JuliaHub, the AI infrastructure company built on the Julia programming language, closed a $65 million Series B led by Dorilton Capital on April 30. General Catalyst, AE Ventures, and former Snowflake CEO Bob Muglia also participated. Alongside the funding, JuliaHub launched Dyad 3.0, an agentic AI platform that deploys autonomous agents to design, simulate, and validate industrial hardware systems.
The Product
Dyad 3.0 positions itself as “Claude Code for the physical world,” according to PR Newswire. Engineering teams feed a full specification into the platform, and autonomous agents construct digital twins, run physics simulations, tune control systems, and generate embedded code. The company claims this compresses design-to-test cycles from months to days.
The platform covers aerospace, automotive, HVAC, utilities, and semiconductor design. Several Fortune 100 companies are already using Dyad and Julia across these sectors, according to the PR Newswire release.
“It’s not about helping engineers complete one small task at a time. It’s agentic engineering at scale, where teams can feed a full specification to Dyad and have it design the complete system. Spec in. Design out,” JuliaHub CEO Viral Shah told PR Newswire.
Why It Matters for Agent Infrastructure
Dyad 3.0 solves a gap that general-purpose AI coding tools cannot. In recent agentic benchmarking for chemical process modeling published in ScienceDirect, general LLM systems including Codex, Claude Code (Opus), and Gemini barely completed initial setup tasks. Dyad almost entirely automated the full process of creating model-predictive controllers for chemical plant yield optimization, a task that typically takes weeks, according to the PR Newswire release.
The core technical differentiator is Scientific Machine Learning (SciML), which constrains AI outputs to obey physical laws. General-purpose models have no guarantees that generated designs are physically valid. In hardware engineering, JuliaHub argues, a design error means a bridge collapse or a battery fire — not a software bug you patch and redeploy.
The Investors
Dorilton Capital, which led the round, is an investment arm of Dorilton Ventures. Daniel Freeman of Dorilton called JuliaHub “one of the defining companies in Physical AI” and described Dyad as the point “where physics, control logic, and AI converge,” according to PR Newswire.
The round also attracted a strategic endorsement from Synopsys. Prith Banerjee, SVP of Innovation at Synopsys, said the Dyad and Synopsys Ansys TwinAI integration “enables high fidelity hybrid digital twins by integrating physics-based simulation with data-driven models,” according to the release.
David Joyce, former CEO of GE Aviation, described the transition in the release: “Previous generations of tools do not provide the promised productivity, or integration to unlock the value of AI. With Dyad, you can model the physics, develop control algorithms with auto code generation, and create accurate digital twins.”
The Infrastructure Investment Gap
McKinsey estimates a cumulative $106 trillion in infrastructure investment will be needed through 2040, according to data cited in the PR Newswire release. The engineers building that infrastructure currently rely on legacy simulation tools. JuliaHub’s bet is that agentic AI can compress the engineering cycle fast enough to match the pace AI has already set in software development.
Dyad 3.0 builds on two prior versions: Dyad 1.0 launched June 2025, Dyad 2.0 in December 2025. The $65 million gives JuliaHub capital to expand into additional industrial verticals and scale its cloud-based agent infrastructure for enterprise deployment.