In late June, a hundred-odd tech workers gathered on the sixteenth floor of San Francisco’s old Standard Oil building for a classic Friday night in Silicon Valley: the kick-off for a 24-hour hackathon, hosted by OpenAI, cloud platform Crusoe, and Lowercarbon Capital.
The assorted engineers and designers were tasked with coming up with a way to use artificial intelligence to get more clean energy projects online — by streamlining permitting, simplifying tax credit monetization, or easing access to home electrification rebates, for instance.
Teams (clad in an assortment of San Francisco tech casual attire including a “Microgrids are dope” hoodie) competed for $4,500 in cash prizes, more than $20,000 in Crusoe compute credits, and lunch with a founding partner of Lowercarbon.
Hackathons, the tech industry’s classic tool for creating software-based solutions for seemingly intractable problems, may not be as ubiquitous as they were in the mid-2010s. But the AI boom is sweeping through every corner of the tech sector — and climate tech is no exception. As evidenced by the excitement in the room in June, the energy transition is an area ripe for hackathon-derived innovation.
And it’s perhaps unsurprising that AI heavyweights like OpenAI are throwing their authority behind the clean energy bottleneck. After all, AI computing is predicted to create massive electricity demand in coming years, and that sector’s success is predicated in part on the ability to get more renewable energy onto the grid.
“I’m looking for the next Peter Reinhart,” Lowercarbon partner Shawn Xu told Latitude Media at the event, referring to the Segment founder who later sold his customer data platform to Twilio before founding carbon removal company Charm. Reinhart was the winner of the first hackathon Xu ever organized a decade ago, in what he described as the “hackathon heyday.”
The classic hackathon involves martialing a troop of computer engineers to come up with software solutions to a specific problem. According to Xu, many of the engineers at the OpenAI hackathon have never worked in clean energy or climate. Accordingly, part of the event’s definition of success was simply to expand the climate tech tent to include more people.
“There are many great engineers who are climate-curious, who have not had an opportunity to study and learn about the energy transition,” Xu said. “How many of the folks here today…are going to decide they want to switch tracks because they’d rather not work on ads or games, and [instead] build some meaningful software that could help the energy transition or accelerate it?” Xu said.
Unlike many of the bets venture capitalists make on climate technologies, the hackathon was geared at near-term fixes to bottleneck problems.
“We’re thinking a lot about deployment,” he said, adding that the focus was specifically on streamlining solutions that have already been proven: getting wind and solar through permitting pipelines or electrifying low and medium income households, for example.
Winning projects included an AI-powered tool to identify and connect eligible households with financing for electrification, and an AI voice assistant trained on data from the Department of Energy and Pacific Northwest National Labs to help developers identify key decision makers for individual projects.
As with any hackathon, the hope is that real companies spin out from the event, Xu said. And in the midst of an election season that is swiftly gathering speed and controversy, he hopes that the event’s focus on near-term solutions will mean that those spin-outs will be “insulated” from much of the partisan politics often associated with clean energy.
Lowercarbon specifically is looking to invest in companies that are able to “sidestep” political hurdles, he added, by being closely tied to overall economic opportunity.
“To the extent that we have great folks who come out of this weekend with interesting projects that could turn into real companies, we’d love to keep track of those folks and see if we can support them,” he said.
