nTop and JetZero Are Building the Next Generation of Aircraft Design with NVIDIA NemoClaw
JetZero is designing a blended wing body (BWB) aircraft, one of the most geometrically complex configurations in commercial aviation, on a program schedule that demands a fundamentally different approach to engineering.
To get there, JetZero needed two things to work together: geometry generation that keeps up with an automated workflow, and an engineering intelligence layer capable of managing that workflow without human hand-holding at every step. They built both with nTop and the NVIDIA NemoClaw blueprint.
The Geometry Problem
Blended wing body design requires constant iteration across a deeply interconnected set of parameters. Change the fuselage profile and the wing sweep changes with it. Adjust an airfoil section and every downstream analysis, including aerodynamics, structures, and high-fidelity CFD, needs to run again. In traditional CAD and CAE, each of those updates is a manual process, and the geometry breaks often enough that engineers spend a meaningful share of their time just keeping the model intact.
JetZero's nTop model works differently. A fully parametric, logic-driven notebook encodes the entire aircraft as a single interconnected workflow: outer mold line, structural wingbox, airfoil sections, nacelles, winglets, landing gear, and interior volumes. Parameters change, the model regenerates correctly, and the output is simulation-ready every time. The model currently feeds aerodynamic panel analysis via AVL, with Flexcompute flow360 for high-fidelity CFD.
JetZero is bringing an entirely new aircraft configuration to market for the first time in decades. NVIDIA NemoClaw gives us the AI infrastructure to connect every engineering tool in that workflow. nTop is the geometry engine that makes all of it run.
Brad Rothenberg
CEO, nTop
The Automation Layer
The nTop geometry engine is now running as a native skill inside NVIDIA NemoClaw.
NVIDIA NemoClaw is an agentic engineering blueprint – an open-source reference stack that integrates with harnesses like OpenClaw, running on accelerated computing in the cloud, datacenter and local systems like NVIDIA DGX Spark personal AI supercomputers. Enhanced with NVIDIA technologies for accelerating CAE simulations, agents introspect the nTop notebook schema, populate design parameters, dispatch geometry jobs to GCP virtual machines, scaling to available compute as the design of experiments grows, and route outputs to the appropriate analysis tool, without manual intervention. The loop from engineering requirement to analyzed design variants runs autonomously.
What this means in practice: a single engineer can set up a design of experiments across dozens of BWB configurations and let the stack run overnight. The geometry generates cleanly, the analysis tools receive valid inputs, and results come back structured and ready to act on. The engineer's time goes to decisions that require engineering judgment, not to manual configuration, format conversion, or tool handoffs.
Autonomous agents are accelerating engineering workflows, but they need a backbone that keeps coordination intact across tools and environments. Using NVIDIA NemoClaw, nTop and JetZero are demonstrating the kind of throughput and design fidelity that agentic AI can deliver for the most demanding engineering problems.
Timothy Costa
Vice President and General Manager of Computational Engineering, NVIDIA
Where Things Stand
The JZ-BWB-Claw repository is progressing from a baseline airfoil test to a full general arrangement under active development. AVL aerodynamic analysis is wired directly to the parametric model. The nTop and NemoClaw integration is running end-to-end with GCP available for distributed computing.
JetZero CEO Tom O'Leary has engaged directly with nTop leadership on the program, describing nTop as having a significant role not just in the agentic demo but in the ongoing development of the aircraft itself.
"JetZero is a digital-native company, allowing us to develop the design of our all-wing airplane with confidence," said Tom O'Leary, JetZero CEO and co-founder.
nTop and NVIDIA together have designed tools that provide near-instantaneous downline information. The world needs better planes – lower fuel burn, a better experience – and nTop and NVIDIA help us accelerate the future everyone wants to see.
Tom O'Leary
CEO, JetZero
The infrastructure is built, and the full loop is running now, from engineering requirements to analyzed design variants, operating end-to-end without manual intervention.
nTop is the computational geometry engine behind some of the most advanced aerospace and defense programs in development today. Learn more at ntop.com.
