Slide 1: NVIDIA
GI NVIDIA
Slide 2: European ecosystem partner logos
4X Ge ocrsi ABB accenture EP AcILEROBOTS BAkilo (hed) Ansys aws AnKoW astrazenecd 2 7SuS (or1D) AVEVA paesystems (© — waarciays 7 beyond.pl BMW. BNPPARIBAS BOSCH | EEE.) —-BTGroup AJ Business Systems Capgemini T.CHARM suai @ 36 passaur Neuse CINECA ‘cisco’ D —_cmiacner cores2 Gntinentals CY CoreWeav
Slide 3: Aerial, Earth-2, MONAI, Parabricks and More
Aerial Earth-2 MONAI —Parabricks cuQuantum cuPyNumeric Dynamo —cuEquivariance CUDA-Q cuDNN cuTensor Sionna cuLitho ‘AmgX
Slide 4: CUDA-Q Quantum-GPU Classical Computing Across Europe
NVIDIA CUDA-Q Quantum-GPU Classical Computing Across Europe QPU GB200 NVL72
Slide 5: CUDA-Q Quantum-GPU Classical Computing
NVIDIA CUDA-Q Quantum-GPU Classical Computing 4,000X 1,300X Speedup 1,200X vs. CPU 800X 500X QPU cuQuantum cuQuantum cuQuantum CUDA-Q CUDA-Q Quantum Algorithm Qubit Design Quantum Data Hybrid Quantum Error Development EDA Generation Applications Correction GB200 NVL72
Slide 6: First Contact
“FIRST CONTACT” riSzesit a 2012 ALEXNET PERCEPTION Al SPEECH RECOGNITION DEEP RECSYS MEDICAL IMAGING PHYSICAL AI AUTONOMOUS VEHICLES GENERAL ROBOTICS AGENTIC Al CODING ASSISTANT CUSTOMER SERVICE PATIENT CARE GENERATIVE Al DIGITAL MARKETING. _— CONTENT CREATION
Slide 7: Slide 7
Slide 8: Grace Blackwell NVL72
GB300 NVL72 Grace Blackwell NVL72 “A Thinking Machine”
Slide 9: Grace Blackwell NVL72
Grace Blackwell NVL72 “One Giant GPU”
Slide 10: Blackwell Massive Leap in Reasoning Inference Performance
NVIDIA Blackwell Massive Leap in Reasoning Inference Performance EP64PP2+PP2, Batch 832, 47% Context 1,200,000 EP64PP2+PP2, Batch 416, 46% Context 1,100,000 EP64+PP2, Batch 128, 45% Context EP64PP4+PP2, Batch 352, 43% Context 1,000,000 EP64+PP2, Batch 64, 42% Context EP64PP2+PP2, Batch 96, 40% Conte
Slide 11: SSeS oes
SSeS oes peeeee
Slide 12: Grace Blackwell in Full Production
Grace Blackwell in Full Production
Slide 13: One Architecture
One Architecture – From Cloud AI, Enterprise AI, Personal AI, to Edge AI Spectrum-X RTX Storage Switch DGX Servers B200 / B300 GB200 / GB300 RTX PRO NVL72 RTX PRO Server DGX RTX PRO Workstation DGX Station Spark Laptop
Slide 14: One Architecture
One Architecture – From Cloud AI, Enterprise AI, Personal AI, to Edge AI Spectrum-X RTX Storage Switch DGX Servers B200 / B300 GB200 / GB300 RTX PRO RTX PRO Server NEW DGX NVL72 RTX PRO Workstation DGX Station Laptop NEW Spark NEW NEW NEW NEW
Slide 15: NVIDIA RTX PRO Server
NVIDIA RTX PRO Server Enterprise AI and Industrial AI Computing Platform Available Now
Slide 16: A New Industrial Revolution
A NEW INDUSTRIAL REVOLUTION AI FACTORY $100T se SCIENCE ENTERPRISE = HEALTHCARE MANUFACTURING TOKENS “UNITS OF INTELLIGENCE” Sooo0oo000000 ——<_—»>
Slide 17: Stargate
Stargate 200 MW, Scaling to 1.2 GW Abilene, Texas Courtesy of Crusoe
Slide 18: European Telcos Building AI Infrastructure With NVIDIA
European Telcos Building AI Infrastructure With NVIDIA
Slide 19: European Cloud Service Providers Building AI Infrastructure
European Cloud Service Providers Building AI Infrastructure With NVIDIA
