Modernizing Cloud Infrastructure in 2025: Key Insights for Every CEO

2025 is reshaping the game: cloud has become the core of corporate strategy, influencing financial efficiency, independence, and resilience. This article explores how to manage costs through FinOps, strengthen trust with sovereign cloud and ESG practices, and leverage Generative AI and Serverless 2.0 to accelerate innovation.

We examine the evolution of trust, from Zero Trust to Adaptive Trust, the growing importance of multi-cloud strategies, and the need for systematic technical-debt management. A special focus is placed on the role of architects and high-performing teams as the foundation of future growth.

2025: A Turning Point for the Cloud

Over the past decades, cloud computing has become the foundation of the digital economy. Ten years ago, only about 30% of corporate data resided in the cloud; by 2022, that share had doubled.

By 2025, the cloud has firmly established itself as the core of IT strategies: more than half of all workloads now run in public and hybrid environments, and the global market has surpassed $912 billion.

If these numbers still don’t convince you, let’s look at a few key facts from recent years:

• IT budgets have been restructured.

Companies now spend more on public cloud services than on on-premise solutions. SaaS alone reached nearly $300 billion in 2025, while the entire public cloud segment approached $723 billion. These figures make it clear: cloud has become one of the main drivers of competitive advantage, although success still depends on many other factors, from team capability to business model.

• AI sets the pace.

Around 72% of organizations are using Generative AI (from chatbots to content creation) and 79% are implementing AI/ML more broadly. None of this would be possible without the computing power of the cloud.

• Multi-cloud as portfolio diversification.

Nine out of ten companies now combine multiple cloud providers to reduce risk and avoid dependency on a single vendor.

But what has driven such rapid progress? Throughout history, disruptions have always been catalysts for growth. Even everyday culture offers its reminders.

According to legend, the exiled Polish king Stanisław Leszczyński, who ruled Lorraine in the 18th century, was so weary of wartime hardships that he demanded something sweet every day. The popular Kugelhopf cake seemed too dry to him. One day, in frustration, he struck the table and accidentally spilled a bottle of rum, soaking the cake and giving birth to the dessert we now know as rum baba. Historians may debate the details, but the story itself illustrates a simple truth: even an accident in difficult times can spark something new.

The same applies to science and technology. World War II gave rise to the Manhattan Project, led by Robert Oppenheimer. Under the pressure of circumstances, humanity achieved in just a few years what would have otherwise taken decades in peacetime. But with that breakthrough came a moral dilemma. After the first nuclear tests, Oppenheimer remarked: “Now I am become Death, the destroyer of worlds.” He understood that his team had created a weapon capable not only of ending a war, but also of threatening the very existence of humanity.

The same is true for the cloud. In 2020, the world was struck by a devastating pandemic that pushed everyone out of their comfort zone. Education went online, companies shifted employees to remote work, and familiar processes had to be rebuilt within weeks. To survive, businesses needed instant flexibility, launching online services, supporting remote sales, and avoiding costly hardware investments.

That’s when the cloud took center stage. The domino effect was clear: as industry leaders like Amazon and Google doubled down on cloud services, others rushed to follow. This wave of migration brought valuable lessons, learning from others’ mistakes and best practices helped organizations understand how to do things better.

By 2025, the challenge has changed: it’s no longer just about moving to the cloud, but about managing it as a strategic asset and extracting real business value. That’s why today’s CEOs focus on three key priorities: FinOps for optimizing cloud costs; Sovereignty for ensuring independence from vendors; ESG for driving sustainable growth.

From Implementation to Optimization: Three Key Priorities

FinOps: Every Dollar Under Control

FinOps is a discipline that helps organizations manage the cloud as a business asset.

If the first generation of FinOps was mostly about turning off forgotten servers, then FinOps 2.0 is about culture, where IT, finance, and business teams work together, linking every dollar spent to measurable business value.

According to the Flexera 2025 State of the Cloud report, the challenge is real: 84% of organizations name cost management as their top concern. Nearly one-third of companies spend over $12 million per year on public cloud services, roughly equivalent to maintaining an entire IT department, often exceeding planned budgets by 17%.

