• AI startup P-1, founded by former DeepMind and Airbus executives, has raised $23 million to develop an AI engineering assistant named Archie.
  • The AI system is designed to handle complex engineering tasks, reflecting a shift in AI research focus toward real-world applications.

P-1, an artificial intelligence startup that has operated in stealth until now, has secured $23 million in funding, with the round led by Radical Ventures. The company is developing an AI agent called “Archie” aimed at automating engineering tasks typically handled by humans in the physical world.

Founded by Aleksa Gordić, a former researcher at Google DeepMind, and Paul Eremenko, former CTO of Airbus, P-1 is positioning Archie as an artificial general intelligence (AGI) system capable of advanced reasoning and engineering design.

“Our vision is simple: we want to build an engineering AGI for the real world to help us design airplanes, Dyson spheres, cars, HVAC systems, etc.,” Gordić stated in a post on X.

Archie is designed to handle tasks that require multi-physics and spatial reasoning. According to P-1, the AI can derive key design drivers from given requirements, generate and compare design concepts, conduct early-stage trade-offs, and choose suitable tools for detailed engineering.

The company says these functions are intended to streamline and augment workflows currently dominated by human expertise.

“Our aim is [to ensure] that every engineering team at every major industrial company has an Archie as a team member, focusing initially on the dull and repetitive tasks, enhancing the team’s bandwidth and productivity, learning from real-world feedback and data, getting smarter and smarter, and ultimately helping humankind build things we don’t know how to build today,” Eremenko said.

The announcement comes amid growing industry conversations about the limitations of large language models (LLMs) and the need for AI systems capable of operating in the physical domain. Speaking at the recent NVIDIA GTC 2025 event, Meta AI Chief Scientist Yann LeCun pointed to a shift in focus within the AI research community.

“I am not so interested in LLMs anymore,” LeCun remarked, highlighting instead four core pillars of machine intelligence: understanding the physical world, persistent memory, reasoning, and planning. He emphasized the importance of “world models”—internal representations that support reasoning and predictions.

“We all have world models in our minds. This is what allows us to manipulate thoughts, essentially,” he said.

With the launch of Archie, P-1 joins a growing number of startups looking to apply AI to industrial and engineering challenges, suggesting a widening scope for AI applications beyond text and image generation.

Edited by Harshajit Sarmah