Cloud as a Story - Vunvulea Radu

 In recent years, many organisations have focused on GenAI pilots to demonstrate feasibility, with 2024 and 2025 centered on testing possibilities. By 2026, the main argument is clear: leaders must shift their emphasis from experimentation to delivering measurable ROI, strong unit economics, and predictable costs. AI's future depends on moving from innovation initiatives to products that drive sustainable margins. The FinOps Foundation notes that AI and ML growth is accelerating, introducing higher-cost resources and new spending models such as tokens, GPUs, and managed AI services. FinOps is key to managing these costs. The Foundation cites Gartner’s $644B GenAI spending estimate in 2025 and IDC’s projection that by 2027, 75% of organisations will integrate GenAI with FinOps processes. AI challenges traditional cloud cost models. While classic applications focus on CPU, memory, and traffic, GenAI introduces new cost drivers, including token usage, retrieval latency, vector search,...

 Generative AI is reshaping digital platforms. Previously, cloud programs focused on migration, stability, and cost. Now, leaders seek intelligence, automation, and proactive experiences. This is where Generative AI and agents come in — and why modern cloud platforms must evolve again. Many companies believe GenAI is simply calling a large language model through an API. In reality, GenAI is not just a feature you add on top of the platform. It becomes a new layer in the architecture. It changes how systems are built, how data flows, and how operations work. In our Intelligent Cloud Modernisation for AI-Native work at Endava, we describe this as moving from “cloud-hosted” to “cloud-modernised” and then to “AI-Native”. The difference is important: migration puts workloads in the cloud, modernisation changes how they behave, and AI-Native embeds intelligence into the platform itself. A good way to explain it is to compare AI-enabled versus AI-native. AI-enabled means adding a chatbot...

 With over 1,800 cloud projects delivered by one organization, it’s hard to answer a simple question: what cloud services are actually used where? As projects grow, teams change, and repositories multiply, tracking which AWS or Microsoft Azure services are used in each project, or across industries, becomes a real challenge. This knowledge often exists only in people’s heads or scattered across codebases, making it harder to find than expected. For me, this is not just curiosity. I needed this information for three practical reasons. First, I want to understand patterns across industries and customers. One customer uses a lot of storage and messaging. Another one is more about identity, monitoring, and eventing. When you have many repos, your brain cannot keep this map in its head. Second, RFP questions can be very specific. They don’t ask “Do you have cloud experience?” They ask, “How many projects have you delivered with DynamoDB?” or “Do you have experience with Azure Functions...

 Many organisations struggle to get value from AI, even after moving to the cloud. The main obstacle is outdated cloud infrastructure, which impedes the use of AI. Only with a modern cloud foundation can AI deliver real and lasting business value. But there is one big question that always comes up when people consider investing in modernisation: “How can we show the business value in a simple way, not just with technical terms?” In this post, I will share five metrics we often use with clients. These are easy for non-technical leaders to understand and clearly show how updating the cloud helps unlock AI’s potential. 1. Customer-Facing Throughput First, this metric shows how many customer requests, predictions, or transactions the system can handle in a short period. If an AI recommendation service slows down or cannot scale, customers notice the impact right away. Modernising the cloud increases throughput by allowing systems to scale and process data faster. This results in a bett...

 When adopting Azure Spot Virtual Machines, one of the first challenges is understanding how billing and eviction actually work. Many people initially believe Spot VMs require a minimum runtime, such as 5 minutes. This is not correct. Spot VMs follow the same billing model as standard Pay-As-You-Go VMs: they are billed per second, with a 60-second minimum charge. If a VM runs for less than a minute, you pay for the full minute. After that, you pay only for the exact seconds it runs. The 5-minute intervals you may see on Azure pricing pages are only part of the UI graphs—they are not related to billing. Spot pricing is based on Azure’s unused capacity. Because of this, Spot VMs often come at massive discounts—sometimes 65% to 90% off On-Demand pricing. The exact discount varies by VM family, region, availability zone, and overall demand. When provisioning a Spot VM, you can set a max price bid. Setting it to “-1” means you agree to pay up to the normal On-Demand price. Your VM will ...

