#1NVIDIA GTC FALL 2020 FINANCIAL ANALYST EVENT#2Forward Looking Statements Except for the historical information contained herein, certain matters in this presentation including, but not limited to, statements as to: our expanding ecosystem and the estimated number of Al and autonomous systems developers adopting NVIDIA Al; artificial intelligence as the most powerful technology force of our time; data center workload growing exponentially; our strategies, growth, position, opportunities, and continued expansion; the performance and benefits of our products and technologies, including the NVIDIA Bluefield-2 DPU, NVIDIA DOCA, the EGX Edge Al platform, NVIDIA Clara Discovery, and the Cambridge-1 Al supercomputer; the NVIDIA DPU roadmap; the impact of NVIDIA accelerated computing on VMware cloud foundation; the impact of GSK's partnership with NVIDIA to build an Al drug discovery lab; our TAM; and other predictions and estimates are forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995. These forward-looking statements and any other forward-looking statements that go beyond historical facts that are made in this presentation are subject to risks and uncertainties that may cause actual results to differ materially. Important factors that could cause actual results to differ materially include: global economic conditions; our reliance on third parties to manufacture, assemble, package and test our products; the impact of technological development and competition; development of new products and technologies or enhancements to our existing product and technologies; market acceptance of our products or our partners' products; design, manufacturing or software defects; changes in consumer preferences and demands; changes in industry standards and interfaces; unexpected loss of performance of our products or technologies when integrated into systems and other factors. For a complete discussion of factors that could materially affect our financial results and operations, please refer to the reports we file from time to time with the SEC, including our Annual Reports on Form 10-K and our Quarterly Reports on Form 10-Q. Copies of reports we file with the SEC are posted on our website and are available from NVIDIA without charge. These forward-looking statements are not guarantees of future performance and speak only as of October 5, 2020, based on information currently available to us. Except as required by law, NVIDIA disclaims any obligation to update these forward- looking statements to reflect future events or circumstances. © 2020 NVIDIA Corporation. All rights reserved. NVIDIA and the NVIDIA logo are trademarks and/or registered trademarks of NVIDIA Corporation in the U.S. and/or other countries. Other company and product names may be trademarks of the respective companies with which they are associated. Features, pricing, availability, and specifications are subject to change without notice.#3AGENDA Computing for the Age of Al NVIDIA AI NVIDIA Enterprise Al NVIDIA Edge Al NVIDIA Healthcare GTC Fall 2020 Paresh Kharya Al Computing HERE WITH YOU TODAY Jensen Huang Founder & CEO Manuvir Das Enterprise Computing Colette Kress EVP & CFO Justin Boitano Edge Computing Kimberly Powell Healthcare#4NVIDIA RTX COMPUTING FOR THE AGE OF AI NVIDIA HPC arm ☐☐☐☐☐☐ S NVIDIA AI www. NVIDIA Enterprise Al CREDIT CARD 1234 5438 1932 Mellanox TECHNOLOGIES NVIDIA Edge Al VIDIA#5DLSS 2.1 RTX DI Optix 7.2 HPC SDK 20.9 RAPIDS 0.16 Parabricks 3.5 