Post-cloud strategy: the cloud bill shock

For more than a decade, the enterprise technology conversation revolved around a single assumption: move to the cloud. Amazon Web Services, Microsoft Azure, and Google Cloud promised scalability, agility, and reduced infrastructure complexity.

Organizations embraced cloud-first strategies at speed, accelerated further by remote work demands, digital transformation initiatives, and the rise of AI-powered services.

Now, however, the mood is shifting.

Enterprises are not abandoning hyperscale cloud platforms, but they are becoming more selective about how and where they use them. Research from Gartner, suggests that by 2027, 90% of organizations will adopt hybrid cloud infrastructure. Cloud repatriation is growing in popularity.

Latest Videos From

Escalating operational costs, mounting concerns around digital sovereignty, and growing complexity in cloud environments are forcing IT leaders to ask a more nuanced question: what workloads actually belong in the cloud?

This article is the first in a three-part series examining the rise of post-cloud strategies. In part two, we'll explore the geopolitical and regulatory forces driving sovereign cloud adoption, while part three will look at how enterprises are designing hybrid architectures that balance cost, resilience, and control.

The hidden economics of public cloud

Cloud computing was originally sold as an economically more efficient model. Rather than investing heavily in on-premises infrastructure, businesses could pay only for what they used. That flexibility remains valuable, especially for rapidly scaling services or unpredictable workloads. The problem is that hyperscale pricing models have evolved into highly complex ecosystems that can be difficult to predict and manage.

Jay Litkey, SVP of cloud and FinOps at Flexera, says enterprises are increasingly being caught out not by headline compute costs or rising storage fees, but by secondary charges that quietly accumulate over time.

“Hyperscaler pricing has become much more granular over the last few years,” he explains. “The surprises are usually not the obvious things, such as compute or storage. It is often the layers underneath that catch people out, such as data transfer between regions, storage sitting in the wrong tier, duplicate environments, or commitment-based discounts that looked smart when they were signed but are no longer aligned to actual usage.”

Those hidden costs are becoming especially problematic for organizations running data-intensive workloads. Analytics platforms and AI systems generate enormous amounts of data movement, which can trigger substantial egress and replication charges.

Indeed, AI adoption can complicate costs massively. AI inference workloads and large-scale model training are driving fresh waves of unpredictable spending. James Brooks, lead for hybrid solutions at Hewlett Packard Enterprise, emphasizes that AI costs are now emerging as a major contributor to cloud bill escalation.

“The most common source of ‘cloud bill shock’ is rarely the primary service,” Brooks says. “Secondary and consumption-based elements such as data egress charges, long-term storage costs, AI inference and token usage, and over-provisioned or idle resources, are where costs quietly compound.”

For many enterprises, the result is a growing mismatch between expected cloud savings and operational reality.

Governance failures are amplifying cloud sprawl

While hyperscale economics are under scrutiny, many experts argue the biggest issue is not the cloud itself but how organizations govern it.

The ease of provisioning resources has created a culture of rapid consumption. Litkey says the core problem is that cloud adoption often outpaces the maturity of governance. “Cloud makes provisioning incredibly easy,” he explains. “Teams can spin things up in seconds and, without the right controls in place, that flexibility can create sprawl very quickly.”

This challenge is particularly acute for enterprises that migrated workloads quickly during earlier cloud transformation initiatives. In many cases, organizations simply lifted and shifted legacy infrastructure into hyperscale environments without redesigning architectures for cloud-native efficiency.

Justin Sharrocks, general manager, EMEA at Trusted Tech, says governance immaturity remains one of the biggest contributors to overspending. “Weak workload design, unclear ownership, and low FinOps maturity tend to drive inefficiency,” Sharrocks says. “Cloud pricing hasn’t necessarily increased; it’s become easier to mismanage.”

The challenge for IT leaders is that cloud costs are often distributed across multiple departments and teams. Engineering prioritizes speed and innovation, finance focuses on predictability, and operations teams concentrate on performance and resilience. Without shared accountability, visibility becomes fragmented.

James Peet, Practice, director for cloud and digital transformation at Ensono says organizations are increasingly integrating financial controls directly into architecture decisions. “FinOps needs to shift from looking backwards at what’s already been spent to shaping decisions before that spend happens,” he explains. “The organisations getting this right are making cost part of how systems are designed and built.”