Slide 20: European Supercomputing Centers Building AI Infrastructure
European Supercomputing Centers Building AI Infrastructure With NVIDIA
Slide 21: European Union Building AI Factories
European Union Building 20 AI Factories Barcelona Supercomputing Center (Spain) CINECA (Italy) CSC (Finland) GENCI/CEA (France) GRNET (Greece) Jülich Supercomputing Center (Germany) Linköping University (Sweeden) LuxProvide (Luxembourg) Poznan Supercomputing Center (Poland) Sofia Tech Park (Bulgaria
Slide 22: 10X Growth Projection
10X Growth Projection Pan-European AI Infrastructure 3X+ 2024 2025 2026 GPU Die AI FLOPS – Ampere FP16, Ada/Hopper FP8, Blackwell FP4
Slide 23: NVIDIA Establishes European AI Technology Centers
FINLAND SWEDEN NVIDIA Establishes European AI Technology Centers to Grow Regional Research and UNITED Startup Ecosystem KINGDOM GERMANY FRANCE ITALY SPAIN
Slide 24: NVIDIA AI Infrastructure in United Kingdom
Startups Upskillers Developers Enterprise Researchers GPU Cloud Providers NVIDIA AI Infrastructure in United Kingdom CSP ISVs Compute, Storage, Networking AIOps GSI
Slide 25: NVIDIA AI Infrastructure in Germany
Startups Upskillers Developers Enterprise Researchers GPU Cloud Providers NVIDIA AI Infrastructure in Germany CSP ISVs Compute, Storage, Networking AIOps GSI
Slide 26: NVIDIA AI Infrastructure in Italy
Startups Upskillers Developers Enterprise Researchers GPU Cloud Providers NVIDIA AI Infrastructure in Italy CSP ISVs Compute, Storage, Networking AIOps GSI
Slide 27: NVIDIA AI Infrastructure in France
Startups Upskillers Developers Enterprise Researchers GPU Cloud Providers NVIDIA AI Infrastructure in France CSP ISVs Compute, Storage, Networking AIOps GSI
Slide 28: Planning and Concept Reference Designs
Planning | Concept 1st Fully Integrated Reference Designs AI Factory Electrical Digital Next Gen Twin Simulation CC DC Architectures EcoStruxure IT Design CFD Build
Slide 29: Announcing Mistral and NVIDIA Build AI Cloud
Announcing Mistral and NVIDIA Build AI Cloud
Slide 30: Open-Source AI Models Are Unlocking Global Innovation
Open-Source AI Models Are Unlocking Global AI Innovation 100 DeepSeek R1 0528 90 GPT-4o Gemini 2.5 Pro Llama 3 405B GPT-4 Gemini 1.5 Pro Mistral Large 2 DeepSeek R1 671B MMLU Qwen 3 235B Llama 3 70B Llama Nemotron Ultra 80 Mixtral 8x22B Mixtral 8X7B 70 GPT-3.5 Llama 2 Llama 60 2022 2023 2024 2025
Slide 31: NVIDIA Nemotron Further Advances Leading Open Models
NVIDIA Nemotron Further Advances Leading Open Models Open Frontier NVIDIA Nemotron Model
Slide 32: NVIDIA Nemotron Increases Open Model Efficiency and Accuracy
NVIDIA Nemotron Increases Open Model Efficiency and Accuracy *Nemotron Llama Nemotron Ultra 253B Llama Nemotron +21 Super 49B Llama Nemotron Nano 4B +16 +21 40 35 24 Llama 8B Llama 70B Llama 405B Next Open Model Artificial Analysis Intelligence Index MMLU-Pro, GPQA Diamond, Humanity's Last Exam, Liv
Slide 33: NVIDIA Nemotron Achieves World-Class Benchmarks
NVIDIA Nemotron Achieves World-Class Benchmarks #1 #1 Models #1 Llama Nemotron Ultra Leading GPQA-Diamond Accuracy Datasets Agents Llama Nemotron Nano Llama Nemotron Post-Training Math 500 Nemotron-CORTEXA Trending Reasoning Dataset Leading SWE-Bench Accuracy Llama Nemotron Vision OpenCode OCRBench
