Operations | Monitoring | ITSM | DevOps | Cloud

March signified a very important period for the Konstruct team, where we were able to focus on something we’ve heard consistently from teams: reduce the time to value without compromising control. In the previous post, we walked through how Konstruct 0.1–0.3 established the core platform model, introduced templates, and expanded GitOps into something that can represent both infrastructure and applications. With 0.4, we’re taking a more opinionated step forward.

The UK’s Competition and Markets Authority (CMA) has now set out its latest actions under the Digital Markets Competition Regime (DMCR), following its multi-year Cloud Services Market Investigation. While the regulator has now expanded its focus into business software ecosystems, we must not lose sight of the core issue: the entrenched dominance within the UK's cloud infrastructure.

Back in 2016, Jonathan Ross founded Groq, the AI chip startup, which went on to enter a non-exclusive licensing agreement with NVIDIA for Groq’s inference technology (as part of a $20 billion deal). The name ‘Groq’ is commonly confused with X (formerly Twitter)’s Grok, which was launched in 2023 as a Gen AI chatbot. As demand for real-time AI continues to grow, inference has become one of the most important and expensive parts of the machine learning lifecycle.

While GPUs remain among NVIDIA's flagship products, they also offer a range of other compute products beyond the dedicated graphics cards for which they are known. If you are unfamiliar with the words DGX or HGX, this blog is for you. Throughout this blog, we will cover what these terms mean in practice and when you should be using them.

For seven years, I've watched the same pattern. An organization decides it needs a platform and assigns two of its best engineers. They estimate it will take three months, but eighteen months later, they're still integrating ArgoCD with their secrets manager, still debugging Crossplane providers, and still arguing about how to structure the GitOps repo. What’s happened is they’ve built something that works for one team and can't be repeated for a second.

Every year, NVIDIA GTC offers a glimpse into the future of computing. But this year felt different. The conversations from the past few days point to something bigger than faster GPUs or larger models. The industry is shifting its mindset entirely. GTC 2026 made it clear that the goalposts for AI haven't just moved, they’ve been uprooted. We’re past the point of talking about "faster chips." Everything points to a total shift in the industry's DNA.

This blog is based on the webinar, “AI panel: The next wave of AI technology”. You can watch the full recording by clicking here! The pace of AI innovation is reshaping research, business, and everyday life. However, as breakthroughs in Large Language Models (LLMs) and high-performance computing accelerate, they bring new technical challenges around scale, efficiency, and reliability.

Civo has been identified by ABI Research as one of the world’s leading NeoCloud providers. The report underscores our focus on cloud and AI infrastructure that blends high performance, technical innovation, and strong sovereignty. Being included in the ABI NeoCloud report validates the work we have done to support modern AI workloads while giving organisations control over where and how their data is handled.

Westminster is starting to recognise the importance of technology sovereignty. The recent Westminster Hall debate on technology sovereignty was encouraging to see. For those of us working in the UK technology sector, it felt like an important moment. Conversations about cloud infrastructure, data control and platform dependency have been happening inside the industry for years.

Enterprise private cloud procurement tends to generate impressive security documentation. SOC 2 reports, penetration test summaries, ISO 27001 certificates, detailed descriptions of network segmentation and encryption standards. What it doesn't always generate is clarity on the question that actually matters: does this infrastructure make it possible to operate securely at the level your organization requires, given your specific workloads, your regulatory context, and your threat model?