[Volume 28. Neural Concept: AI-First 3D Deep Learning Platform Transforming Engineering Design vs Traditional CAD/CAE Tools]

Neural Concept: Real-Time Physics Simulation Platform with AI-Based Engineering Intelligence

Executive Overview

Founded: 2018 (EPFL spinoff)

Headquarters: Lausanne, Switzerland (EPFL Innovation Park)

Website: https://www.neuralconcept.com

Status: Private Company

Core Service: 3D Deep Learning-based Engineering Intelligence Platform

Total Funding: $38M across 3 rounds

• Seed: $2M (2020)

• Series A: $9.1M (March 2022, led by Alven)

• Series B: $27M (June 2024, led by Forestay Capital)

  
  

Annual Revenue:

• 2021: $1.4M

• 2023: $3.0M

• 2024: $3.6M

• 2025: ARR growth significantly outpaces burn rate, achieving capital efficiency

Leadership:

• Dr. Pierre Baqué (CEO & Co-founder) - EPFL PhD, Computer Vision & Deep Learning expert

• Théophile Allard (CTO & Co-founder)

• Dr. Jonathan Donier (VP Vertical Solutions)

• Thomas von Tschammer (General Manager USA & Co-founder)

  
  

Team Size: 65 employees (as of 2025)

Key Technology: 3D Deep Learning, Geometric Deep Learning, Physics-based Predictive AI

Major Clients: 70+ OEMs, 40% of top 100 Tier-1 suppliers

• Automotive: Subaru, Bosch, OPMobility, Renault, MAHLE

• Aerospace: Airbus, General Electric, GE Vernova

• Others: LG Electronics, 4 out of 10 Formula 1 teams

Industry Recognition:

• World Economic Forum Technology Pioneer 2025

• CB Insights AI 100

• NAFEMS Conference keynote speaker

Business Model: SaaS subscription + Professional services + Enterprise custom deployment

Neural Concept Inc., established in 2018 as an EPFL spinoff in Lausanne, Switzerland, provides an AI-First platform that revolutionizes engineering product development processes through 3D Deep Learning technology. The company has demonstrated proven results across automotive, aerospace, semiconductor, and consumer electronics industries: reducing product development time by up to 75%, accelerating simulation speed by 10x, and improving product performance (efficiency, safety, aerodynamics) by up to 30%.

What is Neural Concept? From Traditional CAD/CAE Tools to AI-First Engineering Platform

Traditional engineering design processes follow a sequential design-simulation workflow where designs are created in CAD (Computer-Aided Design) tools and then validated through CAE (Computer-Aided Engineering) simulations. This process takes hours to days to simulate a single design iteration, requires designers to depend on simulation engineers, and demands hundreds of iterations for design optimization.

Neural Concept replaces this with a real-time AI-based physics prediction platform. Using 3D Deep Learning models, it predicts physics simulation results directly from CAD geometry in approximately 30 milliseconds, enabling engineers to evaluate thousands of design variations within minutes. This represents a shift to "knowledge-driven" design processes.

Practical Example: Traditional approach: For automotive cooling system design, a single design iteration's CFD (Computational Fluid Dynamics) simulation takes 12-24 hours on HPC clusters. Engineers can only evaluate 20-30 design alternatives over several weeks.

Neural Concept approach: An AI model trained on historical simulation data predicts thermal performance of new designs in 30 milliseconds. Engineers can explore 10,000+ design variations in the same timeframe and identify optimal solutions in real-time. Subaru has significantly reduced development time for press forming predictions using this approach.

Market Leadership and Proven Track Record

Neural Concept is one of the fastest-growing companies in the engineering AI sector as of 2025, with enterprise customer base doubling since January 2025. Over 70 OEMs and 40% of the top 100 Tier-1 suppliers use the platform, with particularly strong market position in automotive (approximately 50% of customers).

Technology Approach: Neural Concept is based on Geometric Deep Learning technology developed at EPFL's world-class AI laboratory. This technology directly understands 3D geometric data and learns its interaction with physical laws. Unlike typical parametric models, it directly processes mesh-form original 3D data, enabling learning from any CAD/CAE tool data.