Slide 34: Model Builders Across Europe Creating Sovereign LLMs
Model Builders Across Europe Creating Sovereign LLMs with NVIDIA Nemotron
Slide 35: Perplexity Makes European Nemotron Models Available in Region
Perplexity Makes European Nemotron Models Available in Region
Slide 36: Cae
Cae THINKING a TOOL USE i COMPUTER USE OTHER AGENTS €@e MEMORY =
Slide 37: NVIDIA Enterprise AI Agent Platform
NVIDIA Enterprise AI Agent Platform Specialized AI Agents NeMo Agentic AI Toolkit AI-Q Agentic AI Blueprint Nemotron Reasoning LLM NeMo Retriever RAG
Slide 38: AI Defense
Al Defense Training Al Defense Safety Models Eval Dataset Safety Models Cisco Safety NVIDIANIM NVIDIA NeMo Framework | Al Threat Intel Data Cisco Security NVIDIA NIM NVIDIADGX Cisco Privacy NVIDIANIM Al Defense Runtime Protection i User Al Defense Proxy NVIDIA NeMo Guardrails Model Assessment. Needs
Slide 39: Connecting Developers to Global AI Compute
Connecting Developers to Global AI Compute Multi-Cloud | Multi-Tenant | Scalable Training | Global Inference Global Cloud Providers Developer Local Environment GPU Cloud Providers GPU Marketplaces DGX Cloud Lepton Andromeda The San Francisco Compute Company Developer Local Environment DGX Spark BYO
Slide 40: Connecting Developers to Global AI Compute
Connecting Developers to Global AI Compute Multi-Cloud | Multi-Tenant | Scalable Training | Global Inference Global Cloud Providers Developer Local Environment GPU Cloud Providers GPU Marketplaces DGX Cloud Lepton Andromeda The San Francisco Compute Company Developer Local Environment DGX Spark BYO
Slide 41: Slide 41
Slide 42: SAP Applications
SAP Applications Joule = Customer Partner A Cloud ERP RM pend xm Pps Al Foundation — Al Operating System Development Tooling Generative Al Hub Serving/inference Engine Trust, Orchestration © © © Generative Al Hermonized APL Hub in SAP Al Data Masking NVIDIA NeMo NVIDIA NIM Mistral With Launchpad Pro
Slide 43: Sana agent architecture
User Sana Main Agent Llama Nemotron Plan Reflect, valuate Uses Documents Ly Indexing NVIDIA NeMo Retriever Extraction NVIDIA NeMo Retriever Embedding NVIDIA NeMo Retriever Reranking Vector Search Internet Search Mcp ‘Third Parties
Slide 44: Training and Inference
Training Inference NVIDIA Megatron Core NVIDIA Transformer Engine NVIDIA TensorRT-LLM. LoRA Inflight Batching Optimized KV Cache Pytorch Training Code Pe Weight ‘naSeale & raster LLM Model sD Parallelism Distillation EPs RLHF Kubernetes $ LLM Model Workflow Orchestration
Slide 45: Photoroom inference pipeline
Kubernetes Cluster ® Photoroom Triton Ensemble Models Data Dependent Acyclic Computation Control Flow © PyTorch — Torchscript NVIDIA TensorRT Triton Ensemble Models Data Dependent Acyclic Computation Control Flow © PyTorch — TorchScrint NVIDIA Tensor — =— i Server Type 1 8, £4, Pruning — = i Server
Slide 46: Retrieval and Generation Pipeline
