#1NVIDIA Investor Presentation Q3 FY24 November 27, 2023#2Except for the historical information contained herein, certain matters in this presentation including, but not limited to, statements as to: our financial position; our markets, market opportunity, demand and growth drivers; a broadening set of GPU-specialized CSPs; entering the holidays with our best-ever line-up for gamers and creators; generative Al emerging as the new "killer app" for high-performance PCs; being on track to exit the year at an annualized revenue run rate of $1 billion for our recurring software, support, and services offerings; Al emerging as a powerful demand driver for Professional Visualization; Foxconn incorporating Omniverse into its manufacturing process; our financial outlook, and expected tax rates for the fourth quarter of fiscal 2024; our expectations of sequential growth to be driven by Data Center, continued strong demand for compute and networking, and Gaming likely declining sequentially; the U.K. government building one of the world's fastest Al supercomputers; Jülich building its next-gen Al supercomputer; the combined Al compute capacity of all the supercomputers built on Grace Hopper across the U.S., EMEA and Japan next year; the benefits, impact, performance, features and availability of our products and technologies; the benefits, impact, features and timing of our collaborations or partnerships; NVIDIA accelerated computing being broadly recognized as the way to advance computing as Moore's law ends and Al lifts off; accelerated computing being needed to tackle the most impactful opportunities of our time; Al driving a platform shift from general purpose to accelerated computing, and enabling new, never-before-possible applications; trillion dollars of installed global data center infrastructure transitioning to accelerated computing; broader enterprise adoption of Al and accelerated computing under way; Al and accelerated computing making possible the next big waves of autonomous machines and industrial digitalization; a rapidly growing universe of applications and industry innovation; Al's ability to augment creativity and productivity; generative Al as the most important computing platform of our generation; data centers becoming Al factories; full-stack and data center scale acceleration driving significant cost savings and workload scaling; the high ROI of high compute performance; our belief that every important company will run its own Al factories; our dividend program plan; Al factories expanding our market opportunity; our Automotive design win pipeline, ramp and production expectations; our aim to engage manufacturing suppliers and goal of effecting supplier adoption of science-based environmental targets by fiscal 2026; and our plan for 100% renewable electricity for our operations and data centers by fiscal 2025 and annually thereafter are forward-looking statements. 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. NVIDIA has based these forward-looking statements largely on its current expectations and projections about future events and trends that it believes may affect its financial condition, results of operations, business strategy, short-term and long-term business operations and objectives, and financial needs. These forward-looking statements are subject to a number of risks and uncertainties, and you should not rely upon the forward-looking statements as predictions of future events. The future events and trends discussed in this presentation may not occur and actual results could differ materially and adversely from those anticipated or implied in the forward-looking statements. Although NVIDIA believes that the expectations reflected in the forward-looking statements are reasonable, the company cannot guarantee that future results, levels of activity, performance, achievements or events and circumstances reflected in the forward-looking statements will occur. Except as required by law, NVIDIA disclaims any obligation to update these forward-looking statements to reflect future events or circumstances. 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 most recent Annual Report on Form 10-K, Quarterly Reports on Form 10-Q, and Current Reports on Form 8-K. Copies of reports we file with the SEC are posted on our website and are available from NVIDIA without charge. Many of the products and features described herein remain in various stages and will be offered on a when-and-if-available basis. The statements within are not intended to be, and should not be interpreted as a commitment, promise, or legal obligation, and the development, release, and timing of any features or functionalities described for our products is subject to change and remains at the sole discretion of NVIDIA. NVIDIA will have no liability for failure to deliver or delay