These figures reflect the aftershocks of rushed migrations, when companies, chasing speed, failed to keep spending under control.

Why does this happen? Because cloud provider invoices can feel like a labyrinth. For example, up to 49% of storage costs don’t come from the data itself, but from fees for operations and access. A forgotten test database or an extra API call can easily turn into thousands of dollars in unexpected charges. To visualize it more simply, imagine your home hard drive, it’s only 100 GB, but half the space is taken up by old apps and files you no longer need. Instead of cleaning it up, you just buy another 50 GB. The same thing happens in the cloud when forgotten files turn into financial ballast. And 42% of companies end up transferring more data than they initially planned, leading to multiplying costs as a result.

About 42% of companies migrate more data than planned, only to face cloud bills that multiply as a result. FinOps 2.0 is designed to break this cycle. Instead of reacting after the fact, it introduces a systemic model built on three pillars:

• Unified metrics: Cost per transaction, per session, or per user: clear, comparable indicators that make sense even at the CFO level.
• Standards like FOCUS: A consistent format for analyzing and comparing invoices across multi-cloud environments.
• ML-powered automation: Detection of usage anomalies long before a cloud bill turns into a seven-figure surprise.

Turning cloud usage from chaos into a controlled process isn’t about cutting every cent. It’s about making the real connection between cloud spend and business value visible. That’s why FinOps 2.0 is increasingly viewed as a strategic function: it doesn’t force teams to save at any cost, but helps them decide where paying more drives speed and profit and where waste can be eliminated with zero impact.

But money is only one side of control. The other is jurisdiction and independence. Today, roughly 70% of cloud budgets go to a single provider, creating critical dependency. That’s where the next priority comes into focus: sovereign cloud.

Sovereign Cloud and the New Rules of the Game

Storing data “in Europe” is not the same as storing it under European jurisdiction. A data center may be physically located in Lyon, but if the cloud provider is registered in the United States, the CLOUD Act allows U.S. authorities to request access to that data. For businesses, it’s like building a fortress on someone else’s land.

And what happens to a fortress during political tension? Exactly: a company relying on foreign jurisdiction risks losing control over its own data at any moment.

A sovereign cloud is an environment where both data and governance fall fully under local jurisdiction, and where infrastructure decisions reflect the interests of the region. At the European Sovereign Cloud Day 2025, experts highlighted a growing concern: the absence of European players among the world’s top eight cloud providers is now considered a strategic risk. The trend is clear - organizations are increasingly seeking independence.

Demand for sovereign solutions is accelerating under pressure from new regulations, the Data Act, Digital Markets Act, and EUCS certification, all of which require strict oversight and control over data. The U.S. sanctions against Huawei (2019–2023) and the EU’s regulatory responses demonstrated a simple truth: cross-border infrastructure is vulnerable.

A sovereign cloud acts as a jurisdictional shield. It protects data, reduces systemic dependency, and, importantly, fuels economic growth. Companies investing in technological independence are simultaneously opening new opportunities for the region’s digital ecosystem.

As a result, the sovereign IaaS market is expanding fast: from $37 billion in 2023 to an expected $169 billion by 2028 (a 36% CAGR, compared with 24% in the traditional cloud segment). This shift not only creates jobs and protects intellectual property, but also strengthens customer trust: 84% of European companies already use or plan to adopt sovereign solutions.

In practice, the sovereign cloud takes several forms:

• Hybrid models built in partnership with hyperscalers.
• Federated architectures, enabling data exchange under EU-defined standards.
• Alliances of local European providers, forming region-first cloud ecosystems.

Beyond Europe’s major cloud players, specialized providers focused on local markets and strict compliance are gaining importance. These providers fill a critical niche: they enable companies to build scalable services without the risk of cross-border data exposure.

For mid-sized and fast-growing companies, this is becoming a real alternative to hyperscalers preserving control while lowering the barrier to adopting sovereign solutions. All of this points to a clear shift: cloud strategy in Europe is no longer just about price and SLA. Today, it is defined by jurisdiction, geopolitical risk, and certifications such as EUCS and SecNumCloud. The sovereign cloud is evolving from mere infrastructure into a new dimension of competitiveness.

