This is an announcement of the start of my research into digital twins for data centers.

A digital twin is a continuously updated virtual model of a physical system, synchronized with real‑time data from sensors, analytics platforms, and AI systems. This live replica mirrors the behavior, performance, and interdependencies of the system, enabling teams to test, validate, and optimize decisions before making real‑world changes.

Data centers are complex systems of resources, infrastructure, and personnel that house the data and computational capability to deliver outcomes for an organization. The increasing importance of data centers along with their increasing need for resources (e.g., power, water) makes their availability critical to maximize their utility. By applying digital twin technology to data centers, organizations will become data-driven data center operators: optimizing performance, reducing risk, and working at machine speed instead of human speed.

The focus of my research will be exploring what digital twin for data center solutions are available, what to expect from them, and what use cases justify the use of this technology and provide the best ROI. The critical aspect of this research is finding use cases and determining the digital twins for data centers that are more effective than other solutions. When beginning my research, I focused on what can be achieved quickly to show value, such as scenario-testing infrastructure decisions (i.e., which vendor to go with, what servers to deploy).

When having discussions with vendors about demonstrating the value of digital twins for data centers, virtually every vendor was on board, but one person challenged me: Dr. Tim Shedd (senior distinguished engineer in the office of the CTIO at Dell Technologies). In his pushback, he pointed out that the use cases I promoted did not provide enough value to the customer — they were essentially “a pretty interface put on top of Excel.” His challenge was to find use cases that justified the use and cost of a digital twin platform; it needed to go beyond what static analysis could achieve but remain below needing a supercomputer. In taking this position, Dr. Shedd is advocating for customers. They should not adopt a new technology just because it is there; it needs to provide a real return on investment. That position is what will inform this research.

Expect research that sets expectations for this technology and service, identifying whom you should speak to and what use cases provide an ROI. If you would like to participate in this research or have a specific request, please reach out to me or my research associate, Melissa Chhay, at mchhay@forrester.com.