takes collaboration across the entire ecosystem, and we're excited about the conversations that will continue long after the event.

Our Engineering team was on the ground at the @RaiseSummit in Paris last week. Packed sessions highlighted the demand for high-level discussions among senior leaders shaping the future of AI infrastructure. Throughout the week, we had valuable conversations with teams from @Vertiv, @Scality, Vanta, @MirantisIT, @ifs, @andromeda_ai, @Lenovo, and Global Scale. A highlight of the week was meeting Ryan Deans, AI Lead at @DDNintelligence, over dinner following their very own AI Summit, and connecting with Vladimir Prodanovic, Principal Program Manager at @nvidia, and Scott Constable, Director of European Sales at @netris_io.

“AI's natural limit is electricity, not chips." - Former @Google CEO @EricSchmidt, @RaiseSummit 2025. A year on, that conversation has only become more relevant. As organisations move from experimentation to production, the challenge is no longer just building better models. It's securing the infrastructure needed to deploy and scale them successfully. Compute capacity, power availability, networking and operational expertise are becoming the foundations of enterprise AI. That's why the CUDO Compute team is in Paris this week for RAISE, joining leaders from across the AI ecosystem to discuss the infrastructure strategies shaping the next generation of AI deployments. If you're attending, come and meet Barry Kick, Aleks Fox, Rafael Griman and Mohamed Alibi. We'd love to discuss how organisations are building resilient, production-ready AI infrastructure.

The CUDO Compute team is in Paris today for the @DDNintelligence AI Data Summit, joining partners, industry leaders and the wider AI ecosystem to discuss the future of AI infrastructure. As organisations move from AI ambition to production, the conversation is shifting beyond GPU access to the infrastructure needed to deploy and scale AI successfully. If you're attending the summit, we'd love to connect and talk about the infrastructure strategies shaping the next generation of AI deployments.

CUDO Compute's HPC Engineering team will be at @RaiseSummit in Paris next week. Joining us at the event are Aleks Fox, Rafael Griman and Mohamed Alibi, alongside Barry Kick, AI Business Development Director. If you're planning AI infrastructure, scaling GPU capacity or exploring enterprise AI deployments, we'd welcome the opportunity to connect. See you at the Carrousel du Louvre: raisesummit.com

AI infrastructure has outgrown the cloud model. Training workloads now demand thousands of tightly coupled GPUs, far higher rack densities, and infrastructure built around power, cooling, networking and data visibility, not just compute. Read more: cudocompute.com/blog/how-ai-in…

There’s no one-size-fits-all blueprint for AI infrastructure. Every deployment has unique technical, operational and commercial requirements. Join Barry Kick, AI Business Development Director at @CUDO_Compute, at the @DDNintelligence AI Data Summit in Paris on 7 July as he explores what it takes to build a modern AI factory. Register now: ddn.com/lp/events/rais…

The UK can currently supply around 6GW to data centres under existing clean energy plans. Meanwhile, around 125GW of demand sits in the grid connections queue. As AI investment accelerates, infrastructure is becoming just as important as policy. Power availability, data centre capacity and long term infrastructure planning will shape where AI is built and deployed. In his latest open letter to the next Prime Minister, our CEO @HawkinsTech explores why infrastructure may be the UK's defining AI challenge, and what needs to happen if the UK is to turn AI ambition into delivery. Read the full article: linkedin.com/pulse/dear-pri…

For years, AI strategy was largely a software conversation. Today, it's becoming an infrastructure conversation. A recent Fortune piece examining Europe's growing focus on AI sovereignty cited CUDO Compute research showing that geopolitical uncertainty is already influencing where organisations choose to deploy AI workloads. The shift is bigger than sovereignty alone. Power availability. Infrastructure access. Capacity constraints. Geopolitical risk. These factors are increasingly shaping AI deployment decisions just as much as performance and cost. The organisations that win in AI won't simply have access to the best models. They'll have access to the infrastructure required to run them. Read more: fortune.com/2026/06/23/eur…

Cloud computing taught organisations to treat infrastructure as someone else's problem. AI is changing that. Today's training and inference environments place demands on power, cooling and physical infrastructure that the cloud era was never designed to accommodate at this scale. It turns out the hard part wasn't getting access to compute. It was everything underneath it. Read more: cudocompute.com/blog/how-ai-in…

AI priorities are diverging across regions. While UK organisations continue to focus heavily on GPU availability and cost, organisations in Europe and the US are increasingly focused on model performance. The findings suggest that AI markets are maturing at different rates, with infrastructure access, cost and deployment priorities varying significantly by geography. Explore the full findings in our Land. Power. Compute. report: tinyurl.com/Land-power-com…

For decades, technology strategy was measured in software, hardware and performance. Today, AI infrastructure is increasingly measured in megawatts. Power availability is becoming one of the defining factors influencing where AI workloads are deployed, how quickly capacity can be brought online, and which regions will emerge as AI leaders. The conversation is shifting from chips to infrastructure. From GPUs to megawatts. Discover the findings in our latest industry report, Land. Power. Compute. report: tinyurl.com/Land-power-com…

96.7% of UK organisations expect power availability to influence where AI workloads are deployed in future. The conversation around AI infrastructure is changing. For years, the focus was on GPUs. Today, organisations are increasingly grappling with a broader challenge: aligning land, power and compute quickly enough to support AI growth at scale. Land takes years. Power takes years. Compute evolves annually. Alignment is now the challenge. Explore the report today: cudocompute.com/land-power-com…

For years, AI infrastructure conversations have centred on GPUs. But the economics of AI are changing. Research from our industry report, Land. Power. Compute. found that long term power contracts, network and data transfer costs, and power availability now rank above GPU pricing as drivers of total AI compute cost. The question is no longer just whether you can access the GPUs you need. It's whether you can power them. In our latest blog, we explore why power is becoming one of the defining factors in the cost of AI at scale. Read more: cudocompute.com/blog/why-power…

The AI conversation has spent years focused on chips. But securing GPUs is only part of the equation. In his latest article for @UniteAi, @HawkinsTech explores the growing attention being paid to energy availability, grid capacity and the infrastructure needed to support AI at scale. As demand continues to accelerate, these are becoming strategic considerations rather than operational ones. Read the full article today. unite.ai/ai-infrastruct…

Final day at Datacloud Global Congress 2026 by @BroadGroup_Intl . For @CudoPete and Michael Jones, the week has been less about the headlines and more about the conversations happening behind them. From AI infrastructure deployment and data centre capacity to power, operations and delivery, it's been valuable to hear first hand how organisations across the ecosystem are approaching the challenges and opportunities ahead. A huge thank you to everyone who took the time to meet with the team this week. It's been a great opportunity to reconnect with familiar faces, make new connections and exchange perspectives with the people building the infrastructure that will underpin the next generation of AI. Until next time, Cannes.

The AI race is increasingly being decided by physical infrastructure. Power. Land. Capacity. Delivery. That's why we're at Datacloud Global Congress this week. @CudoPete and Michael Jones are on the ground discussing the challenges and opportunities shaping the next generation of AI infrastructure. If you're in Cannes, let's talk.

As GPU clusters become denser and power requirements continue to increase, cooling is becoming a critical consideration in how AI infrastructure is designed, deployed and scaled. In our latest blog, we explore the technologies shaping the next generation of AI environments, from immersion cooling and direct liquid to chip systems through to advances in air cooling, and what they mean for performance, efficiency and future capacity planning. Because delivering AI infrastructure at scale is no longer just about compute. It's about managing the heat that comes with it. cudocompute.com/blog/cooling-i…