JuliaHub's most recent round is a $65M Series B announced on April 30, 2026, led by Dorilton Ventures, with participation from General Catalyst, Menlo Ventures, AE Ventures, HighSage Ventures, and Founder Collective.

Earlier rounds include a $4.6M seed in 2017, a $24M Series A in July 2021 led by Dorilton Ventures, with Bob Muglia joining the board, and a $13M strategic investment in June 2023 from AE Industrial Partners HorizonX.

SEC filings from early 2026 show $20M in SAFEs that converted into the Series B at closing. A May 2026 Form D reflects $69.9M in total offering amount and $46.7M sold at that point, excluding the converted SAFEs. JuliaHub has raised $106.6M in total disclosed funding across all rounds.

JuliaHub is a cloud platform for technical computing built around the Julia programming language, with Dyad as its industrial simulation and digital twin product on top of that infrastructure. In practice, an engineer or scientist uses JuliaHub much like a software team might use GitHub, cloud notebooks, managed HPC, and compliance tooling, but the stack is tuned for numerical computing, simulation, and physics-based modeling rather than general software development.

A typical workflow starts in a browser-based Julia IDE or Pluto notebook, with work organized into shared Projects where teams collaborate on code, datasets, and model outputs, then moves to batch jobs or larger compute instances launched on demand through CloudStation. Machine options range from small CPU instances up to 32-vCPU, 256GB memory configurations and V100 GPU-backed instances, and the platform can scale Julia programs to thousands of nodes for large HPC workloads. Each completed job captures its code, environment, inputs, and outputs for exact later reproduction through Time Capsule, which matters in regulated industries like pharma, where 21 CFR Part 11 compliance, audit trails, and IQ/OQ qualification tooling are platform features.

Dyad is the layer focused on industrial modeling. It is an AI-native environment for physical systems such as controls engineering, digital twins, battery design, HVAC, and multibody systems, where engineers can build models through either a visual drag-and-drop interface or a VS Code-based text workflow, with a one-to-one mapping between the two so physical engineers and software developers work from the same source of truth.

Within Dyad, engineers assemble components from physics libraries spanning electrical, rotational mechanics, thermal, hydraulics, and signal flow domains using acausal modeling, so components can be connected without hand-specifying every directional signal flow. They can then run transient simulations, steady-state analyses, PID autotuning, surrogate training, and parameter calibration, and export outputs as FMUs or embedded control code. The AI layer operates inside this constrained physics environment rather than through a generic chat interface: agents can derive governing equations, assemble models, run simulations, and verify physical consistency, with the generate-simulate-validate loop grounded in typed models and unit-aware structure.

A key adoption feature is migration tooling. Dyad can automatically translate MATLAB, Simulink, and Modelica models to Julia and Dyad syntax using LLMs, which matters because many industrial engineering teams have years of legacy models they cannot afford to rewrite from scratch. JuliaHub also sells professional services around migration, model development, and digital twin implementation to help customers through that transition.

JuliaHub sells to enterprise technical organizations through a B2B model that combines software subscriptions, usage-based cloud compute, domain application licenses, and professional services. Its go-to-market starts bottom-up, with individual engineers or small teams using free or low-cost access, then expands top-down when governance, compliance, and broader deployment require enterprise contracts.

Monetization has three layers. The base layer is access to JuliaHub as managed technical computing infrastructure, priced by subscription tier and compute consumption. The middle layer is domain applications, Dyad for industrial modeling and Pumas for pharmaceutical modeling, which are priced separately from the base platform. The top layer is professional services for migration, custom model development, and implementation, which can speed adoption in complex accounts and identify bespoke workflows that may later become product modules.

Because the product includes cloud compute, including GPU instances, large-memory machines, distributed HPC jobs, and hosted applications, JuliaHub has infrastructure cost exposure that makes its gross margin profile more mixed than pure software. The software and application layers carry higher margins, compute-heavy usage lowers blended margins, and services fall between those two. Heavy simulation workloads can also drive usage-based expansion within accounts as engineering programs scale.

The model has three reinforcing loops. Adoption of the open-source Julia and SciML ecosystem builds developer familiarity and package depth that can feed commercial platform conversion. Platform customers can expand into Dyad, Pumas, or internally deployed apps on the same substrate. Services-led migration work can also open accounts using MATLAB or Simulink, which then convert to recurring software and platform revenue. JuliaHub's vertical integration, from language runtime to cloud execution to domain application to embedded code generation, gives it more control over performance and user experience than vendors that assemble third-party components.

JuliaHub competes across two broad groups: legacy engineering software incumbents that already control simulation and controls budgets inside large industrial accounts, and cloud or infrastructure players that can intercept adjacent HPC and digital twin spend without requiring a new modeling language.

MathWorks is the most direct competitor because JuliaHub is positioning Dyad against workflows historically owned by MATLAB and Simulink. Simulink now has cloud and cluster execution, an online IDE, and a Simulink Copilot in its latest release cycle, so JuliaHub is competing against a platform that is still modernizing while retaining an enormous installed base, training footprint, and procurement familiarity inside automotive, aerospace, and industrial organizations.

MathWorks' structural advantage is workflow lock-in: models, toolboxes, code generation, verification, and requirements traceability are already embedded in engineering programs where switching costs are high. JuliaHub's counter is migration tooling that translates MATLAB and Simulink artifacts to Dyad automatically, plus an AI-native architecture that incumbents are retrofitting rather than building from scratch.