Core Technology Architecture: 3D Deep Learning Platform Neural Concept Shape (NCS) Platform Technology

Neural Concept's flagship product, Neural Concept Shape (NCS), is a 3D Deep Learning platform specifically designed for engineering, taking a fundamentally different approach from traditional CAD/CAE software.

1. 3D Deep Learning Architecture

The platform's core is a proprietary deep learning architecture that directly processes 3D geometric data:

Direct Shape Processing: Learns directly from mesh-form 3D data without parametric models or feature extraction Physics Learning: Neural networks learn relationships between CAE simulation results and CAD geometry Universality: Can use data from all CAD tools (CATIA, NX, SolidWorks, Fusion, etc.) and all CAE tools (Ansys, Abaqus, LS-DYNA, etc.) as input Real-time Prediction: Trained models predict physics simulation results for new designs in approximately 30 milliseconds

2. Proprietary Technology Capabilities

Neural Concept's technology has the following differentiating characteristics:

Non-parametric Learning: Learns directly from original 3D geometry without design parameterization Multi-physics Support: Supports structural, fluid, and thermal analysis Hybrid Models: Combines Physics-based Predictive Models with 3D Geometry Generative Models MLOps Integration: Complete MLOps pipeline for model training, deployment, and monitoring

3. Deployment and Integration Options

The platform supports various deployment methods:

Web Application: Deploy trained models as web-based interfaces for design teams to use directly in browsers CAD Integration: Direct integration within existing CAD tools (e.g., Siemens NX, CATIA) providing real-time physics feedback in design environment API Integration: Integration with enterprise PLM, CAD, CAE systems through RESTful APIs Private Cloud: Deployable as SaaS or in customer's private cloud

Neural Concept vs. Traditional CAD/CAE Tools

Limitations of Traditional CAD/CAE Workflow

Traditional engineering design processes have the following structural limitations:

1. Sequential Process: Designer creates design in CAD → Transfer to simulation engineer → Mesh generation → Run simulation → Analyze results → Feedback to designer This process takes hours to days per design iteration

2. Expert Dependency: Only specialized engineers can perform simulations Designers must wait for simulation results, making real-time feedback impossible

3. Limited Exploration Space: Time and cost constraints allow evaluation of only tens to hundreds of design alternatives Settle for "good enough" designs rather than finding optimal solutions

4. Underutilized Data: Tens of thousands of simulation data accumulated over years by companies go unused Must re-run identical types of simulations from scratch every time

Neural Concept's Innovative Approach

Neural Concept overcomes these traditional limitations as follows:

1. Real-time AI-based Design: Designers receive immediate physics feedback within CAD environment Predict structural, fluid, and thermal performance in 30 milliseconds Evaluate thousands of design variations within minutes

2. Design Democratization: Designers can independently optimize designs without simulation experts Simulation engineers focus on AI model training and validation New engineers can use platform after just 2 hours of training (vs. years of simulation expertise acquisition)

3. Large-scale Design Exploration: Traditional: Evaluate 20-50 design alternatives Neural Concept: Real-time exploration of 10,000+ design alternatives Can discover truly optimal designs

4. Knowledge Recycling: Utilize all historical simulation data for AI training Company's engineering knowledge becomes embedded in AI models Dramatic efficiency gains in repetitive design tasks

Comparison with Major CAD/CAE Vendors Neural Concept vs. Traditional CAD/CAE Tools

