VXN
// SURFACE-NATIVE AUTONOMOUS SYSTEMS

Robots that move
like the surface
is their world.

One unit. Any surface. Any angle. Learns while it works.

GPU-native swarm units that traverse real geometry — not approximated floor plans. Built for environments that defeat every other system on the market.

MAX SURFACE ANGLE
90°
VERTICAL · INVERTED · CURVED
UNIT INTERFACE
CONFORMABLE
ONE UNIT · ANY SURFACE
FLEET BEHAVIOUR
SWARM
SPECIALIZED ROLES · NOT COPIES
IMPROVEMENT CYCLE
CONTINUOUS
FIELD-TRAINED · COMPOUNDS OVER TIME
Custom PCB Design
3D Fabrication
Conformable Unit Architecture
Brushless DC & Servo Drive
GPU-Native Planning
Continuous Field Learning
Calgary, AB
UNIT CAPABILITY · 01
The environment is fully known before the first unit moves.

Every competitor navigates an approximation — a 2D projection of a 3D world. VOXARN units work from the actual geometry at full fidelity. Walls are walls. Curves are exact. Inclines are not rounded down.

UNIT CAPABILITY · 02
Units coordinate roles. One stabilizes. One fastens. One monitors.

A swarm is not a fleet executing the same instruction. VOXARN units divide labour based on what the task actually requires. The collective does what no individual unit can — and compensates when one cannot.

UNIT CAPABILITY · 03
Every hour of field operation improves the next.

VOXARN fleets compound. The units on your site in twelve months are meaningfully more capable than those on day one — trained on your specific surfaces, your geometry, your operational patterns.

SYSTEM WHY VOXARN OUTPERFORMS THE FIELD OVERVIEW

What the field gets wrong.

Two decades of industrial automation have produced systems optimised for environments that don’t exist — flat, static, pre-mapped. The real world is none of those things.

Current generationVOXARN
SURFACE RANGEFloor-plane only. Ramps at reduced confidence. Walls, ceilings, and curved surfaces are not navigable.Any surface at any angle — walls, curves, inverted surfaces. Full geometric fidelity from the actual structure.
UNIT INTERFACEFixed hardware geometry. Each unit is designed for one surface type. Multi-surface deployment requires multiple hardware configurations.Conformable architecture. Brushless DC drive and servo-actuated contact surfaces adapt to whatever geometry the unit encounters.
UNIT BEHAVIOURHomogeneous fleet. Every unit runs identical logic. One unit fails — nothing compensates.Specialized roles, dynamic coordination. Units divide labour and compensate for failure in real time without manual intervention.
ENVIRONMENT INPUTSensor-mapped during deployment. Weeks of integration before a unit touches the real environment.Direct from your existing engineering models. The structure your team already has becomes the traversal map.
FIELD IMPROVEMENTStatic from deployment. Performance on day one equals performance on day three hundred.Continuous improvement through operation. Field time builds capability specific to your environment.
FLEET SCALABILITYFixed architecture. Adding or replacing a unit typically requires downtime and system reconfiguration.Hot-swap capable. Add a unit — it joins immediately. Replace one — the fleet adapts. No reconfiguration. No downtime.
CAPABILITIES FIVE CORE CAPABILITIES RESTRICTED
CAP-01 · GEOMETRY
True-scale surfaces, known before deployment.

The actual 3D structure of your environment becomes the traversal surface. No information loss. No translation step. No weeks of pre-deployment integration.

SURFACE-NATIVE
CAP-02 · INTERFACE
One unit. Any surface. Any interface condition.

Brushless DC drive and servo-actuated contact surfaces adapt to the geometry encountered. Flat deck, curved pipe, vertical wall — no hardware modification required.

CONFORMABLE
CAP-03 · LEARNING
Field operation builds capability. It doesn’t just consume it.

Every deployment cycle compounds. Performance improves continuously — without manual retraining, without update cycles, without downtime.

CONTINUOUS IMPROVEMENT
CAP-04 · COORDINATION
Division of labour at swarm scale.

One stabilizes a structure while another executes precision work. One monitors environmental conditions while another reroutes around an obstacle. The swarm functions as a trained crew.

ROLE-SPECIALIZED
CAP-05 · MODULARITY
Add a unit. Remove a unit. The fleet adapts.

Hot-swap a unit out — the fleet compensates immediately. Expand capacity — new units integrate without reconfiguration. Built for operational continuity.

HOT-SWAP · MODULAR
FIELD SPECS TECHNICAL ARCHITECTURE RESTRICTED

Navigation built for the real world’s actual geometry.

