FlexSysAI launches platform to ease data centre power strain

FlexSysAI has launched a platform that lets data centres shift GPU-based AI workloads across locations and time to ease pressure on electricity grids during periods of high demand.

The Australian energy technology company links live electricity market signals and grid conditions with workload balancing, allowing operators to move computing tasks to places and times where power is cheaper and more available. It also lets operators reduce non-critical workloads when the grid is under strain and prices rise.

The launch targets a growing problem for data centre operators as electricity demand from artificial intelligence workloads rises and access to power constrains expansion. FlexSysAI says its system is intended to help operators connect to the grid sooner, lower power costs, use renewable electricity when supply is strong and earn revenue by supporting the grid.

Founded by Victor Feoktistov, Angelo Perera and Sean Senvirtne, the business has spent several years developing its demand-response technology and securing a retail licence that gives customers direct access to energy markets. It is backed by early investors EnergyLab and Sean Senvirtne, and is part of Nvidia's Inception programme.

Australia is the company's first market. FlexSysAI points to the structure of the National Electricity Market, the pace of data centre growth and the country's high share of renewable generation as favourable conditions for testing the product. It has a pipeline of multiple Australian data centres exploring adoption.

Policy backdrop

The launch comes as governments and regulators examine how to manage the electricity demands of large data centres. In Australia, the Australian Energy Market Commission recently recommended a new national framework for handling rising data centre demand, including an operational flexibility pathway.

FlexSysAI says its model gives data centre operators a way to respond to those policy pressures without giving up control over when and where workloads are moved. Operators can see a live price for flexibility and decide whether to participate.

Global electricity demand from data centres is expected to climb sharply as artificial intelligence use spreads. FlexSysAI cited projections that annual demand could exceed 1,700 TWh by 2035, with electricity costs reaching USD $170 billion, underscoring the commercial and infrastructure pressures facing the sector.

Physical limits on the grid are already affecting how quickly operators can secure power connections in some markets. That has pushed energy access higher up the list of strategic concerns for the industry, particularly for facilities running power-intensive AI workloads on GPUs.

Victor Feoktistov, Co-founder of FlexSysAI, set out the company's view of the challenge facing operators.

"Data centres need a strategy for the power crunch that is holding back the sector. Physical grid constraints are beginning to bite, and time to power is already a major constraint for operators. These pressures are likely to worsen over the next two to three years, while regulatory pressure is moving even faster as more jurisdictions look to require flexibility from large energy users," Feoktistov said.

He also described the company's case for a voluntary approach to flexibility.

"Flexibility needs to be built into data centre operations today to stay competitive. FlexSysAI can support both voluntary and mandatory approaches, but we believe giving data centres a voluntary path to participate is critical to driving early adoption. Operators stay in control, see a live price for flexibility and choose when to shift workloads, while connecting to the grid sooner and getting ahead of mandatory requirements that continue to emerge across key markets," Feoktistov said.

Technology focus

Chief Technology Officer Angelo Perera said Australia offers a difficult but useful environment in which to prove the model. He argued that the market's volatility and structure reward demand that can respond quickly to price and supply conditions.

"We are starting in Australia's National Electricity Market, as it has one of the world's most challenging grids, and its structure and volatility strongly reward smarter, more flexible demand. This makes Australia an ideal market to prove the technology before deploying it globally, as we provide a long-term solution to the accelerating electricity-demand problem," Perera said.

"We believe this technology has a critical role to play in managing electricity costs and supporting grid reliability, while helping data centre operators connect sooner, pay less for power and get paid to support the grid, while maintaining the support of communities, governments, regulators and utilities around the world," Perera said.