JuliaHub Raises $65M Series B and Launches Dyad 3.0 — Agentic AI Platform Targets Simulink (April 30, 2026)

JuliaHub on April 30, 2026 raised a $65 million Series B led by Dorilton Capital, with participation from General Catalyst, AE Ventures, and former Snowflake CEO Bob Muglia, and simultaneously launched Dyad 3.0 — an agentic AI platform that pulls hardware engineering inside the same loop developers already use for software.

What Happened

Boston-based JuliaHub — the commercial backer of the open-source Julia programming language — announced its Series B alongside the public release of Dyad 3.0. Dorilton Capital led the round; participants include General Catalyst, AE Ventures, and Bob Muglia, who built Snowflake's enterprise business as CEO from 2014 to 2019.

Dyad is described by JuliaHub as the first agentic AI platform purpose-built for hardware engineering. Where traditional industrial design tools like MathWorks Simulink rely on a human engineer dragging blocks around a canvas, Dyad's modeling language is grounded in physical laws — fluid flow, gravity, electromagnetism, thermodynamics — and is designed to be read and written by autonomous AI agents as well as humans. The platform compiles those models all the way down to production control code in a single environment, compressing what JuliaHub claims is a months-long design-test-build cycle into minutes.

Key Details

• $65 million Series B — led by Dorilton Capital with General Catalyst, AE Ventures, and Bob Muglia participating
• Daniel Freeman of Dorilton Capital takes a board seat and called Dyad "a platform that doesn't just model systems but compiles them"
• Dyad 3.0 compiles physics-based models all the way to production control code, targeting Simulink as the primary incumbent it's trying to replace
• Customers already using Dyad include Fortune 100 companies in aerospace, automotive, HVAC, utilities, and government, plus a SciML-powered digital twin built with Binnies and Williams Grand Prix Technologies that predicts pump faults in water distribution from four sensor inputs at over 90% accuracy
• Live launch event scheduled for May 19, 2026 with full product demos and customer case studies across aerospace, robotics, HVAC, and utilities

What Developers and Engineers Are Saying

Reaction in the Julia community was immediate. On the official Julia Discourse forum, the announcement thread filled with questions about open-source vs commercial Dyad licensing and how Dyad models interoperate with existing ModelingToolkit.jl code. The general sentiment was cautiously positive — long-time Julia users have been waiting years for the language to find its enterprise foothold, and a $65M round at this stage is the clearest signal yet that scientific computing is JuliaHub's beachhead. On Axios, Bob Muglia framed Dyad as a direct attempt to take the Simulink market, calling MathWorks' tooling "a productivity ceiling that hasn't moved in 20 years." Skeptics on Hacker News pointed out that displacing Simulink in regulated industries like aerospace and automotive will require certification work that no AI-first tool has cleared yet.

What This Means for Developers and Engineers

For Julia developers, the round is the strongest commercial validation the language has had. JuliaHub now has the runway to keep funding core Julia maintenance, the SciML ecosystem, and the enterprise distribution that Simulink-replacement deals require. For mechanical, electrical, and aerospace engineering teams already evaluating modern alternatives to MATLAB/Simulink, Dyad 3.0 is the most credible challenger to enter the market in over a decade — but the pragmatic move is to wait for the May 19 product demo and the Dyad licensing details before committing a new project. For AI agent builders, the most interesting technical bet is Dyad's claim that its modeling language was designed agent-first: if true, it sets a precedent for purpose-built DSLs that LLMs operate natively rather than retrofitted natural-language wrappers around legacy tools.

What's Next

JuliaHub will officially unveil Dyad 3.0 at a live event on May 19, 2026, with customer demos spanning aerospace, automotive, HVAC, utilities, and robotics. The funding will go toward expanding the engineering team behind Dyad, scaling enterprise sales, and continued investment in the open-source Julia language and SciML ecosystem. The company has not announced an open-source release of the Dyad agent runtime itself, but has confirmed that the underlying modeling primitives remain part of the open Julia ecosystem.

Sources

• JuliaHub Blog — primary announcement from the company
• PR Newswire — official press release
• Axios — Bob Muglia interview framing Dyad vs Simulink
• SiliconANGLE — independent reporting and customer breakdown
• Engineering.com — engineering-industry context
• Julia Discourse — community discussion thread