Operations | Monitoring | ITSM | DevOps | Cloud

There are cloud outages, and then there are us-east-1 outages. That distinction matters because failures in AWS’s Northern Virginia region rarely feel like ordinary regional incidents. They tend instead to expose something larger and more uncomfortable: too much of the modern internet still behaves as though one place is an acceptable concentration point for infrastructure, control, recovery, and communication. When us-east-1 goes wrong, the problem is not only that workloads fail.

Welcome to April’s Cloud Economics Pulse, CloudZero’s monthly look at cloud spend as AI moves from cost problem to strategic commitment. March’s Pulse called 4.01% a record. It lasted all of 31 days. Why? February’s billing data came in at 4.84% aggregate AI/ML share. That’s another high, another acceleration. You’ve heard it before and it’s getting a bit boring now, but the story isn’t in the numbers; it’s now in the behavior.

A sovereign cloud is a cloud environment that keeps data, infrastructure, and access under the control of a specific country or region. It lets organizations meet strict data residency and privacy laws without giving up cloud speed, automation, or modern DevOps practices. As regulations tighten and AI adoption grows, sovereign cloud is becoming the go‑to model for governments, regulated industries, and global enterprises that need both compliance and agility.

This post is also available in German and in French. Tracking environmental impact can be fragmented, time-consuming, and disconnected from operational data. Beyond simply checking ESG reporting boxes or making sure your company is CSRD compliant, actively monitoring environmental impact is the foundation for building an effective sustainability strategy. At Upsun, we know that measuring progress is the first step toward improvement.

The early era for AI was defined by experimentation, standing up isolated environments, and finding the first practical use cases. Today, the conversation is different. Enterprises are no longer asking whether AI matters. They are asking how to scale it sustainably, securely, and economically. That shift is giving rise to the AI factory: a repeatable, governed, production-ready environment where data scientists, platform teams, and application teams can build, train, deploy, and operate AI at scale.

TL;DR: Most Kubernetes clusters waste GPU compute through over-provisioned pod requests and suboptimal node selection. This guide covers 10 tools that fix this across four layers: resource lifecycle (Kubex, ScaleOps, Cast.ai), hardware partitioning (GPU Operator, MIG, time-slicing), inference serving (Triton, KServe), and observability (DCGM Exporter, NFD). For most teams, the biggest gains are at the resource lifecycle layer: no model changes required.

Key takeaway: Hyperscaler pricing models often penalize e-commerce growth due to unpredictable egress fees and unbounded auto-scaling, but moving to a resource-based allocation model allows teams to treat infrastructure costs as a deliberate business decision rather than a post-campaign surprise. Ecommerce traffic doesn't grow linearly. It spikes, and every spike rewrites your cloud bill.

The regulatory landscape governing UK sovereign cloud security has shifted more dramatically in the past 12 months than in the preceding decade. New legislation, tightened procurement frameworks, and an intensifying cyber threat environment are collectively raising the compliance floor for organizations running cloud workloads in the UK.

Imagine if you had a magic box where you could keep all your business information — sales numbers, customer feedback, everything — safe and sound, but also easy to look at whenever you needed. That’s kind of what Snowflake does, but for big organizations and using the cloud. It’s a new way for companies to store and use their data without getting bogged down by the techy details.

Not long ago, adding a new cost connector to CloudZero was a serious undertaking. We’d task multiple engineers, build in extended review cycles, run a private preview period. But a single connector could take up to two months from kickoff to customer hands. For the major cloud providers, that timeline was acceptable. The size of the investment matched the scale of the integration. But the tools landscape has changed. Our customers’ teams don’t just run on AWS and Azure.