Khosla, Pritzker’s Inspired Capital Back AI for Power Systems, Pest Control Startup

(Bloomberg) —

While many artificial intelligence tools are aimed at software developers and office workers, venture capitalist Vinod Khosla and former Commerce Secretary Penny Pritzker’s Inspired Capital are investing in a startup deploying the technology for dirty, time-consuming infrastructure maintenance tasks like pest control and inspecting power poles and air-conditioning systems. 

San Francisco-based BrightAI has raised a $51 million Series A round led by Khosla Ventures  and Inspired Capital, along with BoxGroup, Marlinspike and VSC Ventures. The round valued the company at around $300 million, according to a person familiar with the company who didn’t want to be named discussing financial details. Upfront Ventures led a previous round, and the company has raised a total of $78 million. 

“So many people are so focused on the future of digital AI, but we’re excited about this new layer of AI: the physical world AI,” said  Alexa von Tobel, co-founder and managing partner at Inspired. “Think about every electricity pole that exists. If it’s tilting a little bit, that’s dangerous. That can start a fire.” 

Six-year-old BrightAI was founded by Alex Hawkinson, who previously started home technology company SmartThings. BrightAI’s platform —  called Stateful — relies on custom semiconductors and AI models that analyze data. Several different devices the company offers collect data, from sensors embedded in stickers that can be placed on power poles to a custom-built sensor head BrightAI offers for existing commercial robots. There’s also a bodycam-type device that workers who need their hands free can use. 

“When you look at these essential services, there’s big problems — most the infrastructure itself was kind of built around World War II, and now 80 years hence, it’s duct tape and barbed wire taping things together,” Hawkinson said. Plus the methods of inspecting aging infrastructure are time-consuming, expensive and aren’t based on any knowledge of what might actually be failing.

“I jokingly say we’re sort of still in the Roman era of infrastructure management,” Hawkinson said. “It’s that old way: You send people to periodically inspect it and it’s brutal, dirty, dangerous work.”

Some equipment, like underground water pipes, can’t even be inspected in person. That can mean leaks go undetected for a long time, negatively impacting performance and costing companies money.

One customer, Azuria Water Solutions, uses Stateful to train autonomous robots to scan pipes and then act by making precise cuts in those that are being repaired. Another, Pelsis, has created an AI-powered insect trap with sensors and cameras that warn facilities about potential infestations — a camera takes pictures of the trap’s glue board and an AI hub in the trap figures out the number, size and species of insects and sends that data to the cloud. 

To Khosla, part of the appeal is the atypical nature of BrightAI’s market — unlike many segments of the AI market, it’s ripe with opportunity and has little competition. Other startups offering AI to meet utilities’ needs include AiDash, which does satellite imagery analysis to help protect against wildfires, and a growing number of weather forecasting companies that feed the energy industry.

Hardware and physical use cases are more difficult to do well than software, Khosla said.

“When you come to high-value infrastructure — remote stuff like pipelines and power poles and electric utilities — there’s just far less competition for a greater need and it can be a very large market,” he said. “There’s lots of businesses like this that are not what I would call your ‘Silicon Valley hype put some software together’ kind of business, but they are also much more resilient and harder to displace once established.”

(Adds name of Pritzker’s investment fund in first paragraph. A previous version corrected amount of funds raised in subhead.)

To contact the author of this story:
Dina Bass in Seattle at dbass2@bloomberg.net

© 2025 Bloomberg L.P.