A single question changed how Singularity viewed its market

The startup says most companies’ carbon footprint math is bogus

When Wenbo Shi started Singularity Energy, a carbon intelligence platform that today raised a $4.5 million seed round, he never thought he would focus the company on a greenhouse gas. But one conversation with a customer changed the way he viewed his product and, ultimately, his company and the type of customers it now serves.

“The journey was really customer driven, to be honest. When I started the company three years ago, I wasn’t thinking of carbon at all,” Shi said. “The first idea that I had for Singularity was that we’d do intelligent control for batteries, for EV charging, for those types of things. The objective for battery control is always going to be, ‘How can I save money for the customers?'”

A few years ago, Shi and Singularity had that goal in mind when working with the Harvard Innovation Lab, which houses entrepreneurial resources for Harvard Business School students. The university was looking to pair a battery with solar panels on the building’s roof.

“During one of the conversations, they brought up carbon. ‘Can you actually consider carbon as a signal?’” Shi recalls them asking. The university wanted to install a battery not just to save money, but to lower the campus’s carbon footprint.

“I had never thought of carbon because I was like, ‘Oh, I’m a power system guy,’” Shi recalled. But after the conversation with Harvard, “then I was like, ‘Oh, that’s a very neat idea.’ If I know how clean or how dirty the power grid is, then to me it’s another control signal. It’s an optimization objective, which should be pretty straightforward to integrate with the software.”

It turned out that incorporating carbon as a control signal changed the math for Harvard’s battery project. Shi had discovered that optimizing for cost alone would increase pollution, a revelation that occurred after he started analyzing the grid’s carbon emissions on an hourly basis as opposed to the more commonly used annual averages.

What’s more, when Shi shared his company’s data with some scientists at the University of California, Davis, the researchers found that annual averages were probably wrong, potentially by a great amount. That’s because the energy sources powering the grid are constantly changing. Wind might be strongest at night in some regions, while in others, a cloudy afternoon could reduce the output of solar panels, driving up demand for electricity from a gas power plant. At the same time, the spread of solar panels has helped reduce the overall need for fossil fuel power during peak summertime hours, which used to be some of the dirtiest.

“If you look at the real-time generation fuel mix, the grid actually changes very rapidly,” Shi said. “If you look at the carbon intensity of the electricity, it’s not necessarily true that peak hours are dirty hours. You need to look at the real data because a lot of the assumptions or heuristics about power grids are no longer true.”

In a paper published last month, the UC Davis researchers showed that annual averages might overestimate the carbon intensity of the grid by 33% in some cases while underestimating it by 22% in others. Most of the time, annual averages are within ±5% of hourly estimates, the researchers found. That may not seem like much, but it can be low-hanging fruit for companies looking to cut emissions.

Few companies are under as much pressure to cut emissions as utilities. Several states have laws or directives for the electrical grid to reach net-zero by 2050, 2040, or, in the case of Rhode Island, 2030. Many of those targets were put in place in the last few years, while Shi was rethinking Singularity’s business model.

“When I founded the company, when I pivoted the company to carbon, I was thinking, ‘Oh, this is a very fascinating area, because technically, it’s a very interesting problem,’” Shi said. “But at the same time, we struggled with monetization. Who cares about this? Who is going to pay money for this? Initially, our hypothesis was that it’s going to be the large companies.” Utilities were on his list, but up until a year ago, they weren’t showing any interest in the startup’s product.

That changed when Eversource, a Fortune 500 utility serving New England, reached out to Singularity. They wanted to figure out the carbon intensity of their line losses, or the amount of electricity that is lost along the transmission and distribution system. Nationally, about 5% of electricity is lost in this way, and depending on where those losses occur, they can contribute different amounts of carbon emissions. A transmission line from a gas-fired power plant, for example, will produce a certain amount of pollution, while one from a solar installation will produce none.

“Eversource has one of the most aggressive net-zero carbon objectives in the industry. We have an objective to be net-zero by 2030,” said Vandan Divatia, vice president of transmission policy, interconnections and compliance at Eversource Energy. “A large part of the carbon footprint for Eversource is based on the carbon footprint of the losses on the grid.”

“Today’s methodology includes very gross approximations of the entire regions — emission factors based on whatever is being generated throughout the region and whatever is being consumed throughout the region,” he added. “Singularity allows us to use the much better location-based calculation. So if a particular part of the region has more solar or more nuclear, or more wind, then it should have a better carbon footprint.”

Divatia said that Eversource is going to begin using Singularity’s data on a state-by-state basis. “But ultimately, we could go down to an individual customer level if desired.”

That would bring Shi and Singularity back full circle to the customers he initially envisioned the company serving.

“There is no way for you to achieve or even try to achieve grid decarbonization without collaboration from grid operators or utilities, because they own the system. They have the best view of kilowatt hours flowing through their system,” Shi said. “What we’re trying to help them with is overlaying a carbon lens, a carbon layer, so that people can know more about the carbon emissions in the power system. Then the next step is, can we actually take action? Then, the end-user side is still going to be part of the picture.”