ThinkLabs secures $28-million USD Series A to help power grids manage data centre demand

ThinkLabs has secured a $28-million USD ($39-million CAD) Series A round to help utility providers modernize power grid infrastructure as a wave of electricity-hungry AI data centres comes online.

The Canadian-led, New York-based company provides a “physics-informed AI” software that helps utility control room operators and planners maintain grid reliability and reduce bottlenecks and outages.

The annual power demand of data centres in the United States is expected to exceed the entire power output of Texas by 2028, according to S&P Global. In Canada, the problem is also getting attention; Alberta last year added its utilities minister to the team behind its $100-billion AI data centre push.

“We have been crawling for the past decades, and so suddenly now [that] we’re at the end of the road we have to run.”

ThinkLabs’ AI-powered platform maps out and simulates a power grid, and then puts the “digital twin” through the proverbial ringer to learn how the real grid can allocate power or be improved. This can help prepare for extreme scenarios, like what to do if the power grid gets overloaded, or how to budget power for big draws like a data centre.

“We’re turning a quarter-million-dollar study that will typically take six months to perform into less than a day of analysis and just minutes to hours of GPU costs,” ThinkLabs founder and CEO Josh Wong told BetaKit in an interview.

A power systems engineer by training, Wong previously led the smart-grid department at Toronto Hydro and, in 2012, founded smart-grid software startup Opus One Solutions. Wong eventually sold Opus One to General Electric (GE) in 2022. The multinational’s energy business, GE Vernova, went on to incubate and then spin out Wong’s newest idea, ThinkLabs, with $6.8 million CAD ($5 million USD) in seed funding in early 2024.

While ThinkLabs is incorporated in Delaware and has its headquarters in New York, the company has established a Canadian subsidiary. About one-quarter of ThinkLabs’ 20 employees, including Wong, are based in Toronto, he said. Wong added that he’s looking to bring his headcount up to around 50 people before raising a Series B round.

ThinkLabs’ all-equity Series A round was led by Energy Impact Partners, with participation from NVentures and Edison International. NVentures is AI chip giant Nvidia’s venture fund, and Edison is one of the largest electric utility businesses in the US.

Returning investors included GE Vernova, Powerhouse Ventures, Active Impact Investments, Blackhorn Ventures, and Amplify Capital, as well as an undisclosed North American investor-owned utility.

Wong argued that the participation in ThinkLabs’ Series A shows that the company is from the industry, backed by the industry, and for the industry.

“The pain points are just extremely high right now, just given the amount of electrification, data centres, ageing assets, [and] adverse weather events,” Wong said. “But these pain points are really forcing the utility to have no choice but to modernize faster than ever before.”

“We have been crawling for the past decades, and so suddenly now [that] we’re at the end of the road we have to run,” Wong added.

While ThinkLabs has made strides beyond proof of concept since its inception, Wong said the company will use the funding to mature its product for general availability. The funding will also help build out new features, which Wong said is another side effect of the AI race.

“Every week, we are getting demand for new use cases from our customers, and so we have a lot to do in terms of training models, and [applying] these AI models to new use cases,” Wong said.

He added that ThinkLabs is also getting lots of interest internationally, so some of this funding will help prepare for entering new markets such as Asia-Pacific and Europe. For now, however, ThinkLabs will “discipline” itself to focus on North America due to its familiarity with regulations and standards.

“We can apply these AI models universally, but we do need to tweak how [they apply], especially the standards, regulations, and workflows,” Wong said.

Following ThinkLabs’ seed round, Wong told BetaKit that he expected the next 10 years to be critical for the energy transition. Less than two years later, that timeline has tightened significantly.

“Since we are an AI race, including AI for grid, I will say the next two years is enough to dictate and define the next 50 years of the grid, and I’m seeing that right now,” Wong said. “The developments that we have … are just infinitely faster than the pace of development and change that I’ve seen in my last company, and even in the incumbents.”

Feature image courtesy Unsplash. Photo be Andrey Metelev .