 These days, AI tools like GitHub Copilot or ChatGPT feel less like helpers and more like real partners in our engineering work. Now comes vibe coding—a funny-sounding idea with huge impact. Vibe coding means working with AI to express our software intentions and context, rather than focusing solely on precise code. It shakes up how we think about writing, fixing, and running software. Building on this, for years, code was our bridge to machines. We sweated over rules, patterns, and clean syntax to make them understand us. But AI gets our intentions now, straight from the vibe, the goal, the context, what we really want. No more line-by-line translation needed. Of course, this shift isn't without its quirks. The code AI generates isn't always pretty; it might look messy or hard for humans to read. But that's okay. Future SRE and support won't be people, they'll be AI agents. These agents regenerate logic, fix bugs, and tweak things without whining about style. They ...

 Working on legacy applications has always been intimidating. When I began a large migration of a .NET  project or an old Java system to Azure, it felt like going back in time to unravel years of decisions made by people who might no longer be there. Before you can write any code, you often spend days or weeks just figuring out how everything fits together. By the time you understand it all, it’s easy to feel your motivation slipping away. That’s why Microsoft’s new AI-powered tools are such a relief. With the latest GitHub Copilot and Azure features, modernization begins with clarity instead of worry. One major improvement is how the GitHub Copilot App Modernization Agent reviews legacy code. It highlights outdated frameworks, risky patterns, library issues, and migration risks. It’s like having a teammate who already understands the system. When I use it with Dr. Migrate, which checks the app on Azure’s side, I get a much clearer, faster view of what I’m dealing with. The ne...

AI’s true potential comes from advances in cloud platforms, not just from building better models.  Many AI projects run into problems because their cloud environments are not prepared to support them. In this article, I’ll share a practical guide on how organisations can move from a traditional cloud setup to an AI-Native platform in five clear steps. This process is based on what we see with clients whose cloud foundations are not ready for AI adoption. Let’s look at how these five steps can turn a basic cloud into a platform that learns, adapts, and grows. Step 1: Cloud-Native Refactoring Many organisations begin by lifting and shifting workloads into virtual machines (VMs) without changing how their applications are built. These apps still act like they’re running in a traditional data centre, with local data storage, slow scaling, tight dependencies, and all functions bundled together. When you add AI workloads, these systems often can’t keep up. Refactoring for the c...

  In technology conversations, “best practices” are mentioned everywhere—architecture reviews, governance frameworks, and delivery checklists. They are part of how we design and operate digital platforms. But in many projects, especially those with low or moderate workloads, best practices may feel theoretical. They look good on paper, yet the business impact is not always visible.   I recently worked on a project that challenged this perception. We pushed Azure Batch to operate at over 100,000 vCores, stretching the service's limits and placing significant pressure on Azure Storage, Azure Container Registry, and the networking layer. At this scale, every detail matters. And suddenly, all those Microsoft recommendations that previously seemed optional became essential.   1. Best Practices Deliver Real Value When Systems Become Truly Intensive For smaller systems or early-stage products, it is easy to overlook best practices. Everything works fine. For example: ...

For the last decade, the 6 Rs of cloud migration have been used to describe how enterprises should adopt the cloud: Rehost, Replatform, Refactor, Retain, and, sometimes, Retire. The 6 Rs of cloud migration have guided enterprises in adopting the cloud. However, with AI now central to digital transformation, these Rs alone are no longer sufficient. Cloud migration is just the first step; true AI-Native status requires a deeper cloud-native transformation. Customers labelling their migrations as Cloud-Native often have applications that still behave like on-premises systems, resulting in manual operations, static systems, and locked data that hinder AI programs. This is where a new perspective is required to build AI capabilities on top of the 6Rs. Pure cloud-native solutions are difficult for large enterprises. Realistically, we need to identify gaps and what is needed to prepare for AI integration. In the next part of the article, each R will be analysed in terms of AI-Native needs. R...