Growth of the Sovereign IaaS Market (2023–2028)

The source: BCG, 2025.

Independence helps solve the problems of data control and trust. But that alone is no longer enough. Companies now recognize that both customers and regulators expect the cloud to be environmentally responsible as well. ESG is emerging as the next competitive factor: if FinOps is about money, and sovereignty is about independence, then ESG is about building a sustainable future.

ESG: Green Cloud as a Competitive Advantage

ESG is about environmental responsibility, people, and fair governance. In the cloud context, ESG means data centers that consume less energy and initiatives that save both resources and budget.

According to Flexera, 36% of companies already track the carbon footprint of their cloud usage, and 57% plan to start this year. This shift is partly voluntary, partly driven by regulation. The EU has introduced the Corporate Sustainability Reporting Directive (CSRD), a new, union-wide reporting standard that requires companies to disclose data on emissions, energy consumption, and climate-related risks. Initially, from 2024, the directive applies to large organizations. Starting in 2026, the same reporting obligations will extend to mid-sized companies as well.

Ignoring ESG is risky. In 2022, Goldman Sachs Asset Management was fined $4 million by the U.S. Securities and Exchange Commission (SEC) for failing to follow its own “green” policies. Investors were offered ESG-branded products that were not backed by real internal processes. The case sent a clear signal: regulators will no longer tolerate “green” claims without substance - sustainability statements must be supported by concrete practice.

In this context, people often talk about “incentives and penalties.” We’ve just looked at the penalties side. Now it’s time to examine the other side of the scale, the incentives.

A recent example from Microsoft shows that green cloud is not just about sustainability, it’s also about profit. Their study published in Nature found that switching to direct chip cooling cuts emissions by 15–21%, reduces energy consumption by 15–20%, and almost halves water usage. For businesses, that means lower operating costs and more room to scale: it becomes easier to launch new data centers in regions where regulators and local communities strictly enforce sustainability standards. Technologies like this set a new industry benchmark and turn ESG from a “burden” into a growth driver.

Companies that keep energy consumption and emissions under control win tenders, attract green investment, and build customer trust. With CSRD making environmental reporting mandatory, ESG is becoming as integral to strategy as finance and security.

The green cloud is setting the tone but the future of the industry will ultimately be shaped by technology.

The Technologies Shaping What Comes Next

Generative AI as a Service: Power Without Capital Expenditure

Everyone remembers the initial hype in 2020, when generative models like GPT-3 first appeared. They could write text, generate code, create images, and, at the time, many saw them more as a playground for experiments than as a serious production tool.

By 2023, the story had changed. The Future of Life Institute published its open letter “Pause Giant AI Experiments”, signed by more than 30,000 experts, including Elon Musk and Yoshua Bengio. The authors called for a six-month pause on developing systems more powerful than GPT-4 to properly assess the risks. That moment became a symbol of how AI had moved beyond research labs and started to shape the global agenda.

By 2025, generative AI stopped being an experiment and became a full-fledged service for businesses. The shift is visible in the numbers: 71% of organizations already use intelligent algorithms in marketing, sales, or product development. More importantly, it’s not just about adoption, it’s about transformation. CEOs are taking AI governance under direct control, processes are being redesigned, and KPIs are explicitly tying technology to profit.

The core advantage is access to massive compute power without massive upfront spend. Building a proprietary large language model can cost hundreds of millions of dollars in servers, data, and engineering talent. Now, companies simply connect to AI via an interface and pay per use.

This removes the traditional barriers. Even mid-sized businesses can work with enterprise-grade AI without blowing up their budgets. At the same time, scale is reshaping the economics of AI: the more users there are, the lower the unit cost of each operation. Industry estimates suggest that in just two years, the cost of generating a single AI response has dropped by roughly a factor of 1,000, making it comparable to a standard web search.

The practical impact is clear: chatbots, marketing automation, and faster product development are all just an integration away but they demand thoughtful architecture and strict quality control.

