Dyad, the world’s first agentic AI platform for hardware engineering to be brought to market, brings physical AI to the design and testing of complex systems, reducing R&D time from months to just days.
JuliaHub announced the launch of Dyad 3.0 and a $65 million Series B funding round led by Dorilton Capital, with participation from General Catalyst, AE Ventures, and technology investor and former Snowflake CEO Bob Muglia. Dyad represents a fundamental shift in how physical systems are designed and built, introducing autonomous artificial intelligence agents into the digital design and testing of industrial machinery. From heat pumps and satellites to semiconductors, engineering teams can reduce design, test, and build cycles from months to minutes. Several Fortune 100 companies are already using Dyad and Julia across various industries, including aerospace , government , automotive , HVAC , and utilities.
Daniel Freeman, who led the Series B funding round for Dorilton Capital, commented: “Systems modeling is one of the most strategically important layers in the AI-native engineering stack, as it’s where physics, control logic, and AI converge. JuliaHub has created something extraordinary with Dyad: a platform that not only models systems but compiles them, enabling engineers to move from concept to production control code in a single environment. We believe JuliaHub has the potential to become a game-changer in the field of physical AI, and we are proud to support the team as they accelerate Dyad’s time to market.”
The difficult problem of innovation in the field of computer hardware
Physical engineering is one of the largest sectors that has yet to fully benefit from the AI revolution. While tools like Claude Code, Codex, and Gemini have transformed software development, industrial engineers remain limited by existing tools. McKinsey estimates that $106 trillion in cumulative investments will be needed by 2040 to meet the demands of new and modernized infrastructure. Engineers planning and implementing these upgrades need a solution that allows them to evolve at the pace of AI-enhanced software. That’s where Dyad comes in.
Dyad offers engineering teams an AI-powered environment to model, test, and validate industrial systems: think Claude Code for the physical world. Dyad 3.0 launches and builds upon Dyad 1.0, launched in June 2025, and Dyad 2.0, launched in December 2025. Dyad connects autonomous agents with scalable physical simulations, rigorous controls, safety analysis, and the ability to generate code for embedded systems to bridge the gap between software and the real world. Whether it’s a wastewater treatment plant or a car, you no longer need a PhD to develop highly detailed digital twins, fine-tune controllers for specialized deployment scenarios, and iterate on hardware designs to build the most efficient machine the first time.
“This isn’t about helping engineers accomplish a small task at a time. This is about large-scale agentic engineering, where teams can hand over a complete specification to Dyad to design the entire system. You input the specifications, he takes care of the design,” said Viral Shah, CEO of JuliaHub.
Also Read: apexanalytix Launches QubitOn for Global Risk and Validation
Digital twins with scientific machine learning
Dyad’s cloud-based agents are designed to continuously analyze global scientific knowledge to constantly improve models. AI-powered automated lab tests are performed to ensure the models accurately reflect physical reality. Streaming data combined with scientific machine learning ( SciML ) allows models to evolve automatically as the system learns from the real world. Dyad’s ecosystem and simulation language provide a foundation upon which all this learning is fed back to engineers to verify processes, determine if assumptions meet customer requirements, and act as the human element in the loop, ensuring the safety of the final product. Dyad’s design means engineers don’t have to write every line of code to test millions of designs, while still giving them the right tools to keep aircraft airborne.
Prith Banerjee, Senior Vice President of Innovation at Synopsys, said of the partnership with JuliaHub : “Dyad is transforming system-level engineering by combining scientific AI, agentic modeling, and a powerful build pipeline into a unified workflow. Integrated with Synopsys’ Ansys TwinAI™ simulation software, it enables high-fidelity hybrid digital twins by integrating physics-based simulation with data-driven models. What once required significant manual effort can now be accomplished much more efficiently, accelerating the entire digital engineering lifecycle and redefining how software-defined intelligent systems are designed and validated.”
Dyad is implementing AI for real-world science
General-purpose AI cannot guarantee that a model obeys the laws of physics. In physical engineering, an error is not a bug to be fixed; it’s the collapse of a bridge or the burning of a battery. This is what has so far prevented AI from playing a significant role in hardware engineering. In the recent comparative analysis of agents for modeling chemical processes, general-purpose LLM systems such as Codex, Claude Code (Opus), and Gemini barely completed the initial setup. Dyad has almost entirely automated the process of creating predictive controllers for models to optimize the yields of a chemical plant—a task that would normally take weeks.
“Engineering systems design software is undergoing a transformation, and Dyad is at the forefront of this evolution. Previous generations of tools don’t deliver the promised productivity or integration to unlock the value of AI. With Dyad, you can model physics, develop control algorithms with automatic code generation, and create accurate digital twins and surrogates for the rapid development of deep learning inference models—all powered by AI. Dyad comes in where physics meets analytics, and where customers and shareholders win,” said David Joyce, former CEO of GE Aviation and Vice Chairman of GE.
Dyad’s modeling language is designed to be easily understood by artificial intelligence agents. Its fundamental logic is based on the laws of physics, enabling its agents to reason about how fluids move within machines, how wind speed and temperature affect components, and how fundamental forces such as gravity influence design. This results in physically valid models that engineers can rely on. For example, in partnership with Binnies, a century-old water management company, and Williams Grand Prix Technologies, JuliaHub developed a SciML-powered digital twin that uses just four sensors to predict pump failures in water distribution systems with over 90% accuracy.
“Dyad represents a radical shift for the water industry, enabling a move from reactive operations to predictive, system-level decision-making,” said Tom Ray, Director of Digital Products and Services (Digital Twins and AI) at Binnies. “It has the potential to transform how businesses model real-world complexity, anticipate failures, and optimize performance on a daily basis.”
Source: PRNewswire