DeepStream 5.0 NSIGHT 2020.5 cuDNN 8.03 TensorRT 7.2 CUDA 11.1 NCCL 2.7.8 GPUDirect Storage RTX HPC RAPIDS Al CLARA CUDA-X CUDA MAGNUM IO EXPANDING NVIDIA ECOSYSTEM Al and Autonomous Systems Developers Adopt NVIDIA AI METRO DRIVE 80 New SDKs ISAAC AERIAL 5G 2.3M Developers 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 EST 6M CUDA Downloads in 2020; 20M downloads to date 1,800 GPU-Accelerated Applications 6,500 Al Startups#6NVIDIA NVIDIA AI INFERENCE LEADERSHIP Paresh Kharya#7Petaflop/s - Days 1.E+04 1.E+03 1.E+02 1.E+01 1.E+00 1.E-01 1.E-02 1.E-03 EXPLODING MODEL COMPLEXITY 30,000X in 5 Years | Now Doubling Every 2 Months AlexNet 2012 ARTIFICIAL INTELLIGENCE - THE MOST POWERFUL TECHNOLOGY FORCE OF OUR TIME 2014 ResNet Megatron-GPT2 GPT-2 BERT 2017 GPT-3 Megatron-BERT Turing NLG 2020 HUMAN-LEVEL READING COMPREHENSION 24% Better than Avg Human NVIDIA Megatron BERT FIGHTING CORONAVIRUS Mapping Millions of Proteins Technical U. of Munich NLPs HUMAN-LIKE CHATBOTS Half of Users Preferred Al Facebook BlenderBot CYS 112 DRUG DISCOVERY 1,000X Faster Simulations Caltech OrbNet DRONE DELIVERIES 6X Safer Landings Caltech Neural Lander ALL-TERRAIN WALKING ROBOT On the Fly Surface Adaptation NVIDIA RL Controller#8E May '20 NVIDIA NEWSROOM Press Release SHARE in Monday, June 22, 2020 World's Top System Makers Unveil NVIDIA A100-Powered Servers to Accelerate Al, Data Science and Scientific Computing = SE Cisco, Dell Technologies, HPE, Inspur, Lenovo, Supermicro Announce Systems Coming This Summer June '20 A100 TAKES OFF Fastest GPU Ramp Ever Blog Google Cloud COMPUTE Menu v New Compute Engine A2 VMs -first NVIDIA Ampere A100 GPUs in the cloud July '20 Microsoft Azure Blog Virtual Machines Bringing Al supercomputing to customers Posted on August 19, 2020 lan Finder Senior Program Manager, Accelerated HPC Infrastructure Aug '20 O = Q Press Release Oracle Cloud Infrastructure First to Make NVIDIA A100 GPUs Generally Available in the Cloud Best-in-class HPC cloud combines price-performance of on-premises with fle of cloud Developers and data scientists can boost data science, HPC, and machine learning performance by 2.5-6x, at lowest cost Redwood Shores, Calif-September 22, 2020 Sept '20#9HugeCTR RAPIDS DATA ANALYTICS & TRAINING mxnet Paddle Paddle APACHE Spark CUDA-X-AI CUDA MAGNUM 10 ONNX RUNTIME PyTorch TensorFlow 3 PILLARS OF NVIDIA AI Pre-Trained Model PRE-TRAINED MODELS SOTA Models Data Conversational Al Computer Vision TRANSFER LEARNING TOOLKIT Zero Coding | Speech, Vision, & NLU INFERENCE Recommendations Reinforcement Learning TLT Customized Model Frameworks TRITON Inference Serving TIS - Query TensorRT 7.2 TRT Result CLARA Healthcare ISAAC Robotics APPLICATION FRAMEWORKS DRIVE Autonomous Vehicles MERLIN Recommendation Systems JARVIS Conversational Al METROPOLIS Smart City#107TH GENERATION TensorRT (Inference Compiler) 2000+ Optimizations TensorFlow, PyTorch, ONNX 1.3M Downloads 16,000 Companies TRT ACCELERATING ADOPTION OF NVIDIA AI INFERENCE Celebrating 10 Years of NVIDIA GPU in the Cloud ACCELERATED ADOPTION OF NVIDIA GPU (EXA OPS) 2 25 166 <1 2017 2019 2020 2018 Inference ExaOPS shipped: Represents 12-month periods from October of the previous year to September of the year shown 1,000,000 100,000 10,000 1,000 100 10 1 TOTAL CLOUD AI INFERENCE COMPUTE CAPACITY (PETA OPS) 0 aws Total CPU 2010 NVIDIA M2050 Total GPU (-) Alibaba Cloud aws BAIDU AI CLOUD Google Cloud Microsoft Azure ORACLE Cloud Infrastructure Tencent Cloud 2020 NVIDIA