Tech leaders need a cloud reality check before it’s too late

When Gartner forecast that worldwide sovereign cloud infrastructure spending will reach $80bn in 2026, it was the clearest sign yet that the cloud market is entering a more complex phase for tech leaders.

Growth is being driven not only by demand, but by concerns around control, resilience and risk. For CIOs and CTOs, this shifts cloud planning beyond optimization towards more complex decisions about cost, capacity and placement.

Recent results from providers such as AWS show demand for public cloud remains strong, and this will only increase with interest in AI tools. Capacity is expanding, services are multiplying and investment remains intense. But for enterprise technology leaders, that growth does not remove the need for trade-offs.

As AI-driven workloads become more memory- and compute-intensive, assumptions about elastic capacity and predictable economics are increasingly difficult to sustain, particularly outside the hyperscale platforms, where exposure to cost volatility and provisioning delays is more immediate.

AI demand

That pressure is being driven by the sheer speed at which AI demand is moving from experimentation into production. According to analysis from Omdia, global cloud infrastructure spending reached $102.6bn in the third quarter of 2025, up 25% year-on-year, as enterprises scaled AI workloads across core systems.

At the same time, research from Deloitte highlights that AI is no longer confined to individual applications, but is becoming a foundational layer across the enterprise technology stack.

That shift dramatically increases demand for memory-intensive and compute-heavy workloads, changing the assumptions CIOs can make about cost, scale and availability.

As pricing becomes more volatile and provisioning less predictable, CIOs are encountering friction in programs that were designed to make things easier.

Projects are now being delayed, budgets revisited and, in some cases, legacy infrastructure kept in place longer than planned because alternatives are either unavailable or no longer economically viable.

The issue is not simply higher spend, but a widening gap between technology ambition and what underlying infrastructure can realistically support.

Cloud strategy

For much of the past decade, cloud strategy often assumed a steady migration towards public platforms. As AI-driven workloads place sustained demands on memory and compute resources, that assumption is becoming harder to maintain.

CIOs are increasingly having to distinguish between workloads that genuinely benefit from hyperscale elasticity, those that require tighter control over cost or data locality, and those that need the flexibility to move as conditions change.

In practice, this is driving a more selective approach to cloud adoption, one that balances public cloud, private infrastructure and hybrid models to manage cost, performance and risk.

In practical terms, this means CIOs can no longer treat workload placement as a one-off architectural decision. They need a clear view of which systems are genuinely elastic, which are cost-sensitive, and which are mission-critical.

This requires closer scrutiny of memory and compute requirements, more realistic assumptions about pricing volatility, and contingency planning for delays or shortages. It also means avoiding rigid designs that lock workloads into a single environment.

The organizations coping best are those that build in optionality, the ability to rebalance workloads, defer non-essential demand and protect critical systems when capacity tightens or costs spike.

Hybrid models

For many organizations, hybrid architectures are emerging as the most pragmatic way of managing that complexity. Public cloud continues to make sense for workloads that benefit from rapid scaling, burst capacity or access to managed AI services.

Private infrastructure, meanwhile, offers greater predictability around cost, performance and availability for memory-intensive or business-critical systems. Hybrid models allow CIOs to combine those strengths, placing workloads where they make most sense and retaining the ability to adjust as conditions change.

Done well, this is about creating a coherent operating model that aligns infrastructure choices with business priorities rather than forcing everything into a single platform.

Of course, hybrid on its own is not a cure-all. Private cloud projects are themselves exposed to many of the same pressures shaping the wider market, particularly around memory availability, lead times and cost. Hardware constraints do not disappear simply because workloads move off hyperscale platforms.

The difference is that hybrid models give organizations more control over how those constraints are managed. By spreading demand, sequencing deployments and retaining the ability to shift workloads as conditions change, CIOs gain room to maneuver that a single-platform strategy rarely allows.

The goal is to prevent this constraint from becoming a single point of failure.

Risk management

For CIOs, this makes cloud strategy inseparable from risk management. Decisions about where workloads run increasingly affect financial exposure, operational resilience and regulatory compliance, not just performance metrics.

As a result, cloud planning is moving, or should move closer to the center of enterprise governance, demanding closer alignment between technology leaders, finance teams and boards.

Looking ahead, cloud adoption is entering a new phase. AI will continue to drive demand, while constraints around memory, compute, energy and supply chains are likely to persist. In that environment, cloud strategy becomes something that needs regular reassessment rather than periodic overhaul.

The most effective CIOs and CTOs will be those who plan with uncertainty in mind, test assumptions early and retain the ability to adapt as conditions change. Cloud has become a permanent fixture of modern organizations.

What is changing is the level of attention and governance it now requires to support growth without exposing the business to unnecessary risk.

We've featured the best cloud storage.

This article was produced as part of TechRadarPro's Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro