Encube

What it is

Encube is a browser-based collaborative design platform built for hardware engineering teams. It sits between the CAD system where 3D modeling happens and the PLM system where release management lives — a gap where enormous amounts of engineering data currently live in scattered files, inboxes, and individual engineers' heads. Encube turns that space into a shared visual workspace where complex CAD models and heavy engineering drawings are first-class citizens, loading in two to three seconds on a standard laptop without an expensive graphics card. The speed comes from a custom GPU-accelerated spatial compute engine the Encube team built from scratch. The deeper layer is what makes Encube genuinely different: a deterministic manufacturability engine. This is the 'compiler for hardware' that the industry has never had. Software AI agents can self-correct because there is a feedback loop — compilers, linters, and unit tests tell the agent whether the output is right. Hardware has had no equivalent. Encube's engine answers the questions that matter: can this part actually be manufactured, what are the cost drivers, what breaks under stress, what are the von Mises stress concentrations. These answers are deterministic — run the same compute twice, get the same answer twice — which is a non-negotiable requirement in physical engineering. Large language models in Encube are reserved for translation and explanation at the edges, never for the math. For hardware teams at OEMs, tier-one suppliers, and deep-tech startups, Encube compresses the design-to-review cycle from weeks to days and enables concurrent engineering — exploring five design paths in parallel in the time it used to take to explore one. Founded in late 2021 by Hugo Nordell and Johnny Bigert, Encube exited stealth in October 2025 with a $23M round led by Kinnevik, validated through R&D programs with Volvo Group, Scania, and Beyond Gravity.

Who it's for

Hardware engineering teams at OEMs, tier-one suppliers, deep-tech startups, and aerospace companies who are bottlenecked by slow, siloed design tooling — specifically engineers who know their CAD and PLM systems are not talking to each other and who want to shift manufacturability checks earlier in the design process before costs lock in.

Why it's better

• •Large CAD assemblies load in 2–3 seconds on a standard laptop — no GPU required, no 20-minute file-open wait.
• •Deterministic manufacturability engine gives AI agents the feedback loop hardware has never had — the compiler for atoms.
• •Sits between CAD and PLM rather than trying to replace either, so it integrates with existing workflows instead of displacing them.
• •LLMs are isolated to translation and explanation — never touch the geometry, physics, or anything irreversible.
• •Shifts manufacturability insight left to the design phase, where catching a problem costs cents versus the six-month delays it costs at production.