Retrieval & Generation Pipeline Ingest Pipeline @ qodo Query Coding Agent Lama Nemotron Tool Usage Reasoning NVIDIA NIM Code Search & Retrieval Agent Uama Nemotron NVIDIA NeMo Retriever NVIDIA NeMo Retriever Reasoning Reranking Embedding Codebase & Software-Related Content Enterprise Software Files
Slide 47: Triton Ensemble
Triton Ensemble Preprocess Audio VAT ‘Audio Service chunk Unstructured to Structured Spoken Data NVIDIA NeMo_ Retriever LLM NIM | . structured ‘Output Context Manager Enterprise system integration CRM, Notifications, SoM
Slide 48: Clinical and Medical
& Clinical & Medical Ba Commercial & Chatbot Nemo Guardrails Orchestrator Agents Data Review Agent Literature Review Agent Field Companion Agent Data Agents Data Tools Fe Entity Agent Test Pan Agent Pcs 2 : Validation Agent Rx Agent Data iegrtion fan} EE of & cs Savsm Reto uM Uama NVIDIA NeMo NeMo c
Slide 49: Slide 49
Slide 50: Product Design and Engineering
Product Design Secure Edge and Engineering Industrial AI Operational Planning Production and Optimization Optimization Comprehensive Product and Factory Digital Twin
Slide 51: Slide 51
Slide 52: Omniverse Digital Twin
aoe /) = 3 ea Omniverse Digital Twin a
Slide 53: Omniverse Digital Twin
Omniverse Digital Twin
Slide 54: Omniverse Digital Twin
Omniverse Digital Twin
Slide 55: Omniverse Digital Twin
Omniverse Digital Twin
Slide 56: Omniverse Digital Twin
Omniverse Digital Twin
Slide 57: Omniverse Digital Twin
Omniverse Digital Twin |
Slide 58: Omniverse Digital Twin
Omniverse Digital Twin
Slide 59: Omniverse Digital Twin
Omniverse Digital Twin
Slide 60: Announcing NVIDIA Builds World's First Industrial AI Cloud
Announcing NVIDIA Builds World’s First Industrial AI Cloud Siemens Maserati, Siemens BMW Schaeffler Mercedes-Benz CAD CAE Factory Digital Twin Robotics AV RTX CUDA-X Omniverse Isaac and Omniverse DRIVE AV and Omniverse
Slide 61: NVIDIA Serving Europe's Automotive Industry
NVIDIA Serving $2 Trillion Europe Automotive Industry for 15 Years Siemens Maserati, Siemens BMW Schaeffler Mercedes-Benz CAD CAE Factory Digital Twin Robotics AV RTX CUDA-X Omniverse Isaac and Omniverse DRIVE AV and Omniverse
Slide 62: NVIDIA DRIVE
NVIDIA DRIVE Autonomous Vehicle Platform Sensor Data Motion Plan Navigation NVIDIA DRIVE DRIVE AGX Thor
Slide 63: NVIDIA Wins CVPR End-to-End AV Grand Challenge
Announcing NVIDIA Wins CVPR End-to-End AV Grand Challenge 2024 2025
Slide 64: Slide 64
Slide 65: NVIDIA Halos Recognized by Global Safety Ecosystem
Announcing NVIDIA Halos Recognized by Global Safety Ecosystem Comprehensive Chip-to-Deployment Safety System
Slide 66: NVIDIA Isaac
NVIDIA Isaac Open Robotics Development Platform Sensor Input Action Prompt NVIDIA Isaac Jetson Thor
Slide 67: Humanoid robotics partners
apolar AZ REE gt SANE] engen Huang === Menteedot == DualArm = HDCT. NED igure
Slide 68: Deep Research
‘Dienep Research GYNVIDIA J Google DeepMind Newton Fe |< Fe MuJoCo Disney Research Custom Solver Solver Solver MuJJoCo Playground a ro | Other Simulator NVIDIA Warp Isaac Lab
Slide 69: A New Industrial Revolution
A New Industrial Revolution A NEW WAVE OF AI BLACKWELL AI FACTORIES – 10X EUROPEAN AGENTS AND ROBOTS “A THINKING MACHINE” THE NEW INFRASTRUCTURE INFRASTRUCTURE NEXT 2 YEARS
Slide 70: NVIDIA
GI NVIDIA