AI Workload Strategies 2025
Global IT leaders’ Key Considerations for AI Workload Placement
AI’s impact on workload distribution
Artificial intelligence is reshaping enterprise IT, but how are businesses adapting their IT infrastructure to meet AI’s growing demands? To find out, Telehouse partnered with S&P Global Market Intelligence to survey IT decision makers on a global level, with the findings now revealed in the AI, workload placement, and data movement report.
Inside the report, we explore the evolving landscape of AI workload placement, revealing how short-term priorities stack up against long term ambitions, assessing the opportunities for organisations and the major challenges as AI capabilities rapidly grow.
S&P Global Market Intelligence conducted market research on AI, workload placement, and data movement, with the findings released in March 2025. It polled over 900 decision-makers and planners, ranging from CTOs and CIOs to managers, across Asia, Europe and North America.
Download the S&P Global Market Intelligence Report
Key findings from the report include:
76%
Workloads are widely distributed across venues, with 76% of AI workloads currently hosted in the cloud or in different types of data centres
50%+
Over half have experienced significant network issues, and 39% have had to halt AI projects altogether because of networking issues.
90%+
More than 90% of organizations view access to cloud on-ramps as critical or quite important to AI/ML architecture
Top worries include data centre-to-cloud networking, moving data for training and inferencing, and effectively streaming high volumes of information.
40%+
Over 40% of industry verticals – including media and entertainment, telecoms, electricity and oil/gas, are adopting colocation to host their AI/ML workloads
Complete the form to download the S&P Global Market Intelligence Report
Interested in understanding AI’s impact on infrastructure planning and data movement? This detailed report covers the key considerations for infrastructure planning and data movement from model retraining frequency to networking bottlenecks. Gain insights on how organisations are balancing performance, cost, and compliance as AI workloads mature.