Feature	Neural Concept	Ansys	Autodesk (Fusion/Inventor)	Siemens NX	Dassault CATIA
Core Approach	AI-First Physics Prediction	High-precision Numerical Simulation	Integrated CAD/CAM	Integrated CAD/CAM/CAE	3D Experience Platform
Simulation Time per Design	30 milliseconds	Hours to days	Minutes to hours	Minutes to hours	Minutes to hours
Design Exploration Scale	10,000+ designs/day	10-50 designs/week	50-200 designs/week	50-200 designs/week	50-200 designs/week
Expertise Required	Minimal (2-hour training)	High (years)	Medium (months)	High (1-2 years)	High (1-2 years)
Direct 3D Shape Processing	✓ (Direct mesh learning)	✗ (Mesh generation required)	✗	✗	✗
Historical Data Utilization	✓ (AI training data)	✗	Partial	Partial	Partial
Real-time Design Feedback	✓	✗	Partial	Partial	Partial
CAD Integration	API/Plugin	✓	Native	Native	Native
Hardware Requirements	Cloud (NVIDIA GPU)	HPC cluster required	Workstation	Workstation	Workstation
Pricing Model	SaaS subscription	License (expensive)	Subscription	License (expensive)	License (expensive)
Primary Use Cases	Large-scale design optimization	High-precision validation	Product design	Manufacturing-integrated design	Complex systems design
Customer Base	70+ OEMs (startup to enterprise)	Enterprise-focused	SMB to enterprise	Enterprise-focused	Enterprise-focused
Employee Count	65	3,000+ (pre-Synopsys acquisition)	14,000+ (Autodesk total)	Part of Siemens	Part of Dassault
Founded	2018	1970	1982 (Autodesk)	1973 (Unigraphics)	1981

Detailed Comparative Analysis

1. Ansys vs. Neural Concept

Ansys Strengths:

• Highest level of simulation accuracy

• Support for all physics domains (structural, fluid, electromagnetic, thermal, etc.)

• Industry-standard validation tool

• Acquired by Synopsys in 2025, strengthening EDA integration

Ansys Weaknesses:

• Expensive license costs

• Complex interface with steep learning curve

• Hours to days per simulation

• HPC infrastructure required

• Limited AI integration

Neural Concept Differentiation:

• Partnership relationship with Ansys (Ansys Technology Partner)

• Utilizes Ansys simulation data for AI training

• Ansys used for high-precision validation, Neural Concept for initial design exploration and optimization

• Cost-effective: Reduces Ansys simulation count by 90%+

• Enables real-time design optimization

Market Positioning: Complementary relationship - Neural Concept doesn't replace Ansys; rather, it identifies optimal candidates through large-scale exploration in early design stages, then uses Ansys for final validation

2. Autodesk (Fusion, Inventor) vs. Neural Concept

Autodesk Strengths:

• Integrated CAD/CAM environment

• Cloud-based collaboration (Fusion)

• Accessible pricing for SMBs

• Generative design features (Fusion's Generative Design)

Autodesk Recent Developments:

• Announced "Neural CAD" at AU Conference September 2025

• Introducing generative AI-based models to Fusion and Forma

• Developing capability to create CAD objects from text prompts

• Attempting to innovate parametric CAD system unchanged for 40 years

Neural Concept Differentiation:

• Autodesk's "Neural CAD" focuses on shape generation; Neural Concept focuses on physics simulation

• Neural Concept specialized in engineering AI since 2018

• Proven physics prediction accuracy (used by 70+ OEMs)

• Multi-physics analysis support (structural, fluid, thermal)

• Operates independently of all CAD tools (CAD-agnostic)

Market Positioning: Complementary or competitive - Some overlap possible as Autodesk strengthens Neural CAD capabilities, but Neural Concept's engineering physics specialization and proven customer base provide differentiation

3. Siemens NX vs. Neural Concept

Siemens NX Strengths:

• Fully integrated CAD/CAM/CAE environment

• Strong manufacturing (L1-L5) integration

• Tight integration with PLM systems

• Digital twin capabilities

Siemens NX Weaknesses:

• High license costs

• Steep learning curve

• Suited for large organizations; high entry barrier for SMBs

Neural Concept Differentiation:

• Can utilize Siemens NX data as input

• AI-based real-time prediction reduces NX simulation time

• Faster deployment and lower entry barriers

• SaaS model accessible to SMBs

Market Positioning: Integration partner - Neural Concept can integrate as AI layer within NX workflow

4. Dassault CATIA/3DEXPERIENCE vs. Neural Concept

Dassault Strengths:

• Industry standard for aerospace and automotive

• Complex systems design support

• 3DEXPERIENCE platform collaboration features

Neural Concept Differentiation:

• Trains AI models with CATIA users' design data

• Provides real-time physics feedback within CATIA environment

• Cloud-native architecture

• Faster innovation pace (startup agility)

  
  

Market Positioning: Provides AI layer - Offers AI-based acceleration while maintaining existing CATIA investment

AI/ML Models and Technology Partnerships

Neural Concept's Proprietary AI Technology

Neural Concept does not use general-purpose AI models (GPT, Gemini, etc.), instead employing proprietary 3D Deep Learning models specifically developed for engineering physics.