VOXARN · multi-unit surface traversal · toroidal mesh
UNITS 8
MESH_FACES 24,816
PLAN_MS 11
ALGO A*·FACE-COORD
FRAME 0
SYS-ATrue Surface Traversal

Units track position as barycentric face coordinates on the raw triangle mesh — no projection, no approximation. Walls, curves, and inverted surfaces navigated with the same precision as a floor.

SYS-BCAD-Native Environment Ingestion

Your engineers’ existing 3D models become the traversal map directly. No hand-crafted waypoints. No sensor calibration pass. The structure your team already has is all VOXARN needs.

SYS-CGPU-Parallel Unit Planning

All path planning runs on GPU — CUDA and ROCm/HIP, covering both NVIDIA and AMD hardware. Hundreds of units planned simultaneously with no sequential bottleneck.

SYS-DProprietary Stack — Full Ownership

No external navigation dependencies. No third-party planning library. No vendor lock-in on any critical path. Every layer is owned and optimised for swarm performance.

ACTIVE STACK
Python · Taichi (GPU sim)
PyTorch (network training)
CuPy · NumPy (array compute)
CUDA (kernel dispatch)
ROCm / HIP (AMD GPU parity)
C++ (systems layer)
Vulkan 1.3 (render pipeline)
Custom PCB (in development)
DEVELOPMENT MILESTONES
Multi-unit GPU sim on arbitrary mesh
Face-coordinate traversal fundamentals
Physics constraints + energy modelling
CAD / STL environment ingestion
90% simulation fidelity threshold
PCB design — unit v0.1
Motor & servo control loop integration
Hardware validation — first unit
// SIM NOTE
Canvas 2D demonstration only. Production target: GPU compute kernels via CUDA and Vulkan. Set SITE_CONFIG.CAD_DEMO_READY = true to unlock the CAD demo tab when footage is ready.
MEDIA SIMULATION CAPTURES RESTRICTED
SIM CANVAS · LIVE CAPTURE  
SIM STILL · AWAITING CAPTURE
CANVAS SNAPSHOT
HERO CANVAS · LIVE CAPTURE  
HERO CANVAS · AWAITING CAPTURE
CANVAS SNAPSHOT
// Live PNG snapshots from the running canvas renderer. Replace with production renders before launch.
SCREENSHOT · 001
SS-001
ASSET PENDING
SCREENSHOT · 002
SS-002
ASSET PENDING
SCREENSHOT · 003
SS-003
ASSET PENDING
// CAD DEMO footage placeholder — set SITE_CONFIG.CAD_DEMO_READY = true and drop footage at voxarn.com/demo/ to populate.
COMPANY ORIGINATOR UNCLASSIFIED
SUBJECT-ID: VXN-F01
Dev Bhatt
FOUNDER · VOXARN
VERIFIED
CALGARY, AB
CANADA
PRIOR VENTURE
Founder of ShapeSynth — a GPU-native quad retopology engine producing production meshes from raw 3D scans in 269ms end-to-end. Six-stage fully-GPU pipeline with novel sparse data structures for GPU cache hierarchy. Built on CUDA and inline PTX.
OPEN-SOURCE WORK
Builder of Axylith — a native research environment in C++20 with Vulkan 1.3 MTSDF rendering and full CI across GCC × Clang × sanitiser builds. Portfolio at devbhatt.dev runs a hand-written C++ mesh viewer compiled to WebAssembly with LIC field rendering and curvature heatmaps.
EDUCATION
CS + Data Science — dual major. Mount Royal University, Calgary.
DOMAIN EXPERTISE
Active engineer in mine and plant automation — the exact operational context where surface-native autonomy is most constrained by current technology. Employer: . VOXARN is an independent venture with no institutional backing.
FULL STACK
Python · Taichi · PyTorch · CuPy · NumPy · CUDA · ROCm / HIP · C++ · Vulkan 1.3 · GLSL · Emscripten / WASM · OpenGL ES3 · CAN Bus · Cloudflare · PCB Design (in development)
DEPLOY PARTNER INQUIRY RESTRICTED

We’re looking for
the right partners.

VOXARN is in early development and selectively engaging with investors, research institutions, and industrial and defense operators who understand what surface-native autonomy changes.

If you’re operating complex physical environments and current automation falls short — let’s talk about what VOXARN can build for you.

Investors / VC
Industrial
Defense
Research
PARTNER INQUIRY FORM
SECURE
✓ Inquiry received.
Reference ID: —
UNITS 08
MESH_FACES 24,816
PLAN_MS 11
GPU NOMINAL
UTC --:--:--
VXN-2026-001 · REV.A