JuliaHub Inc., an agentic industrial engineering startup, has raised $65 million in Series B funding led by Dorilton Capital, with participation from General Catalyst, AE Ventures, and Snowflake's former CEO Bob Muglia. The capital will accelerate the development of Dyad 3.0, an AI agent platform designed to automate the design, testing, and maintenance of complex industrial hardware like semiconductors, satellites, and lithium batteries.
While software engineers enjoy AI coding tools like Claude Code and GitHub Copilot, industrial engineers are still reliant on legacy design tools and simulation platforms, leading to timelines that stretch months or years. JuliaHub argues this contributes to a massive infrastructure gap requiring over $106 trillion in investment by 2040, per McKinsey & Co.
Dyad 3.0, powered by the Julia programming language, enables "agentic engineering at scale"-a cloud-based environment where AI agents grounded in physics laws can design ultra-realistic systems and stress-test machines. In tests, it automated the entire design process for chemical manufacturing controllers, a task that normally takes months of manual work.
A key challenge is preventing AI hallucinations in safety-critical applications. JuliaHub's solution is scientific machine learning, blending real-world sensor data with physics-based equations to ensure accuracy. Agents can create digital twins and simulate stress tests for systems like bridges or pumps.
Early results show promise: with Binnies, JuliaHub built a digital twin of a water pump system predicting failures with over 90% accuracy using only four sensor inputs. It also worked with Synopsys to enhance chip development, transforming system-level engineering.
The startup aims to make Dyad the industry standard for AI-native engineering, using the Series B funds to scale go-to-market efforts and integrations. Long-term, it envisions autonomous operations where AI agents manage complex machines with minimal human intervention.