The rise of artificial intelligence (AI) and high-performance computing (HPC) technologies has presented the chemical process industries (CPI) with opportunities for improving process efficiency, asset maintenance, product development and plant safety. While not fully realized in the CPI, the potential benefits of AI and HPC tools are considerable. But for as much as the CPI could benefit from these tools, the CPI are also an enabler of the computing industry. Various parts of the CPI provide critical inputs to the computing infrastructure that is allowing AI to become the ubiquitous force that is has become. This page in CE has previously discussed the key chemical inputs required for semiconductor manufacturing (Chem. Eng., May 2025, p. 4), but the demand for AI technologies has also accelerated the construction of data centers to house the IT infrastructure needed for modern digital operations. The construction and operation of those data centers also require critical inputs from the CPI.
Data center construction. Building the physical facilities that underpin AI computations requires a host of materials that are designed for the data center environment, and that provide durability and functionality to the buildings. The wire- and cable-heavy facilities require the processing and refinement of conductive metals for the wire and polymers for insulating the wiring and cabling, for example. Sealants are used for protecting building equipment from moisture and other environmental factors. Specialized coatings for concrete and steel that can resist corrosion and fire are also supplied by the CPI. Concrete admixtures, such as those with hydrophobic compounds, can create waterproof concrete foundations for areas where liquid cooling systems are needed.
Cooling systems. After construction, data-center challenges shift to operations, where the CPI also play a critical role. By a large measure, AI technologies work because of computational heft, and that computation generates a large amount of heat. One of the key challenges of current data centers is the need to efficiently dissipate and manage the heat generated by the computational processing. Many liquid-cooling systems for data centers are based on ethylene glycol and propylene glycol, and are often mixed with deionized water and a cocktail of corrosion-inhibitor additives. Another emerging cooling method involves immersing computer servers in tanks of dielectric fluids, which can conduct heat, but not electricity. Mineral oils and synthetic hydrocarbons and esters are examples of the classes of compounds that can be used. Fluorine-containing chemicals could be used in two-phase (gas and liquid) cooling systems.
Energy and water use. Currently, the expansion of data centers is contending with the burden their energy needs place on the power grid. A 2024 study from Lawrence Berkeley National Laboratory [1] found that data center electricity use nearly tripled between 2016 and 2023, and by 2028, the report’s authors found data centers could account for as much as 6.7 to 12% of total U.S. electricity consumption. The CPI will play a role both in improving the functionality of the energy system, as well as in raising the efficiency of data-center operations and of the computing itself. Among the areas of interest are the following: improved cooling systems; advanced lower-energy microchips; microcapacitors for on-chip energy storage; and the development of optoelectronics materials (using photons instead of electrons). ■
Scott Jenkins, senior editor
1. Shehabi, A. and others, 2024 United States Data Center Energy Usage Report, Lawrence Berkeley National Laboratory, 2024, LBNL-2001637.