Petersson said in a LinkedIn post that he resigned from OpenAI three months ago and that Energy is now launching. He described Energy as a tool that can turn multi-step computer work into a simple instruction, saying users can complete days of work in hours. 2 Min Read
Gabriel Petersson, a former OpenAI AI researcher, has launched Energy, an AI startup focused on helping users execute work across multiple applications.
Petersson said in a LinkedIn post that he resigned from OpenAI three months ago and that Energy is now launching. He described Energy as a tool that can turn multi-step computer work into a simple instruction, saying users can complete days of work in hours. According to Energy’s website, users tell the platform the outcome they want, after which it gathers relevant context from emails, files, conversations and connected tools, completes the required steps and returns the result for review. Energy connects with tools including Gmail, Google Drive, Google Calendar, Slack and Outlook. The company said users can choose which tools to connect and specify what each tool is allowed to access. Petersson also said in a post on X that the company is focused on the gap between the current use of AI for computer-based work and what he expects adoption to become. how we compare:
0.1% of people outside sf are using ai for all their work, and soon 100% will
either the products are not right, or people don’t understand
we are trying to figure both out, and that’s easier if we ship fast and build in publichttps://t.co/9kgpvbSeRQ
— gabriel (@gabriel1) August 6, 2026 He said Energy is trying to determine whether that gap exists because the products are not right or because people do not understand how to use them.
As per its website, Energy said one of its key differences from Cowork and ChatGPT Work is that it allows users to run their work with any model. It also lets users set up assistants with a click instead of manually managing memory, skills and automations.
The company said Energy is aimed at founders, operators and teams whose work is spread across multiple tools and who want to spend more time on decisions than coordination. Also Read: OpenAI’s new device will be hockey puck-sized and cost over $300