Core Technology Foundation:

1. Geometric Deep Learning: Technology developed at EPFL's Computer Vision Laboratory that learns geometric characteristics of 3D shapes directly through neural network architecture. Unlike typical CNNs (Convolutional Neural Networks) used for image processing, it directly processes irregular 3D mesh data.

2. Physics-Informed Neural Networks: Integrates physical laws (fluid dynamics, structural mechanics, thermodynamics) into neural network training process to generate physically valid predictions rather than simple data fitting.

3. Proprietary Architecture: Neural Concept's neural networks use proprietary architecture specifically designed for engineering applications by the company's research team. This is completely different from publicly available general-purpose models (GPT, BERT, etc.).

   
   

Major Technology Partnerships

1. NVIDIA Partnership (2024-2025)

Most significant strategic partnership:

NVIDIA Omniverse Blueprint Integration (announced March 2025):

• Neural Concept platform fully integrated with NVIDIA Omniverse Blueprint for Real-Time Digital Twins

• Runs optimized 3D physics-based deep learning models on NVIDIA GPUs (H100, A100)

• Utilizes NVIDIA CUDA software

• Joint demo at GTC 2025 conference

Technical Benefits:

• GPU-accelerated large-scale multi-physics simulations

• Simultaneous evaluation of thousands of design variations in real-time digital twin environment

• Reduced deep learning model training time (utilizing NVIDIA GPU clusters)

Joint Case Studies:

• General Motors: Neural Concept's pedestrian crash safety prediction model trained on NVIDIA GPUs predicts impact physics in seconds rather than days

• Joint presentation at GTC 2025: "Revolutionize Pedestrian Safety: AI-Powered Crash Worthiness"

• Ansys Technology Partnership (since 2022)

Neural Concept is Ansys Technology Partner Program member:

• Utilizes Ansys simulation data for Neural Concept AI training

• Integration with Ansys workflow

• Complementary relationship: Ansys for high-precision validation, Neural Concept for large-scale initial exploration

• Rescale Partnership (announced December 2024)

HPC-as-a-Service partnership:

• Neural Concept AI model training on Rescale's cloud HPC infrastructure

• Processes large training datasets with elastic computing power

• Customers can instantly scale HPC resources as needed

• Integration with Major CAD Vendors:

Neural Concept integrates with:

• Siemens NX

• Dassault CATIA

• Autodesk Fusion

• SolidWorks

• All other major CAD tools (CAD-agnostic approach)

Relationship with GPT, Gemini, and Other General-Purpose LLMs

Neural Concept does not use general-purpose Large Language Models (LLMs) like GPT (OpenAI), Gemini (Google), or Claude (Anthropic).

Reasons:

1. Domain Specialization: Engineering physics simulation requires understanding 3D geometric data and physical laws, impossible with text-based LLMs

2. Input Data Types: Neural Concept receives mesh-form 3D CAD data and numerical simulation results as input

3. Output Types: Requires physical numerical predictions (stress, pressure, temperature distribution, etc.), not text generation

4. Accuracy Requirements: Engineering safety and performance prediction demands physical validity and high numerical accuracy

Instead, Neural Concept:

• Uses self-developed 3D Deep Learning models

• Leverages Geometric Deep Learning and Graph Neural Networks

• Trains with Physics-based Loss Functions

• Fine-tunes with customer's domain-specific simulation data

Business Model and Revenue Structure

Multi-layered Revenue Architecture

Neural Concept operates a SaaS-based platform business model:

