The Scoop
Sundar Pichai’s former chief of staff has started an AI company tailored to oil refineries and chemical plants, the latest bet that smaller upstarts can pick off specialized industries left behind by the tech giants.
Called Archimetis, the company analyzes data points from thousands of sensors, flags inefficiencies, and uses agents to offer up fixes. Like Cursor did for coding and Harvey did for law, Archimetis’ tool hopes to tap into the $1.7 trillion global market for refineries.
For example, if the data indicates the equipment that performs heat transfers is lagging, it can flag “gunk” built up in the machine and suggest what could have caused the build-up — something that would take engineers “hours,” said Cofounder Paul Manwell, who worked at Google for 14 years before striking out on his own.
Archimetis has been operating in stealth for two years and counts DeepMind’s chief scientist Jeff Dean among its investors, Manwell said. (Dean didn’t respond to a request for comment.) It has raised $11.5 million since it began operations with a pitch focused on saving companies 10% to 20% of their energy costs.
Know More
Unlike the larger AI firms, startups like Cursor and Harvey have grabbed market share by going deep, rather than wide. They’re not promising AGI and aren’t on the hook for building out AI infrastructure. Instead, they’re creating the kind of hyper-tailored products that companies could probably make on their own but don’t have the resources, time, or expertise. General-purpose AI still falls short of completing real-life work projects, and many companies say it isn’t all that helpful in their specific industries — yet.
“At a ‘Google,’ anything like that would just break everyone’s brain. The Gemini team would say, ‘Why are you doing something so specific?’” Manwell said.