One of the big problems emerging right now is how to power the massive AI data centers that are being built to handle equally massive AI loads because the existing grid isn’t sufficient to cover what is expected to be a significantly higher energy requirement for these new data centers. Cloud companies are looking at everything from huge solar and wind farms, to spinning up new or dormant nuclear energy facilities in order to handle the load. Tech companies don’t currently have the skills needed to do this successfully, so they’ll be looking for partners that can bridge their data center skills into well matched energy generation solutions.
IBM’s acquisition of Prescinto, which specializes in renewable energy management, is likely the beginning of a trend of solutions providers acquiring companies and skills to help clients better architect their data centers and increasingly dedicated energy production facilities for higher efficiency and lower cost of operations.
The Future Data Center Is Energy Independent
One of the issues with AI is that it not only requires a great deal of energy to work at scale, but it will also be increasingly fully active 24/7 particularly as AI implementations become autonomous. This means that the current power grid that depends on the energy load going from factories to homes in the evenings just isn’t designed to provide the massive energy draw of AI facilities whose power needs don’t decline much at night.
Given the energy use is localized, building up the grid to handle this load would not only be excessively expensive due to energy loss over distance, it would also be relatively inefficient. It would be better for loads like that to draw from local energy sources. Even before this current AI wave, we had huge data centers from companies like Meta going in places where there were already inexpensive energy sources like hydro-electric power.
But as data requirements increased and climate change has made hydroelectric power less reliable, the more recent move is to use nuclear power. Amazon is developing its own, Microsoft is using Three Mile Island as that facility is brought back online, Google is hooking up with Kairos Power, and Westinghouse is launching a line of micro-reactors to power data centers.
Satisfying the Growing Need for Energy Generation Management
But while several of these efforts connect to firms already using nuclear power at scale, the Westinghouse solution, and others like it, will require the customer to either develop nuclear power expertise or use a vendor that has both the knowledge of what is needed and how to provide it.
Interestingly, IBM already has been helping manage water, natural gas, oil and nuclear assets for its customers, and Prescinto provides expertise in the renewable areas where IBM is currently light. This addition now allows IBM to manage a far wider variety of energy generation sources, making it one of, if not the preferred choice as a vendor for data centers that need large mixed sustainable energy sources in order to operate.
Wrapping Up:
The need for resources that can help make data centers of the future self sufficient is growing rapidly as data centers needs grow to exceed the capabilities of the existing power grid. However, with some exceptions, IT knows little about energy generation and needs substantial help not only in building and managing sustainable energy sources, but in choosing which ones make the most sense to begin with.
IBM’s acquisition of Prescinto adds to its already substantial expertise in energy management and potentially turns it into one of the most comprehensive data center solutions providers in the world. This shouldn’t be a surprise given IBM’s leadership in AI with watsonx and the IBM Mainframe. IBM has always had the ability to understand the coming needs of future data centers, from cooling to power availability. This acquisition just further assures that leadership position.