Leo AI Raises $9.7M for Large Mechanical Model

Cambridge-based Leo AI closed a $9.7 million seed round to expand its specialized AI platform for mechanical engineering. Flint Capital led the oversubscribed round, with participation from an Andreessen Horowitz scout, TechAviv, Two Lanterns VC, Bertrand Sicot (former CEO of SolidWorks), and Prof. Yossi Matias (VP at Google and Head of Google Research).

Leo built what it calls a Large Mechanical Model (LMM), essentially an AI trained on mechanical parts, product designs, and engineering standards instead of general text. Think of it like ChatGPT, but instead of learning from books and websites, it learned from millions of real-world products and engineering manuals.

The company claims this specialized training delivers 95% accuracy on engineering queries, compared to generic AI tools that often provide unreliable information for technical tasks. Engineers at HP, Scania, Siemens, and Mobileye are among the 50,000 users who have generated over 475,000 3D concepts using the platform.

The problem it solves

Mechanical engineers, the people who design everything from car engines to medical devices, currently toil on mundane tasks. According to Leo, engineers spend over 150 workdays per year just searching for information and selecting parts, leaving only half their time for actual innovation.

"When I started out as a mechanical engineer, I thought I'd be designing life-saving robots and next-gen vehicles. Instead, I spent weeks just searching for parts, and that's a harsh reality that most engineers face," said Maor Farid, Leo AI's co-founder. "Leo saves up to 5 hours of repetitive work each week. And, most importantly; it helps engineers get their spark and passion for innovation back."

The platform integrates with existing engineering software, allowing users to type requests like "show me a bolt that fits this hole" and receive instant recommendations based on their company's standards and available inventory.

Why this matters

Mechanical engineering inefficiencies create ripple effects across entire industries. Poor coordination and information silos contribute to nearly half of all product delays, according to Flint Capital's Sergey Gribov.

"Mechanical engineering is one of the most critical, yet underserved sectors when it comes to AI," Gribov said. "Nearly half of all product delays are driven by siloed knowledge. Leo is addressing a truly pressing need by speeding the development process up by as much as 70%. The traction with top-tier customers and the enthusiasm of the engineering community speak volumes."

These delays can increase manufacturing costs by up to 35%, making efficiency improvements particularly valuable in a mechanical engineering services market projected to reach $620 billion by 2032.

The security angle

Leo processes data within customers' existing systems rather than uploading it to external servers, addressing growing enterprise concerns about IP theft through AI tools. This approach becomes increasingly important as data breaches from third-party AI tools surge 28% month-over-month, costing companies an average of $15 million per incident.

Founded in 2023 by Dr. Maor Farid and Moti Moravia, Leo generated over $100,000 in revenue during its first month of monetization and attracted 200,000 site visitors without paid marketing.