A100#11MICROSOFT OFFICE Word Fiew and Bing for CUDA -Saved Le Reve Search was nu eecubes on MS Wows pas Today, NVIDIA and Microsoft's collaboration opens the Public Preview accessibility for those registered in the NVIDIA Developer Program and the Microsoft Windows Insider Program. Program members can access the CUDA Driver/Container package via Downloads and Documentation from NVIDIA'S CUDA on our WSL page Targeting a spectrum of Al enthusiast from expert data scientists, to beginners, and student, WSL and the work that NVIDIA has done now puts pr Get the fingertip of Windows users. With CIA can now deliver CUDA bware libraries such as RAPIDS, and such as TensorFlow, MONet, and PyTorch to a growing wave of new user, leveraging there familiar Windows Please sea our new Developer Blog CUDA on Windows Subsystem for Linux, to leam about the NVIDIA-Microsoft WSL collaboration, how CUDA works with the Windows Subsystem for Linux, NVIDIA Docker containers, and how to get started with Linux deep leaming application like TensorFlow. Corrections Tomawiting Grammar GRAMMAR CORRECTION Conciseness Clarity Formality Puction Conventions Share Comments Half a Trillion Queries per Year 200ms Inference Response 1/3rd Lower Cost Than CPU S 74 1 1 NVIDIA AI INFERENCE CUSTOMERS AMERICAN EXPRESS AMERICAN EXPRESS BUSINESS CF FROST REXPORT INC >))) 7997 MEMBER SINCE 09 AMEX FRAUD PREVENTION Over $1 Trillion in Transactions | 115M Credit Cards 2ms Latency Enables Real-Time Fraud Detection Up to 6% Higher Accuracy in Specific Segments TENCENT PCG 297, EFETTH SABER- 3 #2 ▸ SENSUPERARBEL> ALOO BONIT s.O RECOMMENDATIONS Videos, News, Music, Apps Hundreds of Billions Queries per day GPU Inference In Production#12AUTO BOSCH Continental Ford tu simple ZOOX CONSUMER INTERNET I₁ Byte Dance criteo. NAVER POSTMATES 0 Square NVIDIA AI INFERENCE CUSTOMERS CSP Alibaba Cloud aws Baidu E Google Cloud Microsoft ORACLE Tencent 腾讯 ROBOTICS bossanova brain@ corp 九鼎香 JOHN DEERE KOMATSU MIR TOYOTA MEDICAL DEEP (eff 6) GE Healthcare NUANCE Predible SIEMENS Healthineers SUBTLE MEDICAL UCSF Ci center for intelligent imaging Departent of Radegy & RETAIL & FSI AMERICAN EXPRESS Capital One Domino's ebay KENSHO KION GROUP Kroger INDUSTRIAL AUO FOXConn INNOPEAK TECHNOLOGY Micron MUSASHI SAMSUNG SPIL#13NVIDIA NVIDIA ENTERPRISE AI DATA CENTER IS THE NEW UNIT OF COMPUTING Manuvir Das#14VMs Infrastructure Management Software-defined Storage CPU MODERN DATA CENTERS ARE SOFTWARE-DEFINED Moving from Fixed-Function Appliances to Software Services on Servers NVIDIA NIC Containers Software-defined Security Software-defined Networking Acceleration Engines ㅗ 068 Manual Infrastructure Management Security Appliances Storage Systems Static Networks#15VMs Infrastructure Management Software-defined Storage DATA CENTER WORKLOAD GROWING EXPONENTIALLY CPU NVIDIA NIC Containers Software-defined Security Software-defined Networking Acceleration Engines 0-0-0-0 11 1414 0-0-0 1114 0-0-0-0-0 8 Microservices East-West Traffic Storage Access Zero Trust Security#16TRADITIONAL SERVER VMs Infrastructure Management Software-defined Storage CPU INTRODUCING THE DATA PROCESSING UNIT (DPU) Software-defined Data Center Infrastructure-on-a-Chip NVIDIA NIC Containers Software-defined Security Software-defined Networking Acceleration Engines VMs Infrastructure Management CPU Software-defined Storage NVIDIA DPU with Arm Cores & Accelerators DPU-ACCELERATED SERVER Containers Software-defined Security Acceleration