FlexSysAI pilots AI workload shifting with ResetData to ease grid pressure

Australian energy technology company FlexSysAI has launched a pilot with sovereign AI infrastructure provider ResetData, CSIRO and The University of Queenslan (UQ) to shift GPU-based AI workloads across locations and time in response to electricity grid signals - without impacting service.

The FlexSysAI platform is now live, managing power and workloads on a cluster of NVIDIA H200 GPUs at ResetData's AI-F1 facility, described by the company as Australia's first publicly available sovereign AI factory.

CSIRO and UQ will independently analyse data from the pilot to assess whether flexible AI workloads can support electricity system reliability.

FlexSysAI's early modelling suggests its platform could dynamically adjust workloads on the H200 cluster by 20–50 per cent within seconds of receiving a grid signal. The platform categorises AI tasks into "Flex Tiers", protecting critical workloads while using elastic training jobs to provide load-shifting capacity.

The company said electricity networks could support additional data centre connections if those loads can be curtailed during peak stress periods - typically just 0.25 to 5 per cent of the year - but that network operators currently lack operational data on whether AI data centres can reliably deliver that flexibility. The pilot is designed to generate that evidence.

At scale, FlexSysAI said its platform could help data centre operators connect to the grid sooner, reduce power costs, access renewable electricity when it is abundant and receive payment for supporting the grid.

“Early testing has been extremely promising, showing our platform can rapidly respond to electricity market signals and shift AI workloads to when and where power is more readily available," Victor Feoktistov, co-founder of FlexSysAI said.

"This reduces strain on the energy system, lowers operator’s energy costs and unlocks additional capacity in existing infrastructure, meaning quicker connections to the grid.”

“It’s great to be working with FlexSysAI, CSIRO and The University of Queensland on this pilot," Marcel Zalloua, co-CEO of ResetData said.

"As Australia’s sovereign AI infrastructure provider, we want to enable Australian innovation by providing the infrastructure to test and prove new ideas that can have a meaningful impact across the industry."

"Data centres are forecast to become some of the largest single points of electricity demand, and unlocking demand flexibility could help support the grid during periods of stress," Dr Sean Lawrence, CSIRO Energy Systems Research Scientist, said.

"This pilot will provide valuable real-world evidence about whether flexible AI workloads can help support electricity system reliability as Australia's future energy system evolves."

“Our role at UQ is to take the operational data this pilot generates and use it to examine how data centre flexibility could be represented in future network planning and regulatory frameworks," Professor Frederik Geth, UQ-Springfield Chair in Energy, said.

"That's a gap that's been missing from the conversation around AI data centre growth, and it's the piece of the puzzle that determines whether this kind of flexibility actually gets recognised and rewarded in how networks are planned.”