Siemens and Dassault Systèmes represent the broadest platform threat because both tie system simulation to CAD, PLM, MES, and automation and sell a unified industrial thread from design through operations. Siemens closed its Altair acquisition in 2025, strengthening its simulation, HPC, and AI portfolio under the Xcelerator umbrella, while Dassault's 3DEXPERIENCE platform anchors digital twin procurement in accounts where PLM standardization dictates tool choices.

Synopsys completed its acquisition of Ansys in July 2025, creating a combined entity spanning electronics, systems, materials, and multiphysics simulation. Ansys Twin Builder and TwinAI give that combined company a credible digital twin offering, and JuliaHub has responded by integrating Dyad with Ansys TwinAI through a Synopsys partnership, a coopetition move that lets JuliaHub monetize its SciML strength inside Ansys-heavy accounts without requiring full displacement.

JuliaHub is structurally disadvantaged where digital twin procurement is controlled by enterprise PLM or industrial transformation teams already aligned to Siemens or Dassault. Its best openings are programs where buyers value simulation speed, model flexibility, cloud deployment, and scientific AI more than full-stack PLM conformity.

Modelon is the closest head-to-head competitor among smaller players because it shares JuliaHub's cloud and system-modeling orientation, is built around open standards like Modelica and FMI, and has added AI-enabled workflows to its positioning. Modelon reduces migration friction for organizations already invested in Modelica libraries, which is a meaningful advantage over Dyad's newer language and runtime.

Rescale competes from the infrastructure side, offering a digital engineering platform that supports over 1,250 HPC and AI software packages and can run incumbent solvers from Ansys, Siemens, and COMSOL on cloud compute. In enterprises that already own those solver licenses, Rescale can serve as the modernization layer without requiring any new modeling language, making it a real budget interceptor even if it is not a direct product substitute.

JuliaHub's expansion logic runs in two directions: broadening Dyad from a simulation tool into a fuller engineering system of record, and layering domain-packaged applications on the same JuliaHub infrastructure for additional verticals.

The largest near-term TAM expansion is Dyad's move toward an end-to-end engineering workflow spanning model creation, simulation, calibration, controls analysis, surrogate training, FMU generation, and embedded controller code deployment. Dyad 3.0, launched alongside the Series B in April 2026, frames this as a path from concept to production control code in a single environment, moving JuliaHub from competing for simulation budget into adjacent spend tied to controls software, verification, and operational deployment.

The battery and electrification adjacency is one of the more concrete examples. JuliaHub has already productized Dyad Batteries with electrochemical, thermal, and degradation modeling, and as EVs, stationary storage, and fast-charging systems require tighter coupling of cell physics, pack control, and field data, JuliaHub can extend from battery simulation software toward a broader battery digital twin and battery management workflow.

JuliaHub's current footprint in pharma and early industrial accounts creates an internal cross-sell path. A customer that starts with JuliaHub infrastructure for managed Julia compute can expand into Dyad for industrial modeling or Pumas for pharmaceutical workflows on the same platform, and Dyad's one-to-one mapping between graphical models and Julia code is designed to bring software developers and physical engineers onto the same source of truth, moving adoption from individual modelers to broader engineering organizations.

The ASML deployment of 700-plus engineers on JuliaHub provides a reference point for enterprise-wide standardization at scale. As more industrial organizations building software-defined machines look for a common substrate connecting physical modeling, AI, and production software, JuliaHub's platform architecture can expand from specialist team adoption into engineering-wide standardization.

JuliaHub's air-gapped deployment option, JuliaHub Air, AWS GovCloud support, and on-prem Kubernetes installation path open a TAM that generic SaaS simulation vendors cannot easily address. Government, defense, and export-controlled programs require deployment flexibility and auditability that cloud-only competitors often struggle to match, and once workflows are validated and authorized in those environments, retention tends to be high.

Europe represents a geographic opportunity given JuliaHub's existing industrial customer base there, the concentration of aerospace and automotive OEMs with model-based design programs, and the regulatory environment around digital engineering in sectors like pharmaceuticals and energy. The company's hosting of JuliaCon 2024 in Eindhoven and workshops in Amsterdam and Nuremberg point to an existing presence that could convert into enterprise accounts as Dyad matures.

Ecosystem inertia: Even with LLM-based translation tooling, displacing MATLAB and Simulink in organizations where models, certification processes, supplier interfaces, and engineering talent are already organized around those stacks requires a validated migration path that safety-critical programs in aerospace, automotive, and defense will accept before committing to a new runtime.

Agentic AI credibility: JuliaHub's differentiation now leans heavily on agentic AI for engineering workflows in industries where aerospace, pharma, and government buyers require rigorous human review, traceability, and domain validation before trusting generated models or control logic, so the path from compelling demo to scaled production revenue is gated by assurance processes rather than raw model-generation capability.

Julia adoption ceiling: JuliaHub's entire product stack is built on Julia, a language with over a million users and 13,000-plus packages but still a fraction of the MATLAB and Python ecosystems, which means enterprise standardization decisions can stall when procurement teams, hiring managers, or safety certification bodies treat language familiarity and community size as proxies for platform risk.