Engines Software-defined Networking#17NVIDIA BLUEFIELD-2 DPU Data Center Infrastructure-on-a-Chip Combines 8 64-bit A72 Arm Cores, 2 VLIW Acceleration Engines, and Mellanox ConnectX-6 DX NIC Accelerated Security: Isolation, Root of Trust, Crypto, Key Management, Regular Expression Engine Accelerated Networking: RDMA/ROCE, GPU Direct, SDN/NFV Accelerated Storage: NVMe-oF, Elastic Block Storage, Data Integrity, De-Dup, Compression 2x 100Gb/s Ethernet and InfiniBand LFFFFF Out-of-Band Management PCIe Gen 4.0#18NVIDIA DOCA Data Center Infrastructure-on-a-Chip Architecture SDK for BlueField DPUS Open APIs - DPDK, SPDK, P4 Certified Reference Apps & Third-Party Solutions Infrastructure Management Telemetry www.m F F INFRASTRUCTURE APPLICATIONS Software-defined Storage Software-defined Security Management DOCA Storage SPDK / Kernel SNAP ASAP² Security DPDK / Kernel CRYPTO Software-defined Networking Networking P4/DPDK / Kernel ROT RDMA#19NVIDIA DPU ROADMAP Programmable Data Center Infrastructure-on-a-Chip FOFO 2021 BlueField-2X 70 SPECINT 60 TOPS AI 200 Gbps BlueField-2 70 SPECINT 0.7 TOPS 200 Gbps BlueField-3 350 SPECINT 1.5 TOPS 400 Gbps DOCA One Architecture 2022 BlueField-3X 350 SPECINT 75 TOPS AI 400 Gbps 2023 BlueField-4 1000 SPECINT 400 TOPS 400 Gbps 1000X 100X 10X 1X#20NVIDIA ACCELERATED COMPUTING ON VMWARE CLOUD FOUNDATION Bringing Al to Every Enterprise, Enabling Next-Gen Hybrid Cloud NVIDIA vmware DATA ANALYTICS & MACHINE LEARNING GPU OPERATOR CPU AI TRAINING & INFERENCE CONTAINERS GPUDIRECT RDMA GPU PROFESSIONAL VISUALIZATION DPU vm MIG VMWARE CLOUD FOUNDATION WITH TANZU PRIVATE CLOUD VIRTUAL MACHINES STORAGE ENTERPRISE APPLICATIONS ACCELERATED AI ALONGSIDE ENTERPRISE APPS VCS NETWORKING#21NVIDIA NVIDIA EDGE AI A TRILLION THINGS - THE NEXT INTERNET Justin Boitano#22(1191)) Edge Data Centers Millions of Edge DC Each with Thousands of Nodes THE AGE OF AI - A TRILLION CONNECTED THINGS - you 2M Factories 5M Call Center Agents 13M Stores 40M Miles of Road 570M Farms 10T Miles Driven per Year#23NVIDIA BlueField-2 8x Arm Cores Dual 100 Gb/s Ethernet or InfiniBand Line-speed TLS/IPSec Crypto Engine Time Triggered Transmission Tech for Telco (5T for 5G) ASAP2 SR-IOV and VirtuallO Offload ANNOUNCING EGX EDGE AI PLATFORM 포 Hel 5G AERIAL Telco METROPOLIS Smart Cities Security Storage ISAAC Robotics CLARA Healthcare CUDA-X-AI Networking JARVIS Conversational Al & Meat & POUILIY 19. yo METROPOLIS Smart Retail Kubernetes Management NVIDIA Ampere Architecture GPU 3rd generation Tensor Core New Security Engine for Confidential Al Secure, Authenticated Boot#24ΚΙΟΝ GROUP DEMATIC FANUC CR-71A/L EGX EDGE AI PLATFORM BY LOWER Kroger everseen garden choose $400 CORN $297 OR GREAT & M Northwestern Medicine WHITEBOARD COORDINATOR What time is my procedure?#25포 Hel 5G AERIAL Telco METROPOLIS Smart Cities Security Storage ISAAC Robotics CLARA Healthcare CUDA-X-AI Networking JARVIS Conversational Al Meat&POUILI Y you yo METROPOLIS Smart Retail Kubernetes Management & CATALOG PRIVATE REGISTRY FLEET COMMAND A Applications Monitoring EGX EDGE AI PLATFORM 28 USER MANAGEMENT V SUPPORT <Collapse NGC Version 2.41.2 ALL LOCATIONS Locations 16 02 --+ 45% 02 Add System 02 01 805 Deployments 16 02 (Updating) 00:00:35 System 00001 System 00002 System 00003 Add Application Create Deployment 01 See Al Deployment VISION AI 5G & CloudRAN SECURITY & NETWORKING INFRASTRUCTURE SYSTEMS CHOOCH DawnLight DEEPNORTH ADEEPVISION KOMATSU MUSASH;A ALTRan Check Point CANONICAL Atos ERICSSON FORTINET Red Hat DELL Technologies Safely You MAVENIR Juniper SUSE Lenovo DEMATIC WHITEBOARD Viasat E nuagenetworks vmware#26NVIDIA NVIDIA CLARA COMPUTATIONAL DRUG DISCOVERY Kimberly Powell#27NVIDIA CLARA DISCOVERY $1.25T Pharmaceutical Industry | $2B R&D Per Drug | 12 ½ Years To Market | 90% Failure Rate GENOMICS Clara Parabricks RAPIDS Biological Target Chemical Compounds STRUCTURE CryoSPARC, Relion AlphaFold Literature 0(1060) SEARCH RAPIDS 0(10⁹) b Blood cultures DOCKING AutoDock RAPIDS NLP Fewer and stary symplans pyelonephrt was established. Other cases of for were possible butlly The patient was repensive an Serum lactate was nie Abood reedive Anar was noted for Hemodynamics were supported with ophed A primary diagnosis st was admired (15) but the paint Our colagan 10 www.cod conting and Pplazo was started after teed and urine cursware drawn 12 lectate had need and hemodynamicshed were pas for Cawan to The patient was stepped down to protocopie a total 14 day come of Biomegatron, BioBERT 0(10²) SIMULATION Schrodinger NAMD, VMD, Open MM MELD Real World Data Drug Candidate IMAGING Clara Imaging ΜΟΝΑΙ#28HEALTHCARE & LIFE SCIENCE RESEARCH AstraZeneca CAMBRIDGE-1 AI SUPERCOMPUTER NHS Guy's and St Thomas' NHS Foundation Trust gsk LARGE-SCALE AI & DATA SCIENCE RESEARCH College LONDON Oxford 80 NVIDIA DGX A100 | 400 PETAFLOPS AI COMPUTE | 20 TERABYTES/SEC INFINIBAND | 2PB NVME | 500KW COLLABORATION WITH UK INDUSTRY, UNIVERSITIES, STARTUPS Technologies EDUCATE FUTURE AI PRACTITIONERS#29GSK PARTNERS WITH NVIDIA TO BUILD AI DRUG DISCOVERY LAB NVIDIA gsk#3030X الا AI MOMENT IN HEALTHCARE Accelerated Computing | Data Science & Machine Learning | Deep Learning | NLP Language Models 7X 15X 20X 6X 30X 6X Genomics Structure Screening Docking Simulation Imaging NLP 16X SPEEDUP ACROSS DOMAINS | FULL STACK ACCELERATION GPU Accelerated Applications vs CPU | GPU Optimizes Training Frameworks vs Open Source Vocabulary e.g. COVID 6.1B Words PubMeb Data BIOMEDICAL NLP BioMegatron 345M Param Al Model Named Entity Recognition Relation Extraction Q&A Al Skill NVIDIA BIOMEGATRON ACHIEVES STATE OF THE ART Biomedical Literature | Clinical Notes | Claims Data | Digital Health#31NVIDIA GTC Fall 2020 Colette Kress#32GTC FALL 2020 NVIDIA AI GAINING MOMENTUM 20M CUDA downloads/6M in last year | Most shipped Al inference cloud compute capacity in the world NVIDIA POWERS ENTERPRISE COMPUTING NVIDIA Al integrated into VMware & Cloudera | NVIDIA DPUS "Data Center on a Chip Architecture" NVIDIA EDGE AI FOR THE INTERNET OF A TRILLION THINGS EGX Edge Al + FleetCommand recurring revenue service NVIDIA HEALTHCARE NEAR A TIPPING POINT NVIDIA Clara for Computational Drug Discovery | Cambridge-1 | GSK BROADENING ECOSYSTEM & CUSTOMER ADOPTION Every Server & Storage OEM | Hundreds of ISVS | Thousands of Enterprises | 2.3M Developers#33NVIDIA RTX $8B 2020 Revenue* 18% Trailing 5-year CAGR *Trailing 4-quarters through F3Q 2021 outlook for Gaming and Professional Visualization market platforms COMPUTING FOR THE AGE OF AI NVIDIA HPC $10B 000000 NVIDIA AI as $45B NVIDIA ENTERPRISE AI CREDIT CARD 1234 561 212 3074 $30B Data Center Total Addressable Market by 2024 NVIDIA Al includes estimated Total Addressable Market for accelerated computing platforms used in Hyperscale/Cloud. Source: NVIDIA estimates, incorporating data from Counterpoint, Dell'Oro, Gartner, IDC, IHS, Hyperion and Strategy Analytics NVIDIA EDGE AI $15B#34NVIDIA Q&A Jensen Huang & Colette Kress