13 May, 2021
Battle for AI batteries, renewables heats up as Fluence inks 1.5 GW in orders
| Fluence Energy will steer operations of Pacific Gas and Electric's 730-MWh energy storage station in Moss Landing, Calif., with an AI-driven optimization and trading platform. Source: EKM Metering Inc. |
Fluence Energy LLC, an energy storage technology joint venture of U.S. power producer AES Corp. and German engineering company Siemens AG, is accelerating its rollout of an artificial intelligence-based energy trading platform designed to make more money for operators of renewable energy and battery storage assets than humans can achieve manually in increasingly complex energy markets.
Building on its October 2020 acquisition of San Francisco-based Advanced Microgrid Solutions Inc., which developed an automated software platform as an alternative to conventional trading and bidding of variable renewable energy and energy storage resources, Fluence has secured 1.5 GW of new contracts in Australia, the company announced May 12.
Fluence now has 3.5 GW of trading platform bookings at wind, solar and storage assets in California and Australia, including 18% of all large-scale wind and solar capacity in the Australian National Electricity Market, according to the company.
"It's not possible for human traders to realize the sort of optimal levels of revenue and value for the grid without software like this," Matt Penfold, a Fluence vice president, said in an interview.
Other AI vendors are also making that case. Canadian Solar Inc., for instance, on May 13 announced a partnership with U.K.-based Habitat Energy Ltd., an expanding developer of AI-enabled energy storage optimization and trading. Canadian Solar plans to use Habitat's technology for battery storage and solar projects globally.
"The power grid of tomorrow is not only built by renewable energy and battery storage, but crucially driven by artificial intelligence, big data, machine learning, cloud computing and many more technologies," Canadian Solar CEO Shawn Qu said in a statement.
'Human-centric approach is inadequate'
Fluence sees enormous potential to expand its customer base in the United States, including in California and Texas, both of which are adding large volumes of renewables and battery storage but have struggled to keep the lights on during extreme summer and winter weather events over the past year.
"This technology is central to the role storage can play in stabilizing the grid and avoiding these types of blackouts," Penfold said. "What our technology allows the battery to do is ensure that it's discharging at the most critical time for the grid." Fluence does that by using machine learning to predict price spikes "with far more accuracy than any other forecasts available in the market and certainly far more accuracy than any human trader could do just sort of using their own intuition," he added.
"In the next year or two, I think people will realize that the human-centric approach is inadequate and that software like this will really be fundamental to the economic operation of batteries," Penfold said.