For CEOs, the takeaway is straightforward: generative AI makes it possible to scale innovation without billion-dollar infrastructure investments. The main risks, from unpredictable costs to compliance requirements, largely sit on the provider’s side, while the company gets ready-made AI power on a subscription basis. But success depends less on the technology itself and more on how intelligently it’s embedded into processes and KPIs.

That’s why generative AI is no longer just a tool; it has become a core element of digital transformation strategy. The next step in that transformation is moving away from traditional servers toward a new computing model: Serverless 2.0.

Serverless 2.0: Pay Only for What Actually Runs

In the cloud stack, the next step after IaaS and PaaS is serverless computing. The servers still exist, but their configuration and maintenance are handled entirely by the provider. Developers simply write a function or connect a service, and the cloud decides where and how to execute the code.

This frees developers from thinking about servers or containers: tasks are triggered by events, and the underlying infrastructure stays hidden on the provider side. So far, serverless has gone through two major stages.

The core impact of Serverless 2.0 is cost efficiency and speed. Companies pay only for the milliseconds when their workloads are actually running, not for idle capacity. After migrating, many see their infrastructure costs drop by 60–80%. Scaling becomes automatic, from a handful of requests to thousands, with no manual tuning. That accelerates product launches and lets businesses focus on ideas rather than hardware.

Serverless 2.0 brings together several building blocks:

• FaaS (Functions as a Service): AWS Lambda, Azure Functions, and Google Cloud Functions simplify development by removing day-to-day server management.
• BaaS (Backend as a Service): Ready-made databases, authentication, and APIs speed up application delivery and shorten time-to-market.
• Edge serverless: Cloudflare Workers and AWS Lambda@Edge deliver ultra-low latency for users across the globe.
• AI/ML integration: Real-time models for chatbots and other intelligent features make applications smarter out of the box.

Serverless 2.0 is changing the game for businesses not only at the technical level, but also strategically:

Challenges remain, for example, vendor lock-in and the difficulty of deep debugging, but the ecosystem is catching up fast.

For businesses, this opens up new scenarios:

• E-commerce platforms save millions during peak days like Black Friday, as carts and storefronts scale automatically with traffic.
• Media services process video in minutes instead of hours, cutting the cost of publishing content.
• Fintech companies scale to millions of transactions without rewriting code or buying additional servers.
  

Taken together, this creates an architecture where products launch faster than the competition and grow without pain. Serverless is becoming the foundation for agentic AI and intelligent data pipelines. But in a world where data drives everything, trust and transparency in how that data is managed demand an entirely new model.

Zero Trust: A New Trust Model for the Cloud

Zero Trust is a security approach where access to data is granted only after strict verification, much like entering a bank vault. In the cloud, where traditional network boundaries have effectively disappeared, this zero-trust model has become the new standard.

In practice, the system evaluates every single request across multiple parameters: who is making it, from which device, from which network, and to which data. Trust is never “remembered”, each new request must earn access again.

Companies are rapidly adopting this model in the cloud: 81% already use Zero Trust partially or fully to protect against data leaks and cyberattacks. Most focus on access management and encryption. However, fragmented tools often create gaps in protection, especially in multi-cloud environments.

EU Cloud Infrastructure You Control

Run production workloads on dedicated resources across EU data centres. Transparent pricing, no hidden costs. Full control over compute, storage, and networking.

Explore Infrastructure

Predictable Virtual Machines for Production

Dedicated vCPUs, consistent performance, no noisy neighbours. Deploy and scale VMs in minutes. Full root access, fixed pricing.

View VM Plans

Up to €22,000 to Migrate Your Workloads

Cut OPEX when migrating to EU infrastructure. Get migration support, cost assessment, and infrastructure credits with no upfront commitment.

Apply for Bonus

Kubernetes with Full Control

Deploy production-ready clusters with control over node configuration, storage, and networking. Predictable costs and direct access to infrastructure.

Deploy Cluster

This is where the weakness of the classic model starts to show. Zero Trust was originally designed for more static environments, where rules could be defined and left in place. In the cloud of the 2020s, that no longer works: serverless functions live for fractions of a second, containers spin up and down by the thousands, and automated, AI-driven attacks demand real-time response.

A new layer has emerged: Adaptive Trust. It’s not a replacement for Zero Trust, but an evolution. The core principle, “never trust, always verify,” remains, but it is augmented with dynamic context.

Here’s how it works in practice: the system evaluates not just a login and password, but also behavioral signals, including geolocation, time of day, device type, and historical activity. Depending on the context, the verification level is automatically tightened or relaxed.

For this approach to deliver real value, it has to translate into concrete actions. For CEOs and CIOs, that means:

• Access management: unified policies and time-bound permissions across all clouds.
• Secrets protection: centralized storage of passwords, keys, and tokens.
• Continuous enforcement: every request and every device is verified.
• Full visibility: real-time audit and analytics.
• Smart response: using AI to prevent and contain incidents.

The impact of Zero Trust and Adaptive Trust goes far beyond cybersecurity.

For FinOps, it means more predictable spending and fewer penalties from security incidents.
For sovereignty, it means knowing that keys and data remain under local jurisdiction.
For ESG, it means fewer breaches and less energy spent on remediation, which directly reduces the organization’s carbon footprint.

Taken together, this turns security from a mere insurance policy into a strategic asset. In the future, Zero Trust and Adaptive Trust will form the foundation for GenAI and Serverless: in a world where agentic AI and serverless functions depend on real-time, context-aware protection, static models simply won’t survive.

And with that, it’s time to move from theory to practice.

Strategic Decisions for CEOs in 2025

Multi- and Hybrid-Cloud Strategies as Risk Protection

Are you prepared to depend on a single provider in a market that demands flexibility? Multi-cloud is like a chess strategy: every move is designed to reduce risk. By 2025, 92% of companies are already combining multiple platforms to avoid outages, lock-in, and regulatory constraints.

Hyperscalers still set the pace. Their ecosystems accelerate innovation and shorten time-to-market. Netflix, for example, runs entirely on AWS, delivering billions of hours of content every month while automatically scaling thousands of server groups to match demand.

But relying on a single player is risky: outages, pricing changes, or new regulatory requirements can become extremely costly. That’s why more and more organizations are choosing multi- and hybrid-cloud models.

Diversification is achieved not only by mixing different hyperscalers but also by partnering with local providers that offer infrastructure under national jurisdiction and more flexible terms. Public cloud brings speed and global reach; private cloud delivers control and compliance. For financial services and healthcare, this balance is critical, and even in e-commerce, hybrid architectures are quickly becoming a matter of survival.

To put this strategy into practice, CEOs should:

• Recalculate risk. Run an audit through a FinOps lens to understand the true cost of relying on a single provider and where the company is exposed. This becomes the foundation for conversations with investors and the board.
• Build pilot scenarios. Launch projects in parallel: with hyperscalers for speed and scale, and with local providers for compliance and trust.
• Create a unified security layer. Implement Zero Trust as the security backbone: consistent access policies and data controls across all clouds. This reduces the likelihood of incidents and regulatory penalties.
• Test resilience under real pressure. Simulate peak scenarios, from Black Friday traffic spikes to sudden user growth, to ensure the architecture holds up. Lab tests alone are not enough.
• Invest in the team. Train architects and DevOps engineers to manage multi-cloud as a strategic asset, not just a patchwork of disconnected services.

Your competitors are already making their move. Launch a multi-provider pilot in the near term to reduce risk and strengthen your market position. Make the first move while others hesitate.

But what if legacy systems are slowing you down? Technical debt can choke even the best-designed strategy. That’s the next big challenge for leadership.

How to Beat Technical Debt

What if your business is slowing down because of old code? Technical debt is like rust: it builds up from rushed decisions, poor documentation, and outdated systems, and gradually slows down digital transformation.

So what is it, in simple terms? Put simply, technical debt appears when a company chooses simpler, faster solutions in code or infrastructure to speed up a product launch. These shortcuts work “here and now,” but later they require rework. It’s like taking out a loan: you borrow time, but pay interest in the form of higher maintenance costs, slower development, and a greater risk of failures.

