The study reveals that only 33% of operators believe their electrical infrastructure is fully prepared to support higher-density artificial intelligence loads, while 78% have experienced electrical or power issues in the past year that have contributed to disruptions, delays, downtime, or potentially serious incidents.

The electrical infrastructure, a bottleneck for artificial intelligence

The results of the study indicate that the challenge for the sector is no longer just about ensuring access to sufficient electricity, but also about having facilities capable of transforming that energy into reliable computing capacity.

Among the main indicators collected by Fluke are:

33% of operators have an electrical infrastructure that they consider fully prepared for higher density AI loads.

78% have experienced electrical problems or problems with their power systems during the last year.

29% say that power quality issues have reduced their ability to deploy or operate high-density AI or GPU workloads.

29% indicate that electrical problems are only identified after performance failures, equipment breakdowns, or downtime occur.

Only 45% have real-time visibility of their critical electrical systems.

Parker Burke, president of Fluke Corporation, warns that internal electrical infrastructure is becoming a bottleneck for the growth of artificial intelligence. According to the executive, the industry must pay closer attention to how efficiently data centers use the energy they already have available.

Shortage of electrical technicians delays capacity expansion

The lack of qualified professionals is another obstacle identified by Fluke for the expansion of computing infrastructures. Sixty percent of data center operators say that hiring qualified electrical technicians has become more difficult, while 34% acknowledge that the shortage of skilled personnel has delayed infrastructure upgrades or the addition of new capacity.

This situation highlights the need to combine investments in electrical equipment with the availability of professionals capable of supervising, maintaining, and managing increasingly complex installations.

Power grid limitations are delaying AI projects

Internal problems at the facilities are compounded by difficulties in accessing sufficient power supply capacity. According to the Fluke study, 76% of operators believe that energy availability and grid capacity moderately or severely limit their expansion possibilities. Furthermore, operators indicate an average waiting time of 12 months to obtain additional grid connections or increase the power available for projects related to artificial intelligence and GPU processing. These constraints reinforce the importance of optimizing the use of existing electrical infrastructure.

Electrical monitoring and proactive maintenance

Limited real-time visibility of critical electrical systems can hinder the detection of anomalies before they impact data center operations. Fluke emphasizes the importance of improving electrical monitoring, power quality analysis, and proactive maintenance strategies to maximize available capacity and reduce operational disruptions. Parker Burke notes that while power grid constraints are largely beyond operators' control, they can improve the monitoring, maintenance, and management of their own facilities.

The study highlights that the growth of artificial intelligence data centers will depend not only on access to new sources of electricity supply, but also on the reliability of power infrastructures and the availability of qualified technical personnel to ensure their operation.