1. Platform Subscription:

   - Access to Neural Concept Shape (NCS) platform

   - Cloud-based SaaS or private cloud deployment

   - Tiered pricing based on company size, number of users, project scope

   - Annual or multi-year contracts

2. Professional Services:

   - AI model training and customization

   - Integration with existing CAD/CAE/PLM systems

   - Data preparation and preprocessing support

   - Engineering team training

3. Customer Success and Support:

   - Continuous model optimization

   - Technical support

   - Support for expanding to new application areas

4. Upselling and Expansion:

   - Start with initial pilot projects

   - Expand to entire departments after validation

   - Eventually deploy enterprise-wide

   - 2025: 40% increase in upsells (expanding from custom applications to company-wide use)

Target Market Segments

Neural Concept targets the entire engineering product development ecosystem:

1. Automotive OEMs and Tier-1 Suppliers (approximately 50% of customers):

   - Automotive manufacturers: Subaru, Renault

   - Parts suppliers: Bosch, OPMobility (Plastic Omnium), MAHLE, SPAL Automotive, Kautex

   - Formula 1 teams: 4 out of 10 teams (including Williams F1, Visa Cash App RB F1)

2. Aerospace:

   - OEMs: Airbus, Leonardo

   - Engine manufacturing: General Electric, GE Vernova

3. Electronics and Semiconductors:

   - LG Electronics

   - Semiconductor manufacturing equipment (expanding)

4. Defense:

   - Leonardo

   - Hanwha Ocean (Korean shipbuilding/defense)

5. Energy:

   - GE Vernova (power generation turbines)

6. Other Industries:

   - Sumitomo Wiring Systems (automotive wiring)

   - Multi-Wing (industrial fans)

Competitive Advantages

1. Proven Customer Base:

   - 70+ global OEMs

   - 40% of top 100 Tier-1 suppliers

   - 40% of largest European and Asian OEMs

2. Technology Leadership:

   - Specialized in 3D Deep Learning since 2018 (7-year lead)

   - World-class AI lab alumni from EPFL

   - 60+ completed industrial projects

   - 2 patent filings

3. Proven Results:

   - Up to 75% reduction in product development time

   - 10x acceleration in simulation speed

   - Up to 30% improvement in product performance

   - MAHLE: 15% efficiency improvement, 4dB noise reduction

4. Rapid Deployment:

   - Usable after 2-hour training

   - Integration with existing CAD/CAE systems

   - Minimal infrastructure investment with cloud deployment

5. Capital Efficiency:

   - 2025: ARR growth rate significantly exceeds burn rate

   - Among most capital-efficient companies in the industry

Financial Performance and Market Position

Recent Performance Metrics

2025 Highlights (January-September):

• Enterprise customer base doubled (2x vs. January)

• 40% increase in upsells (custom applications → company-wide use)

• Team size: 65 employees (increased from 60 in 2024)

• Geographic expansion: Switzerland, France, Germany, North America, Japan, South Korea, India

Revenue Trajectory:

• 2021: $1.4M

• 2023: $3.0M

• 2024: $3.6M

• 2025: Achieving capital efficiency with ARR growth significantly exceeding burn rate

Capital Position:

• Total funding: $38M (3 rounds)

• Seed: $2M (2020)

• Series A: $9.1M (March 2022, led by Alven)

• Series B: $27M (June 2024, led by Forestay Capital)

• Major investors: Forestay Capital, D. E. Shaw Group, Alven, CNB Capital, HTGF, Aster Capital

Market Position Analysis

Market Standing:

• Selected as World Economic Forum Technology Pioneer 2025

• CB Insights ESP (Engineering Simulation Software) Matrix: Rated as Challenger (among 15 companies including Ansys, Autodesk, Siemens)

• Fast Company Most Innovative Companies consideration

Customer Base:

• 70+ OEMs

• 40% of top 100 Tier-1 suppliers

• 40% of largest European and Asian OEMs

• Established market leadership in automotive industry (approximately 50% of customers)

Geographic Distribution:

• Headquarters: Lausanne, Switzerland (EPFL Innovation Park)

