Educational establishments investing in AI are allocating two-thirds of infrastructure budgets toward data, storage and compute to power AI applications, but are struggling to see a positive return on their investments.

This is according to the fourth annual 2026 Wasabi Global Cloud Storage Index explores how organisations are managing rising infrastructure costs, scaling AI projects, and strengthening data security, highlighting a growing disconnect between AI investment and efficiency in the industry.

Based on global survey data from 1,700 business and IT respondents, including 241 respondents within the education sector, the study has uncovered key findings over the past three years on buying behavior, user satisfaction, and the impact of hyperscaler fee structures on storage costs.

AI Ambitions Hit Roadblocks

It found that AI ambitions in the education secretary are high but infratructure cost and efficiency are major roadblocks.

There is little hesitation in investment: Nearly all (98%) education organisations expect their AI infrastructure budgets to either increase or remain unchanged over the next year, with 46% planning to increase spending.

A large share of that infrastructure spend (67%) will go to data, storage and compute to run AI applications.

Costs associated with AI are not just high, they’re unpredictable, with 50% citing data storage challenges, like the cost of storage and data access, as the number one challenge associated with AI project and solution implementation.

Just 37% of AI projects currently in place are achieving positive returns. However, education IT decision makers’ expectations for future results show this rate increasing to 47% over the just the next 12 months – an indication of sector demand to achieve positive ROI soon.

“Education institutions are eager to dive head-first into AI, but the survey data illustrates a concerning trend regarding expectations vs. fiscal realities,” said Andrew Smith, director of strategy and market intelligence at Wasabi Technologies and a former IDC analyst.

“To ensure long-term success of AI initiatives, IT buyers in education must consider both the technical challenges associated with their data (i.e., storage, migration, quality); as well as the long-term cost-efficiency of accessing, retaining, and securing this data.

“Avoiding costly, budget-breaking fees from hyperscaler infrastructure services should be a priority.”

Cloud Costs

The report found that cloud storage fees consume a growing share of budgets and spending.

The proportion of spending on fee-related charges such as data egress, API operations, and access—compared to actual storage capacity—has risen year over year in the education sector.

r.

Recommended reading

• How Do Teachers Feel About AI in Education?
• Is AI In Education Putting Children’s Development At Risk?
• Is the UK Education System Falling Behind on AI?
• Teachers Not Keen On Using AI For Education, BCS Says

In 2026, these fees accounted for 54% of cloud storage spending among education respondents, up four percentage points (50%) from 2025.

As organisations in education access, analyse, and utilise their data more frequently, it is increasingly difficult to predict fees.

This surge in data activity drives higher storage retrieval, API, and egress fees—making spending harder to control.

As a result, 41% of education respondents reported exceeding their public cloud storage budgets last year, signaling a shift from storage being a predictable infrastructure cost to a more variable and impossible-to-manage expense.

Data Maturity, Confidence Gaps

Education has a cyber resilience confidence gap. Fewer than half of institutions (47%) feel completely confident in their ability to keep data unaltered and operational after an attack, pointing to a broader lack of confidence and potentially trust in recovery capabilities for the majority of education respondents.

With 44% of education respondents reporting loss of access to public cloud data due to a cyberattack, this threat remains a persistent, real-world issue.

Progress is being made, with 63% of education respondents using immutability to protect their data in the public cloud, an increase from 49% last year; pointing to a movement toward implementation of stronger data protection mechanisms.