JuliaHub’s Dyad Cuts R&D Time From Months To Days Alongside $65M Series B

JuliaHub is launching Dyad 3.0 alongside a 65 million series B funding round led by Dorilton Ventures, signaling significant investor confidence in the application of agentic AI to physical engineering, a sector historically slow to adopt artificial intelligence. The company’s Dyad platform compresses research and development timelines for complex systems from months to days by bringing autonomous AI agents into the digital design and testing of industrial machines, ranging from heat pumps to satellites. McKinsey estimates a cumulative 106 trillion in investment will be necessary to meet the need for new and updated infrastructure, creating a critical need for accelerated engineering processes. “It’s not about helping engineers complete one small task at a time; it’s agentic engineering at scale,” said Viral Shah, CEO of JuliaHub, explaining that Dyad can design complete systems from a full specification.

65M Series B Funds Dyad 3.0 Agentic AI Launch

Led by Dorilton Ventures with participation from General Catalyst and others, this investment signals strong confidence in the potential of artificial intelligence to accelerate hardware development cycles. Several Fortune 100 companies are already utilizing Dyad across sectors including aerospace, automotive, and utilities, demonstrating early market traction. This funding arrives at a critical juncture, as McKinsey estimates a cumulative 106 trillion in investment will be necessary to meet the need for new and updated infrastructure. Dyad 3.0 aims to alleviate pressure on engineering teams by drastically reducing the time required for research and development, compressing R&D time from months to days and cycles from months to minutes. The platform distinguishes itself from software-focused AI tools by applying agentic AI, autonomous agents capable of independent action, to the complexities of physical systems design and testing.

JuliaHub positions Dyad 3.0 as an AI coding assistant for the physical world, drawing a direct comparison to Claude Code and highlighting its novel application to hardware. Dyad 3.0 builds upon previous iterations, launched in June and December, and connects autonomous agents with scalable physics simulations, controls, safety analysis, and code generation for embedded systems. This integration bridges the gap between software and real-world applications, enabling the creation of detailed digital twins without requiring extensive expertise. This is agentic engineering at scale, where teams can feed a full specification to Dyad and have it design the complete system.

“Design out,” said Viral Shah, CEO of JuliaHub. The platform’s cloud-based agents continuously learn from scientific knowledge and real-world data, automatically improving models and ensuring physical validity, a crucial distinction given that errors in physical engineering can have catastrophic consequences. David Joyce, former CEO of GE Aviation and Vice Chair of GE, stated, “There is a disruptive transition occurring in engineering system design software, and Dyad is on the leading edge.

Dyad 3.0 Connects Agents to Physics-Based Simulations

The convergence of artificial intelligence and traditional engineering is accelerating, moving beyond software applications to encompass the complexities of the physical world. While AI has demonstrably reshaped software development, hardware engineering has largely remained reliant on established, time-consuming processes. JuliaHub’s recent launch of Dyad 3.0, coupled with $65 million in Series B funding led by Dorilton Ventures, signals an effort to bridge this gap and introduce agentic AI to industrial digital twins. This investment underscores growing confidence in applying AI to sectors historically slow to adopt the technology. The platform’s core innovation lies in its ability to model, test, and validate industrial systems, compressing R&D time from months to days and cycles from months to minutes. This comparison highlights the platform’s ambition to provide a similar level of automation and intelligence to hardware engineering.

A key differentiator is the platform’s grounding in physical laws, ensuring models are not only efficient but also demonstrably safe; a critical consideration where errors can have catastrophic consequences, unlike software errors that can be patched. In agentic benchmarking for chemical process modeling, Dyad almost entirely automated the creation of model-predictive controllers, a task that typically requires weeks of engineering effort.

90% Accuracy Achieved in Water Pump Fault Prediction

JuliaHub’s recent advancements in agentic AI are yielding tangible results in critical infrastructure management, demonstrated by a partnership with Binnies and Williams Grand Prix Technologies. The collaboration has produced a scientifically-powered digital twin capable of predicting faults in water distribution systems with over 90% accuracy, a significant leap forward in preventative maintenance. This predictive capability relies on a system that requires only four sensor inputs to assess pump health, streamlining data collection and analysis. This approach moves beyond traditional modeling techniques, offering a dynamic system that learns and adapts to real-world conditions. “Dyad represents a step-change for the water industry, enabling a move from reactive operations to predictive, system-level decision making,” said Tom Ray, Director of Digital Products & Services (Digital Twins & AI) at Binnies. The implications extend beyond cost savings; proactive fault prediction minimizes disruptions to essential services and reduces the risk of potentially damaging failures.

This achievement highlights the potential of applying agentic AI, AI designed to act autonomously, to physical engineering challenges. Unlike software-focused AI tools, Dyad is built on a foundation grounded in the laws of physics, ensuring model validity and trustworthiness. The system’s ability to integrate streaming data with Scientific Machine Learning (SciML) allows models to automatically improve as they learn from real-world performance. This automated process, coupled with rigorous safety analysis, promises to accelerate innovation and improve the reliability of complex industrial systems.

Legacy Tools vs. Dyad’s AI-First Engineering Environment

The demand for new infrastructure is escalating rapidly, and the tools used to design and build it are undergoing a parallel shift. While traditional engineering software struggles to integrate artificial intelligence, JuliaHub’s Dyad 3.0 represents a departure, offering an “AI-first environment” designed to accelerate the development of complex physical systems. This contrasts sharply with legacy tools that require painstaking manual coding and iterative testing. The system’s modeling language is specifically designed for AI comprehension, grounded in the laws of physics to ensure the validity of generated models.