ROLLING & SWIVELING

MATERIALS

FLUIDS

LINKED MECHANISMS

ENVIRONMENTS

TOOLS

COOKWARE

INDUSTRIAL ROBOTICS

HINGED ASSEMBLIES

ANTIM LABS / SIMULATION INFRASTRUCTURE

The world layer
for Physical AI

AI-generated 3D assets and environments for robot training and evaluation.

ANTIM LABSRAISING $2M / SAFE

SCROLL TO ENTER ↓

WORLD AGENT

3D assets
and environments

01 / TRACTION

Customers & pipeline

Since July 2026

ONGOING & COMPLETED

Frontier robotics company

Pilot underway · Recurring engagement

Initial project: six figures

Company confidential

Strata Robotics

Recurring engagement · Construction simulations

DestroAI

Six-figure LOI signed · Pilot in scoping

Dimensional OS

Completed project · Five-figure value

SCOPING PILOTS

Maven Robotics

Biggest American Agricultural & Industrial OEM — Autonomy Team

Confidential

Frontier robotics simulation research team

Confidential

Tactile robotics company

Real-to-sim training and evaluation for dexterous manipulation using egocentric factory data.

IN PIPELINE

Tuesday Robotics

Pickle Robotics

02 / WHY SIMULATION

Train and test beyond
the limits of the field

Evaluating every condition and software update in the field is impossible. Simulation makes those tests repeatable and scalable.

Capture the real world

Use field data to recreate operating conditions.

Train & evaluate in sim

Vary conditions and compare builds at scale.

Deploy & capture failures

Bring new edge cases back into simulation.

FIELD FAILURES FEED THE NEXT TRAINING & EVALUATION CYCLE

03 / ASSETS & ARTICULATIONTEXT / IMAGES → 3D / ARTICULATION

3D assets built to interact

Every asset shown was created by our world agent,
compressing weeks of manual work into minutes.

Data center rackExploded assembly

Real garden shears reference photograph — Garden shearsReal → simulation

Lysol wipes real product photograph — Lysol wipesReal → simulation

Real screwdriver reference photograph — ScrewdriverReal → simulation

DishwasherDoor · racks · spray arms

KitchenAid toasterLift lever motion

Adjustable desk lampBase · shoulder · elbow · shade

Sharpie S-Gel real product photograph — Sharpie S-GelReal → simulation

Carrot cuttingMaterial interaction

Pillow deformationDeformable material

Real-to-sim assets, articulation, and material interactions.

01—04 / 10

04 / REAL-WORLD INTERACTION DATA

Real-to-sim
comparisons

Real and simulated grasping, deformation, and squeezing.

REAL CAPTURE / ABOVEISAAC SIM / BELOW

Grasp & deformation / Plastic cup

REAL CAPTURE / ABOVEISAAC SIM / BELOW

Squeeze & dispense / Bottle

ACTIVE SYSTEM IDENTIFICATION

We measure real interactions to estimate physical properties.

05 / ENVIRONMENTS

Simulation
environments

From factory floors to open fields, simulate the worlds where robots operate.

INFRASTRUCTURE

Data centers

Test connector-level interactions.

LOGISTICS

Warehouses

Simulate warehouse operations.

INDUSTRY

Manufacturing

Test robots on the factory floor.

AERIAL ROBOTICS

Drone simulation

Simulate flights through a neighborhood.

AUTONOMY

Urban environments

Test the same city in different light.

FIELD ROBOTICS

Construction

Simulate solar-field installation.

MANIPULATION

Household manipulation

Test manipulation in a kitchen.

01 / 07

Recorded Antim environments

06 / EVALUATION PLATFORM · IN DEVELOPMENT

Automated
evaluation

Turn field failures into repeatable simulation tests.

Generate and vary scenarios.
Compare software under the same conditions.
Keep failures as regression tests.

Hosted inference

01Layout variation

02Lighting & weather

03Perception variation

04Task variation

Scenario-generation demos for the evaluation platform in development.

07 / PRODUCTS & SERVICES

Products & services

01 / REAL-WORLD DEPLOYMENTS

Custom simulation

Simulations built around your robot, task, and deployment environment.

TEXT + IMAGES → SIM-READY ASSETS

02 / FRONTIER ROBOTICS

Generation API

Generate simulation-ready assets and scenes from text and images.

Antim evaluation prototype showing parallel warehouse scenarios and per-scenario results — PROTOTYPE INTERFACE / DEMO RESULTS

03 / IN DEVELOPMENT

Evaluation platform

Automated evaluations on Newton, with hosted inference and regression testing.

08 / TEAM5 FULL-TIME / 2 PART-TIME

Founding team

AI, 3D, and physics experience. Co-founders met at Columbia.

Shreyans Kothari

Co-founder / CEO

Applied Statistics, Columbia

Built machine-learning and animation systems for real-time 3D digital humans.

Gokul Srinivasan

Co-founder / President

Electrical Engineering, Columbia

Physics simulation research and signal-processing experience.

Viswajit Nair

Co-founder / CTO

Computer Science, Columbia

Built applied AI systems and published machine-learning research.

09 / THE RAISE

Raising $2M

$25M cap on a SAFE

RAISED TO DATE

$1.3MActive Capital · Untapped VC

ANTIM LABS / 4WALL AI INC.CONFIDENTIAL / OCTOBER 2026

The world layer for Physical AI