Cloud storage fees are increasingly eating into IT budgets and slowing down infrastructure plans. According to Wasabi’s 2026 Cloud Storage Index, 48% of overall storage costs now go toward fees rather than actual capacity. These fees include things like data egress, API calls, and other usage-based charges.

This fee-heavy model is creating real budget pressure:

• 46% of UK organizations exceeded their allocated cloud storage budgets in 2025.
• 84% of UK respondents cited at least one fee-related reason for going over budget.

For many IT teams, this means they’re forced to slow or rethink infrastructure expansion plans, even as data volumes and AI initiatives continue to grow. Instead of simply paying for the capacity they need, they’re spending almost half of their storage budget on fees that are hard to predict and difficult to explain to finance leaders.

Storage has become a core dependency for AI projects, not just a background utility. AI models rely on large volumes of high-quality, easily accessible data. When storage is expensive or complex, it directly affects AI performance and ROI.

The research highlights a few key points:

• Only one-quarter of respondents say they’ve achieved a positive ROI from AI so far.
• Yet nearly 48% are confident they can reach positive ROI if they improve their infrastructure.
• The top reported challenge in implementing AI solutions is data storage issues, including cost, data access, and management.

In other words, organizations are finding that AI success is less about adding more software and more about getting the underlying data and storage layer right. Cost-efficient, reliable storage that keeps data readily available for AI models can help reimagine how teams design, train, and deploy AI, without letting infrastructure costs erode the business case.

The study shows a noticeable shift in how organizations fund AI initiatives. Instead of concentrating spend on AI software or SaaS tools, more budget is moving toward the underlying infrastructure that makes AI work at scale.

Key trends include:

• Nearly two-thirds (62%) of budgets are now being allocated specifically to data, storage, and compute capacity build-outs.
• Only 36% of spending is going to AI-related software or SaaS solutions.

This is almost the opposite of the broader public cloud market, where most revenue typically comes from SaaS rather than infrastructure services (IaaS). For AI, however, organizations are recognizing that robust, predictable, and cost-aware storage and compute are essential. They’re reshaping their budget priorities to focus on infrastructure as the enabler of AI innovation, rather than assuming software alone will deliver the outcomes they need.