Common examples of technical debt include:

• An old monolith that everyone is afraid to touch: any change risks breaking the entire system.
• Quick hacks in the code that solve a problem fast, but block future scaling.
• A lack of documentation, forcing new team members to spend weeks just figuring out how the system works.

It’s important to understand: technical debt is inevitable. Research from McKinsey and Carnegie Mellon shows that it exists in any company that actively develops products and updates its systems. Like financial debt, it can be justified: sometimes it’s smarter to ship a feature quickly and “borrow” from the future. But if that debt is left unmanaged, the interest accumulates fast and turns into a blocker for innovation and growth. Companies typically spend up to a quarter of their IT budget just servicing technical debt.

Can you get rid of it completely? No. Technical debt is a natural companion of IT development. The good news is that you can manage it and minimize the damage.

You can’t eliminate technical debt entirely, it’s a natural companion of IT development. The good news is that you can treat it as a managed risk rather than a silent threat.

There are several concrete steps to prepare and minimize the damage:

• Make debt a priority. Put technical debt on the board agenda, alongside financial risk and cybersecurity.
• Understand the scope. Bring in architects and use observability tools to audit systems and uncover hidden dependencies.
• Prioritize remediation. Build a risk map and start with the modules that block growth, for example, legacy monoliths that slow down releases.
• Avoid adding chaos. Choose reliable cloud partners, from OVHcloud and T-Systems to Peerobyte, to support migration, infrastructure modernization, and hybrid-cloud setups that reduce debt while staying compliant with GDPR.
• Embed control into delivery. Integrate debt checks into regular sprints and use AI-powered tools to optimize and refactor code continuously.
  

Handled this way, technical debt stops being a silent killer of innovation and becomes a manageable parameter of your growth strategy.

All of this matters because clean architecture frees up budget for ESG initiatives and speeds up product launches. Start small, even with a simple plan. Technical debt is not a sentence; it’s a manageable process. But who actually executes that plan? Exactly, the team.

People Make the Difference: Investing in Architects and Teams

Think of top experts as skyscraper engineers: only the very best can design a future that doesn’t collapse under its own weight. Cutting corners on talent breeds chaos; investing in strong architects and teams fuels growth, supports multi-cloud strategies, and keeps technical debt under control.

The risk of underestimating this is obvious. When a weak link is built into the architecture, it can translate into billions in losses. McKinsey cites a telling example: a company missed out on $2 billion in potential margin because underpowered teams failed to modernize outdated platforms. They spent $400 million just to keep those systems running, and after two years, only half of the planned changes had been delivered.

The best specialists work one step ahead. They design architectures so that refactoring accelerates innovation instead of turning into constant firefighting. But architects alone are not enough: without strong engineers, analysts, and product managers, modernization stalls.

High-performing teams operate systematically. They don’t just fix code, they build a culture where refactoring is built into sprints, and up to 50% of engineering time can be freed for new products and technologies.

What does this mean for CEOs?

• Find the right leaders. Experienced architects with industry background can run audits and spot risks long before they turn into incidents.
• Give them time. Allocate dedicated sprints for refactoring: it’s an investment in the speed of future releases.
  

Strong teams are not a luxury; they’re a necessity. They can unlock up to half of your engineering capacity for new products. Start looking for the best people today, and with that, we’re nearing the conclusion.

Conclusion: Cloud as a Business Imperative

2025 has become a turning point: the cloud is no longer a supporting technology, it is now the core of business strategy. Companies are no longer debating whether they need the cloud, but how to use it: for cost control, independence, or accelerated innovation.

The key shift is that the cloud has become a platform for growth. FinOps ties spending directly to outcomes. ESG turns data centers into an asset for investors. Sovereign solutions strengthen customer trust and protect businesses from regulatory risk.

From here, the difference is in speed. Some companies use GenAI and Serverless to launch products faster. Others are building security on top of Zero Trust. A third group distributes workloads across multiple providers. The tools may differ, but the goal is the same: to build an architecture that is transparent, flexible, and future-ready.

The path is clear: the cloud is a strategic asset that underpins finance, compliance, and market leadership. Start with a FinOps audit today to build a system where technology and teams work together for growth.

Those who act now will define the market tomorrow. The only question is whether your company will be among them.