Sovereignty concerns are changing workload placement decisions

Cost pressures alone are not driving post-cloud strategies. Regulatory scrutiny and current geopolitical uncertainty are increasingly influencing where organizations place workloads.

What was once considered a compliance issue has become a board-level strategic concern. Peet says organizations are paying much closer attention to who controls their data and which legal frameworks apply to it. “Data sovereignty has moved into day-to-day decision-making for leadership teams,” he says. “Organizations are looking more closely at who can access their data, and which legal frameworks ultimately apply.”

What was once considered a compliance issue has become a board-level strategic concern. Peet says organizations are paying much closer attention to who controls their data and which legal frameworks apply to it. “Data sovereignty has moved into day-to-day decision-making for leadership teams,” he says. “Organizations are looking more closely at who can access their data, and which legal frameworks ultimately apply.”

This shift is particularly visible in regulated industries such as financial services, healthcare, government, and critical infrastructure. Organizations operating in these sectors increasingly want guarantees around data residency, operational resilience, and jurisdictional control.

Mark Duff, VP international regions at Mitel, says concerns around sovereignty are now influencing broader infrastructure strategy. “Security, compliance and digital sovereignty are becoming deciding factors in technology decisions,” Duff says. “Data sovereignty, once a regulatory concern, is now becoming a strategic, board-level priority.”

These concerns are accelerating interest in sovereign cloud environments, regional cloud providers, and localized infrastructure models. Some enterprises are also reconsidering private infrastructure for workloads involving highly sensitive data.

James Brooks says geopolitical instability and regulatory complexity are fundamentally reshaping infrastructure decisions. “Data sovereignty and geopolitical risk have moved from secondary concerns to primary drivers of infrastructure strategy,” he explains. “Organizations are reassessing workload placement in response to regional instability, increased geopolitical tension, and growing operational resilience requirements.”

Importantly, most organizations are not moving entirely away from hyperscalers. Instead, they are becoming more selective about where workloads reside and how data is processed.

Hybrid infrastructure is becoming the default enterprise model

As enterprises reassess their dependence on hyperscale, hybrid and multi-cloud strategies are emerging as the dominant operational model.

Rather than centralizing everything in public cloud environments, organizations are increasingly distributing workloads across hyperscalers, private infrastructure, sovereign cloud platforms, and edge environments based on cost, latency, compliance, and performance requirements.

This transition reflects a broader recognition that not all workloads behave the same way.

Krystal Mattich, VP infrastructure and trust at Brain Corp, says the rise of physical AI is accelerating the need for distributed infrastructure. “The cloud-first era was built for digital workflows,” Mattich tells ITPro. “But physical AI introduces a different set of constraints: continuous sensor data, real-time decision-making, unreliable connectivity, and strict requirements around latency, safety, and data control.”

Edge computing is becoming increasingly important in industries that require processing large volumes of real-time data locally. Manufacturing, logistics, healthcare, and autonomous systems increasingly require infrastructure capable of operating independently from centralized hyperscale platforms.

Mattich emphasizes that enterprises are beginning to rethink the role of the cloud itself. “The cloud remains unmatched for scale, but the next phase of infrastructure involves using it more deliberately, not by default,” she says. “The key question shifts from ‘Can we centralize this?’ to ‘Where should this intelligence live to be most effective?’”

That philosophy increasingly defines the emerging post-cloud era. The future enterprise stack is unlikely to belong entirely to hyperscalers, private data centers, or sovereign cloud providers alone. Instead, it will consist of carefully orchestrated hybrid environments where workloads move dynamically based on operational requirements and business value.

Litkey says the era of default cloud adoption is already fading.“I do not think ‘post-cloud’ means moving away from cloud,” he says. “I think it means moving away from the default cloud.”

That shift may ultimately define the next decade of enterprise infrastructure strategy.

That shift may ultimately define the next decade of enterprise infrastructure strategy.

This article is Part 1 of the series Post-Cloud Strategy: What Comes After Hyperscale? In Part 2, we examine how sovereignty, security, and geopolitical pressure are accelerating the rise of regional and sovereign cloud environments.