AI-driven upgrade of communications energy storage empowers a new energy ecosystem for multiple industries

AI datacenters have become a major driver of rising electricity consumption, but AI doesn't just happen in a datacenter. The telecommunications networks connecting datacenters to businesses and consumers are themselves major electricity consumers. 5G and its successors mean ever higher energy demands. Newer comms architectures deliver more bandwidth and richer services, but they also require denser cells, and 5G infrastructure such as multiple-input multiple-output (MIMO) antennas consumes more power than earlier equipment. The International Energy Agency predicts total datacenter energy consumption will double between 2025 and 2030 to 950 TWh, representing three percent of global electricity demand, while AI datacenter consumption alone will triple over the same period. Telecom industry electricity consumption is rising in parallel. The IEA put telecom network consumption at 260 to 360 TWh annually in 2022, up to 1.5 percent of global electricity use at the time, with mobile comms accounting for two thirds of the total. This all plays out against surging electricity demand in general, as sectors such as transportation and industry shift away from fossil fuels. As the IEA puts it, the age of electricity has arrived. Electricity consumption is now projected to grow at least 2.5 times faster than overall energy demand over the five years to 2030, it says. The relationship runs both ways, the IEA notes. AI may prove critical for global industrial innovation and competitiveness. The agency says that "proven applications of AI could help firms in energy-intensive industries reduce their energy costs by three to 10 percentage points." The agency adds, though, that "The energy sector as a whole is not yet taking full advantage of AI's potential... with lack of sufficient digital skills and data availability emerging as key barriers to adoption." That looks like a significant omission, given that traditional electricity generating infrastructure, in the shape of large central generators and national grids, is struggling to keep up with customer demand, particularly from the tech sector. Those same customers, ironically, already own substantial energy storage and even generating capacity of their own. Technology installations, particularly at the edge or in isolated areas, have always needed backup energy in the form of generators and batteries. These static resources are increasingly supplemented with renewable energy such as photovoltaic (PV) arrays and wind, which are critical, indeed the main option, for installations in remote areas beyond the grid. From idle assets to profitable energy nodes To date, those assets have remained dumb, sitting idle in wait of an emergency and, in an ideal world, never being used at all. Add smarter energy storage and AI, and that infrastructure could become a source of power in its own right, generate revenue for telecom operators, government, and enterprises and tech operators, and help stabilize the grid overall. The starting point is a full-stack energy storage system. As Kong Peng, vice president of ZTE Digital Energy, explains, conventional energy storage systems, have generally been assembled from components sourced across multiple vendors and reliant on heterogeneous comms protocols. This in turn leads to "compounded energy losses, inefficient joint debugging and ambiguous accountability". In contrast, ZTE has developed an integrated full-stack energy storage solution. Its core hardware includes a dual liquid cooling cabinet BESS (battery energy storage system) and a containerized BESS. The 261 kWh cabinet BESS is suitable for deployment at edge data centers and core sites. The containerized BESS can be cascaded to form a large-capacity energy storage system ranging from tens of MW to hundreds of MW and applied to AI datacenter, zero-carbon industrial parks, Grid-side independent energy storage and other scenarios. Around that is "native integration of core components including battery cells, BMS, EMS and PCS." (battery management system; energy management system; power conversion system). It also covers "maturing technologies" such as liquid cooling thermal control and cluster-level management. AI ties the whole thing together through ZTE's intelligent energy management system, ingesting data from the generation and storage infrastructure alongside external signals in real time, such as weather forecasts and energy prices. The VPP system tracks real-time electricity prices while AI algorithms work out the optimal charge and discharge strategies to maximize trading. Beyond creating revenue opportunities, the system stabilizes supply and makes better use of available PV energy. ZTE's complete energy storage system has an overall efficiency of over 90 percent, which can completely replace traditional backup power sources such as diesel generators in some scenarios. The infrastructure also opens up multiple revenue-generating possibilities. One route is "arbitrage via peak-valley electricity price differences, grid demand response and frequency regulation subsidies." Another involves using sites for local PV power absorption "to gain revenue from green energy certificates and CCER carbon trading." Or operators could turn to energy storage asset leasing for predictable long-term cash flow. The result is that operators, or rather the