Dyad, the world’s first-to-market agentic AI platform for {hardware} engineering, brings bodily AI to complicated techniques design and testing, compressing R&D time from months to mere days.
CAMBRIDGE, Mass., April 30, 2026 /PRNewswire/ — As we speak, JuliaHub broadcasts the launch of Dyad 3.0 and a $65M sequence B funding spherical led by Dorilton Capital, with participation from Normal Catalyst, AE Ventures, and know-how investor and former Snowflake CEO Bob Muglia. Dyad marks a elementary shift in how bodily techniques are designed and constructed, bringing autonomous AI brokers into the digital design and testing of commercial machines. From warmth pumps to satellites to semiconductors, engineering groups can compress cycles of design, testing, and constructing from months to minutes. A number of Fortune 100 corporations are already leveraging Dyad and Julia throughout a number of industrial sectors equivalent to aerospace, government, automotive, HVAC, and utilities.
Daniel Freeman, who led the Sequence B spherical for Dorilton Capital, commented: “Methods modeling is without doubt one of the most strategically essential layers of the AI-native engineering stack, as a result of it’s the place physics, management logic, and AI converge. JuliaHub has constructed one thing extraordinary with Dyad: a platform that does not simply mannequin techniques, however compiles them, taking engineers from idea to manufacturing management code in a single atmosphere. We imagine JuliaHub has the potential to turn into one of many defining corporations in Bodily AI, and we’re proud to again the group as they speed up Dyad’s path to market.”
‘The arduous drawback’ of {hardware} innovation
Bodily engineering represents one of many largest sectors but to totally profit from the AI revolution. Whereas instruments like Claude Code, Codex, and Gemini have reworked software program growth, industrial engineers have remained constrained by legacy instruments. McKinsey estimates {that a} cumulative $106 trillion in funding will probably be mandatory by means of 2040 to fulfill the necessity for brand new and up to date infrastructure. The engineers planning and constructing these updates want an answer that permits them to maneuver on the tempo of AI-enhanced software program. That is the place Dyad is available in.
Dyad offers engineering groups an AI-first environment to model, test and validate industrial systems: suppose Claude Code for the bodily world. Dyad 3.0 launches right this moment and builds on Dyad 1.0, which launched in June 2025, and Dyad 2.0, which launched in December 2025. Dyad connects autonomous brokers with scalable physics simulations, rigorous controls, security evaluation, and the power to generate code for embedded techniques to bridge the hole between software program and the true world. Whether or not it is a wastewater facility or an vehicle, a scientific PhD is now not required to develop extremely detailed digital twins, tweak controllers for specialised deployment eventualities, and iterate on {hardware} designs to construct essentially the most environment friendly machine proper the primary time.
“It is not about serving to engineers full one small process at a time. It is agentic engineering at scale, the place groups can feed a full specification to Dyad and have it design the entire system. Spec in. Design out,” mentioned Viral Shah, CEO of JuliaHub.
Digital Twins with Scientific Machine Studying
Dyad’s cloud-based brokers are designed to constantly scan by means of the world’s scientific data to continuously enhance fashions. AI-automated lab testing is rising to make sure fashions match bodily actuality. Streaming information blended with Scientific Machine Studying (SciML) makes it potential for fashions to mechanically develop because the system learns from the true world. Dyad’s simulation ecosystem and language supply a basis on which all of those learnings are relayed again to engineers to test the processes, decide whether or not assumptions match buyer necessities, and be the human within the loop that ensures the protection of the ultimate product. Dyad’s design means engineers don’t have to write down each line of code to be able to strive hundreds of thousands of designs whereas giving engineers the best instruments to verify planes keep within the sky.
Prith Banerjee, Senior Vice President of Innovation at Synopsys commenting on the partnership with JuliaHub says, “Dyad is remodeling system-level engineering by combining scientific AI, agentic modeling, and a strong compilation pipeline right into a unified workflow. Built-in with Synopsys simulation software program Ansys TwinAI™, it allows excessive constancy hybrid digital twins by integrating physics-based simulation with data-driven fashions. What as soon as required in depth handbook effort can now be carried out much more effectively, accelerating all the digital engineering lifecycle and redefining how clever, software-defined techniques are designed and validated.”
Dyad to implement AI for Science in the true world
Normal-purpose AI can’t assure {that a} mannequin obeys the legal guidelines of physics. In bodily engineering, an error will not be a bug to be patched; it is a bridge collapse or a battery fireplace. This has been the barrier blocking AI from taking part in a significant position in {hardware} engineering, till now. In recent agentic benchmarking for chemical course of modeling, basic LLM techniques equivalent to Codex, Claude Code (Opus), and Gemini barely accomplished the preliminary setup. Dyad virtually totally automated the entire course of of making model-predictive controllers to optimize yields of a chemical plant, a process that may usually take weeks.
“There’s a disruptive transition occurring in engineering system design software program, and Dyad is on the leading edge. Earlier generations of instruments don’t present the promised productiveness, or integration to unlock the worth of AI. With Dyad, you’ll be able to mannequin the physics, develop management algorithms with auto code era, and create correct digital twins and surrogates for fast growth of deep studying inference fashions, all enabled by AI. Dyad operates the place physics meets analytics, and prospects and shareholders win!” mentioned David Joyce, former CEO of GE Aviation and Vice Chair of GE.
Dyad’s modeling language is purpose-built to be simple for AI brokers to grasp. Its foundational logic is grounded within the legal guidelines of physics, permitting its brokers to motive about how fluids transfer by means of machines, how wind velocity and temperature have an effect on parts, and the way elementary forces like gravity form design. This produces bodily legitimate fashions that engineers can belief. As an example, in partnership with Binnies, an organization with a 100-year heritage in water administration, and Williams Grand Prix Applied sciences, JuliaHub developed a SciML–powered digital twin that makes use of simply 4 sensor inputs to foretell pump faults in water distribution techniques with over 90% accuracy.
“Dyad represents a step-change for the water business, enabling a transfer from reactive operations to predictive, system-level resolution making,” mentioned Tom Ray, Director of Digital Merchandise & Providers (Digital Twins & AI) at Binnies. “It has the potential to rework how corporations mannequin real-world complexity, predict failure, and optimize efficiency on daily basis.”
Be a part of us for the Dyad 3.0 Launch occasion
Dyad 3.0 will probably be officially unveiled at a dwell occasion subsequent month on Could 19. Join us to see dwell product demonstrations and listen to from our prospects on how they use Dyad throughout industries starting from Aerospace to HVAC to utilities to Robotics.
For extra data and media inquiries, contact: [email protected]
About JuliaHub
JuliaHub is a pacesetter in Scientific AI, and its mission is to empower these tackling the world’s hardest scientific and technical challenges with cutting-edge AI-first instruments in a seamless, safe atmosphere. The corporate was based in 2015 by the creators of Julia, the high-performance open-source language developed at MIT and now utilized by over 1,000,000 builders worldwide. JuliaHub combines superior mathematical computing and machine studying experience to allow Scientific Machine Studying (SciML) strategies, Digital Twin options, and next-generation modeling and simulation in aerospace, automotive and different industrial verticals.
Photograph: https://mma.prnewswire.com/media/2970068/Satellite_Photovoltaics_JuliaHub_Dyad.jpg
Photograph: https://mma.prnewswire.com/media/2970067/Cooling_circuit_JuliaHub_Dyad.jpg
Brand: https://mma.prnewswire.com/media/2826187/JULIAHUB_Logo.jpg
