# Survey: Enterprise IT Shifts Off Premises as AI Reshapes Data Centers

> Source: <https://techstrong.it/featured/survey-enterprise-it-shifts-off-premises-as-ai-reshapes-data-centers/>
> Published: 2026-07-30 16:24:50+00:00

Third-party data centers now host a larger share of enterprise IT workloads than corporate-owned facilities for the first time, a major milestone in how companies deploy infrastructure as AI drives new demands on capacity and power.

According to Uptime Institute’s [2026 Global Data Center Survey](https://uptimeinstitute.com/uptime_assets/77a6191bef6c85cb4ac9bfb169ddc313cae517030b48709697b1f80bc10edc90-uptime-institute-global-data-center-survey-results-2026.pdf), 46% of enterprise workloads now run in third-party environments such as colocation facilities, cloud platforms, hosting providers and SaaS services.

Enterprise-owned data centers account for 44% of workloads, while the remaining 10% operate from IT rooms or server cabinets. Uptime projects third-party facilities will expand their share to 48% by 2028, while enterprise-owned data centers remain a core element for hybrid IT strategies.

The survey, conducted this spring, gathered responses from more than 1,600 participants worldwide, including more than 800 data center owners and operators.

**Capacity Planning Concerns**

Management concern over forecasting future data center capacity needs climbed to 76% of respondents, up from 71% in 2025.

Although cost remains the leading management concern for a fourth consecutive year, executives are now balancing financial pressures with new operational challenges created by AI infrastructure. Among the industry’s leading concerns: power availability and supply-chain disruptions.

Higher-density computing is reshaping facility design. Average modal rack density exceeded 11 kilowatts for the first time, crossing what Uptime researchers described as a notable industry milestone.

Much of that increase is due to a growing number of AI-focused deployments operating above 30 kW per rack. Excluding those higher-density installations, average rack density rose more modestly from 7.5 kW in 2025 to 7.8 kW this year.

Nearly one-quarter of surveyed operators now manage at least some racks exceeding 30 kW, compared with 19% a year earlier. A smaller but growing number of facilities are deploying GPU infrastructure that exceeds 100 kW per rack.

**Caution About AI**

The survey indicates that data center operators are cautious about allowing AI to control facility operations. Fifty-two percent believe AI improves facility efficiency, down from 58% last year. Confidence that AI reduces human error fell to 40%, an 11-point decline from 2025.

Facility operators showed the greatest confidence in AI for predictive maintenance and operational analytics, and much lower trust in AI for autonomous control of equipment settings or automated configuration changes.

Reliability continues to improve: for the sixth consecutive year, fewer operators reported experiencing a significant outage during the previous three years. Forty-seven percent reported an outage, down from 50% last year.

But the costs of those outages continue to grow. Seventy-one percent of organizations experiencing an outage said their most costly incident exceeded $100,000, compared with 57% in 2025. Uptime attributes the higher costs to growing enterprise dependence on digital infrastructure, while warning that grid instability and power constraints (sometimes due to extreme weather) keep boosting those risks.

Staffing has long been a challenge for data centers and it remains one. More than half of respondents reported difficulty finding qualified candidates, up from 46% last year. Electrical specialists and junior operations personnel remain the hardest positions to fill as demand grows for the expertise needed for today’s complex AI infrastructure.

Environmental reporting also regained momentum after slowing last year. More than half of operators now track water consumption, indicating sustainability initiatives continue even as companies prioritize expanding AI capacity.