In the last decade, enterprises migrated thousands of workloads to the cloud for elasticity and lower infrastructure costs. Nevertheless, most workloads behave as they did on-premises: tightly coupled, batch-oriented, and blind to data in motion. This creates high virtualised technical debt with few systems ready for AI. IT leaders should identify and prioritize refactoring critical applications. Early steps include adopting containerization, implementing DevOps, and exploring data integration for real-time flow — establishing an agile, AI-ready ecosystem. This growing intent to re-architect for AI highlights a critical gap between cloud adoption and true cloud modernisation. To better understand what is getting in the way, it’s important to recognize a fundamental insight: Migrations ≠ Modernisations , where Migration offers quick lift-and-shift cloud hosting, but not the value of true Modernisation. From Lift & Shift to AI-Native Moving your servers from a data center to the clou...

  AI is everywhere, part of the IT solutions we build and run today. Having an AI service, a good model, and data is not enough. As for cloud, the real difference is how we build, manage and run the whole solution. Microsoft created the Azure Well-Architected Framework for AI Workloads exactly for this reason — to help teams design AI systems that are reliable, secure, and cost-efficient. The assessment has six main categories that we cover in the next section. Based on the results, we can gain a good understanding of the current AI workload estate and a list of actions to improve how you run and manage your AI workloads . Designing the AI Application The first step in building your AI application is to consider how you will structure it. Using containers for tasks like data processing or model inference helps maintain consistency across the system. This approach makes it easier to update, move, and manage different components. When you have multiple steps in your workflow, such as...

  This week, Cluj-Napoca hosted he 13 th edition of ITDays . With over 800 participants and more than 70 speakers, the two-day conference brought together IT specialists from the area. It was a valuable networking opportunity, allowing us to renew existing relationships and connect with new people. The most notable appearance was that of Morgan Stanley, which joined ITDays as one of its sponsors. With a large delivery office in Budapest, they aroused the interest of many people at the conference. Besides networking, good coffee and chatting with extraordinary people, I had the opportunity to deliver a session about AI-native applications inside the cloud. I presented a five-step playbook for preparing cloud environments and businesses for AI adoption—covering modernisation, data readiness, scalability, governance, and continuous innovation to unlock intelligence and agility. Thank you, Ovidiu, and the rest of the team, for making this conference possible!  

  When companies transition to the cloud, they typically begin with applications and virtual machines, which is often the easier part of the process. The actual complexity arises later when databases are moved. To save time and effort, cloud adoption is more of a cloud migration in an IaaS manner, fulfilling current, but not future needs. Even organisations that are already in the cloud find that their databases, although “migrated,” are not genuinely modernised. This disparity becomes particularly evident when they begin to explore AI technologies. Understanding Modernisation Beyond Migration Database modernisation is distinct from merely relocating an outdated database to Azure. It's about making your data layer ready for future needs, like automation, real-time analytics, and AI capabilities. AI needs high throughput, which can be achieved using native DB cloud capabilities. When your database runs in a traditional setup (even hosted in the cloud), in that case, you will enc...

AI requires data that is accessible, organized , governed, and secure . When storage is disorganized or exposed to the internet, AI becomes slow, expensive, and prone to risks . In contrast, when storage is private, structured, and well-described with metadata, AI operates faster, is more cost-effective, and maintains compliance . This focus is not just on sophisticated models ; it centers on how we store and transfer data . The foundation shapes the outcomes. AI alters the risk profile by consuming data rapidly and broadly . It is advisable to treat storage as a private system by default, regularly discover sensitive data, and integrate these insights into your indexing rules . The use of Private Endpoints, combined with Defender for Storage for malware scanning, and applying immutability provides the basic security feature for Azure Storage .   It’s important to implement them from day zero and not to delay them . It is more cost-effective to implement them fr...