AI, can determine when it makes sense to pull energy from the grid, when to store it, and when to supply it back. ZTE also offers a one-stop service to get customers up and running, Kong says, "covering site survey, grid connection, construction, asset custody and carbon asset development." Adapting to the environment There is no one-size-fits-all approach for how operators can use the system to generate revenue. As Kong points out, Southern Europe is "abundant in photovoltaic resources, prioritizes integrated PV-storage base stations and scales up installations after verifying economic returns." Northern Europe, by contrast, "features volatile power prices and a mature frequency regulation market, where operators focus on revenue from energy storage auxiliary services." The technology is particularly relevant for telecom operators, who, as Kong puts it, have "massive base station resources with untapped load regulation potential, delivering win-win outcomes for both power grids and telecom carriers." This can include peak-valley arbitrage as well as "demand response and frequency regulation auxiliary services, substantially cutting electricity expenses." Using vacant land at sites lets operators generate power while aligning more closely with Europe's renewable energy policies. Datacenter operators also have substantial energy storage infrastructure and are increasingly looking at PV and other renewable, behind-the-meter energy options. Türkiye Telecom has put the technology to work in a 128MWp solar power plant covering 130 hectares at Sivas in central Anatolia. The plant uses N-type PV panels and 350kW inverters, and when complete will produce 196GWh of energy, which amounts to 15 percent of the operator's energy consumption, and cut carbon by 88,000 tons. This marks a crucial step in the cell operator's energy transition, and sets a replicable model for similar projects. Other telcos have adopted the technology. In Italy one provider is building out energy storage systems at its base stations and then accessing peak-valley arbitrage and demand-side response revenues, while also providing third-party storage for domestic and commercial customers. Other European partners are implementing the system in countries including Austria, Romania and Finland, amongst others. Not every site is viable, Kong explains. Regions with flat electricity prices offer little chance of arbitrage revenue, while stable grids reduce the need for backup power. Grid approval mechanisms, broader energy storage regulations, and carbon trading policies will all come into play. Operators must of course ensure they comply with EU regulations around grid stability, telecom infrastructure, and renewable energy obligations. They also need to make sure installations are in line with more mundane rules around construction standards and fire safety. This AI-powered integrated intelligent energy storage system is not limited to the telecommunication world. Its strong compatibility and intelligent capabilities, means it can be deployed in a wide range of industrial and livelihood applications. In industrial parks, the system connects with photovoltaic and wind power equipment. With the same AI algorithms able to optimize energy scheduling, and balance peak and off-peak power demand. It can help cut high electricity expenses, ensure stable power supply for precision equipment, and facilitate the zero-carbon transition of industrial parks. Likewise, in commercial buildings and urban complexes, it adapts to fluctuating building power loads to store energy during off-peak hours for electricity savings. Meanwhile, it can participate in grid demand response to create new revenue streams. Energy storage delivers a clear value in industries which operate in remote locations. Smart farms can adopt PV-energy storage integrated systems to address unstable and costly power supply, ensuring steady operation of agricultural equipment and underpinning the development of smart agriculture. In the mining sector, intelligent energy storage can replace energy-intensive traditional diesel generators. Tailored to the intermittent power consumption of mining operations, it can reduces costs and carbon emissions while securing reliable power for high-risk work. And in areas with weak power supplies such as off-grid villages, islands and remote islets, the integrated PV-Energy storage solution enables the construction of independent, clean microgrids, freeing them from reliance on conventional fossil fuel generation. Continuous advances in energy storage technologies, AI algorithms and supportive new energy policies are paving the way for large-scale adoption of all-scenario intelligent energy storage. Moving forward, the industry will focus on exploring technical adaptation, implementation challenges, profitable business models and compliant operation for diverse application scenarios. Though each operator faces a distinct environment, one thing is clear. We may be in the age of electricity, but electricity, at least from the grid, is not a limitless resource, partly because of the demands of AI. Combined with modern electrical infrastructure and renewables, AI can give telecom and datacenter operators the opportunity to do more than tick the sustainability box while offsetting soaring power prices: they can make their services more resilient, cut their power bills sharply, and help stabilize local and regional power ecosystems.