The rise of artificial intelligence has created an energy paradox. Technology leaders behind AI tools like ChatGpt say that large-scale language models can solve some of the world's biggest problems, but the infrastructure that powers technology could potentially cause another problem as a result of its environmental impact. According to Mark Chung, CEO of energy efficiency monitoring firm Verdigris, AI data centers can consume 20 to 30 times more energy than their CPU-based predecessors. Some experts predict that AI will account for more than 10% of US electricity consumption within five years, driving the fear that unchecked AI computational demand could exponentially accelerate climate damage.
However, AI and energy convergence also forces a rethink of the industry's traditional practices and creates opportunities to mitigate environmental impacts by creating grids, which will supply and enable data centers that operate cleaner and more efficiently.
“One of the biggest challenges of delivering energy to data centers is optimizing that energy flow. This is an issue that AI can really help solve,” said Katie Durham, partner at Climate Capital.
One of the biggest players using AI to tackle this efficiency issue is Kraken Technologies. The AI-powered operating system serves more than 70 million customer accounts across 40 utilities around the world. It connects over 500,000 consumer devices, from EV chargers to household batteries, controls flexible energy supply over 5 gigawatts, offsetting 14 million tons of CO in 2024 alone. luck.
Devrim Celal, chief marketing and flexibility officer at Kraken, said the company's success depends on finding the efficiency of renewable energy demand. “When you move to renewable energy, you'll find a whole new problem,” he says, explaining the company's role in analyzing the demand for renewable energy in order to create systems that store or deploy energy based on user-specific consumption patterns.
He also notes that the company clusters consumers based on energy consumption patterns and uses machine learning to efficiently distribute renewable power with 90% accuracy. This means that if a customer typically charges an electric vehicle at 100% from 9pm to 7am every day, energy will be deployed at this point and booked when the vehicle is away from home. “When you balance the grid, it's very powerful,” he says.
Miami-based Exowatt builds a solar energy system designed to be equipped with AI data centers 24 hours a day. According to Exowatt CEO and co-founder Hannan Happi, the company helps utilities deal with the inherent intermittent solar intermittent without relying on carbon-generating energy sources by providing a means to store and dispatch solar power at any time. “We're in a really crazy hurry to bring our products to the market and expand as quickly as possible,” he says. “If not, the energy and power solutions data center customers can use is to put diesel and natural gas on the grid. This really impacts the community where these data centers are being built.”
Exowatt also leaps heavily towards AI internally. Use LLMS to power “digital twin” systems that simulate real-time performance and enable proactive maintenance. The company is replacing traditional SaaS tools with custom built AI software tailored to supply chains and manufacturing needs.
Halcyon is a $10.8 million seed funding startup that uses AI to help energy experts in different ways. The company has ingested regulatory applications from agencies such as the Federal Energy Regulatory Commission and the Department of Energy to create a large, searchable, structured language model.
“We use LLM primarily to read,” says Sam Steyer, head of data science at Halcyon. “We've been thinking about regulatory analysts at energy companies who have had to search for the right 1,000 pages of PDFs in the past, use Control F, and find the right data. We're trying to make that process as efficient and fast as possible, and make sure that person does the same work on a much larger scale.”
Part of Halcyon's mission is to ensure that AI's widening appetite for electricity accelerates the transition of clean energy. The company is building trackers for special data center power bills and tools to help renewable developer site projects faster.
“AI and energy are truly symbiotic,” says Steyer. “AI is driving the growth of electricity demand significantly. It is completely essential to expanding our electricity system.”