• Operations: Switzerland, Germany, France, USA, Japan, South Korea, India

• Major markets: Europe, North America, Asia-Pacific

Competitive Analysis and Market Position

Market Landscape (2025)

Engineering AI/Simulation Software Market:

• Total Addressable Market (TAM): Estimated $10-20B globally

• Traditional CAE: $5-8B (Ansys, Siemens, Dassault)

• AI-based Engineering: $1-3B (emerging market, rapid growth)

Neural Concept Competitive Advantages:

• Specialization in 3D Deep Learning (7 years since 2018)

• 70+ OEM proven customer base

• Physics-based AI prediction accuracy

• CAD-agnostic platform (integrates with all CAD/CAE tools)

Major Competitor Comparison

Direct Competitors

1. PhysicsX (UK):

   - Similar approach: AI-based physics simulation

   - Difference: Neural Concept specializes in 3D geometry, PhysicsX broader scope

   - Market position: Competitive but market large enough

2. Monolith AI (UK):

   - Focus: Materials science and structural simulation

   - Difference: Neural Concept stronger in geometry optimization

3. Key Ward (Germany):

   - Focus: Engineering data intelligence

   - Difference: Focuses on data management, Neural Concept on physics prediction

4. Intelecy (Norway):

   - Focus: Industrial process optimization

   - Difference: Manufacturing processes vs. product design

Traditional CAE Vendors' AI Entry

1. Ansys (now Synopsys subsidiary):

   - Adding AI features but traditional simulation remains core

   - Partnership relationship with Neural Concept (technology partner)

   - Complementary positioning

2. Autodesk:

   - Developing "Neural CAD" (announced 2025)

   - Focuses on shape generation vs. Neural Concept's physics prediction

   - Possible long-term overlap but currently different domains

3. Siemens:

   - PLM/Digital Twin focused

   - Limited AI integration

   - Potential integration partner with Neural Concept

4. Dassault Systèmes:

   - Adding AI to 3DEXPERIENCE platform

   - Enterprise-focused vs. Neural Concept's broader market

Competitive Positioning Strategy

Neural Concept Differentiation:

1. Specialized AI:

   - Specialized 3D Deep Learning for engineering physics

   - Domain expert, not general-purpose AI platform

2. Proven Results:

   - Used in actual product development by 70+ OEMs

   - Specific performance metrics (75% time reduction, 10x speed improvement)

3. Platform Strategy:

   - Augments existing CAD/CAE tools with AI layer rather than replacing

   - "Intelligence layer on top of existing digitalization layer (PLM, CAD, CAE)"

4. Fast Innovation:

   - Startup agility

   - Rapid development and deployment with 65-person team

Customer Base and Market Penetration

Customer Overview

Neural Concept serves over 70 global OEMs and 40% of the top 100 Tier-1 suppliers.

Customer Distribution:

• Automotive: approximately 50%

• Aerospace: approximately 20%

• Electronics/Semiconductors: approximately 15%

• Defense: approximately 10%

• Energy and others: approximately 5%

Major Customers

Automotive OEMs and Tier-1:

• Subaru: Reduced development time and improved quality with press forming prediction

  • "In the past, mold design optimization required CAD and FEM. However, Neural Concept Shape can create surrogate AI learned from past FEM and CAD shapes, and predict new design shapes at high speed" - Sakata Suke, Body Production Engineer

• Renault: Expanded platform usage in 2025

• Bosch: Global automotive parts supplier

• OPMobility (Plastic Omnium): Demonstrated Neural Concept technology for PHEV fuel tank design at CES 2025

  • Reduced fuel sloshing simulation in PHEV fuel tanks from 12 hours → minutes

• MAHLE: Developed AI-optimized bionic blower

  • 15% efficiency improvement, 4dB noise reduction

  • Ideal for EV platforms

• SPAL Automotive: Italian automotive cooling system supplier

• Kautex: Automotive fuel system supplier

• Sumitomo Wiring Systems: Japanese automotive wiring harness

Aerospace:

• Airbus: Europe's largest aircraft manufacturer

• General Electric / GE Vernova:

  • Aviation engines and power generation turbines

  • Joint presentation with Neural Concept at NVIDIA GTC 2025

• Leonardo: Italian defense and aerospace company

Defense:

• Hanwha Ocean: Korean shipbuilding and defense industry

Electronics:

• LG Electronics: Korean global electronics manufacturer

Motorsports:

• Formula 1 Teams: 4 out of 10 teams

  • Williams F1 Team (featured in TechCrunch article)

  • Visa Cash App RB Formula One Team

Others:

• Multi-Wing: Industrial fan manufacturer

• SP80: World record challenge boat (80 knots, 148km/h target)

  • Solved cavitation problem with wedge-shaped hydrofoil designed by Neural Concept

  • Digital twin simulation demo at NVIDIA GTC 2025

Customer Success Metrics

2024-2025 Growth:

• 2023: 60 customers

• September 2025: 70+ OEMs, enterprise customers doubled (vs. January)

• Upsells: 40% increase (custom applications → enterprise deployment)

Geographic Reach:

• Europe: Strong presence (headquarters in Switzerland)

• North America: Expanding (US office)

• Asia-Pacific: Growing in Japan, South Korea, India

Conclusion

Key Technology Characteristics

AI-First Engineering Platform:

• Proprietary 3D Deep Learning technology

• Real-time physics simulation prediction (30 milliseconds)

• CAD-agnostic: Integrates with all CAD/CAE tools

• Cloud-native SaaS architecture

Market Position:

• Leading company in engineering AI sector

• 70+ OEM customers

• World Economic Forum Technology Pioneer 2025

• CB Insights Challenger (ESP Matrix)

Comprehensive Assessment

Neural Concept is driving a paradigm shift in engineering product development. While traditional CAD/CAE tools require sequential and time-consuming design-simulation iterations, Neural Concept enables real-time physics feedback through 3D Deep Learning, allowing engineers to explore thousands of designs within minutes.

The technical foundation originating from EPFL's world-class AI laboratory and proven real-world results across 70+ global OEMs demonstrate the technology's maturity. Strong market position establishment in automotive and rapid expansion into aerospace, semiconductors, and defense industries showcase the technology's versatility and business model scalability.

The strategic partnership with NVIDIA further strengthens technical advantages, while the collaborative relationship with Ansys clarifies complementary positioning with traditional CAE tools. Although Autodesk is entering the market with "Neural CAD," Neural Concept's 7 years of engineering physics specialization and proven customer base provide sustainable competitive advantages.

Implications for Industry Observers and Potential Customers

Neural Concept demonstrates sustainable market leadership through:

Technical Differentiation:

• Specialized 3D Deep Learning for engineering physics

• Proprietary Geometric Deep Learning, not general-purpose AI models (GPT, Gemini)

• Neural network architecture with embedded physical laws

Proven Business Model:

• Used in actual product development by 70+ OEMs

• 75% product development time reduction, 10x simulation acceleration

• 2025: Enterprise customers doubled, 40% upsell increase

Strategic Partnerships:

• NVIDIA: Cutting-edge GPU technology and digital twin integration

• Ansys: Complementary with traditional CAE

• Rescale: Elastic HPC infrastructure

Capital Efficiency:

• ARR growth rate significantly exceeds burn rate

• 70+ OEM customers acquired with $38M investment

• Among most capital-efficient companies in the industry

Market Scalability:

• Expanding from automotive to aerospace, semiconductors, defense

• Geographic expansion from Europe to North America, Asia

• Serving entire market from SMB to enterprise

In competition with traditional CAE vendors (Ansys, Siemens, Autodesk), Neural Concept positions itself not as a replacement but as an AI augmentation layer that provides innovative value while leveraging existing investments. Despite Autodesk's Neural CAD entry, Neural Concept's physics simulation expertise and 7-year leadership provide solid competitive advantages.

For engineering organizations aiming to dramatically improve product development speed, place AI at the center of design processes, and bring sustainable and innovative products to market faster, Neural Concept is the optimal partner offering proven technology and demonstrated results.

© 2025 All rights reserved.