in the delivery of any of the products, features or functions set forth herein. NVIDIA uses certain non-GAAP measures in this presentation including non-GAAP gross profit, non-GAAP gross margin, non-GAAP operating expenses, non-GAAP operating income, non-GAAP operating margin, non-GAAP net income, non-GAAP diluted earnings per share, and free cash flow. NVIDIA believes the presentation of its non-GAAP financial measures enhances investors' overall understanding of the company's historical financial performance. The presentation of the company's non-GAAP financial measures is not meant to be considered in isolation or as a substitute for the company's financial results prepared in accordance with GAAP, and the company's non-GAAP measures may be different from non-GAAP measures used by other companies. Further information relevant to the interpretation of non-GAAP financial measures, and reconciliations of these non-GAAP financial measures to the most comparable GAAP measures, may be found in the slide titled "Reconciliation of Non-GAAP to GAAP Financial Measures”. NVIDIA#3Content Q3 FY24 Earnings Summary Key Announcements This Quarter NVIDIA Overview Financials Reconciliation of Non-GAAP to GAAP Financial Measures NVIDIA#4Q3 FY24 Earnings Summary#5Highlights Record quarter driven by strong Data Center growth • Total revenue up 206% Y/Y to $18.12B, well above outlook of $16.00B +/- 2% Data Center up 279% Y/Y to $14.51B Gaming up 81% Y/Y to $2.86B Record Data Center revenue driven by continued ramp of NVIDIA HGX platform and InfiniBand networking • Consumer internet and enterprise companies drove exceptional sequential growth, outpacing total growth • Strong demand from all hyperscale cloud service providers (CSPs), and a broadening set of GPU-specialized CSPs • Inference is contributing significantly to NVIDIA Data Center demand as Al is now in full production Gaming growth reflects strong demand for GeForce RTX 40 series GPUs for back-to-school and the holidays GeForce RTX available at price points as low as $299 — entering the holidays with best-ever line-up for gamers and creators Gaming has doubled relative to pre-COVID levels even against the backdrop of lackluster PC market performance Gen Al emerging as new “killer app" for high-performance PCS — NVIDIA RTX is the natural platform for Al-application developers NVIDIA#6Q3 FY24 Financial Summary ■Revenue($M) -Non-GAAP GM $18,120 66.8% 66.1% $5,931 $6,051 56.1% $7,192 GAAP Non-GAAP Q3 FY24 Y/Y Q/Q Q3 FY24 Y/Y Q/Q Revenue $18,120 +206% +34% $18,120 +206% +34% $13,507 75.0% Gross Margin 74.0% +20.4 pts +3.9 pts 75.0% +18.9 pts +3.8 pts 71.2% Operating Income $10,417 +1,633% +53% $11,557 +652% +49% Net Income $9,243 +1,259% +49% $10,020 +588% +49% Diluted EPS $3.71 +1,274% +50% $4.02 +593% +49% Q3 FY23 Q4 FY23 Q1 FY24 Q2 FY24 Q3 FY24 Cash Flow from Ops $7,333 +1,771% +16% $7,333 +1,771% +16% All dollar figures are in millions other than EPS. Refer to Appendix for reconciliation of Non-GAAP measures. NVIDIA#7$4,284 $3,833 $3,616 $10,323 Data Center 279% Y/Y and 41% Q/Q $14,514 Q3 FY23 Q4 FY23 Q1 FY24 Q2 FY24 Q3 FY24 Revenue ($M) . Highlights Data Center compute revenue quadrupled from last year, Networking revenue nearly tripled Strong, broad-based demand for NVIDIA accelerated computing fueled by investment in the buildout of infrastructure for LLMs, recommendation engines, and gen Al applications Networking business now exceeds a $10 billion annualized revenue run rate NVIDIA H100 Tensor Core GPU instances are now generally available in virtually every cloud, and are in high demand Vast majority of revenue driven by NVIDIA Hopper HGX, with a lower contribution from the prior-gen Ampere GPU architecture New L40S GPU began to ship; first revenue quarter for GH200 On track to exit the year at an annualized revenue run rate of $1 billion for our recurring software, support, and services offerings NVIDIA#8$1,574 $1,831 $2,240 $2,486 Gaming 81% Y/Y and 15% Q/Q $2,856 . Highlights Strong demand in the important back-to-school shopping season The RTX ecosystem continues to grow; there are now over 475 RTX enabled games and applications Released Tensor RT-LLM for Windows, which speeds on-device LLM inference by up to 4X GeForce NOW surpassed 1,700 PC titles including Alan Wake II, Baldur's Gate 3, Cyberpunk 2077: Phantom Liberty, and Starfield Q3 FY23 Q4 FY23 Q1 FY24 Q2 FY24 Q3 FY24 Revenue ($M) NVIDIA#9$200 $226 $295 Professional Visualization $379 108% Y/Y and 10% Q/Q $416 Highlights Al emerging as a powerful demand driver, including inference for Al imaging in healthcare, edge Al in smart spaces and the public sector Launched a new line of desktop workstations based on NVIDIA RTX Ada Lovelace generation GPUs and ConnectX SmartNICs Mercedes-Benz is using Omniverse-powered digital twins to plan, design, build and operate its manufacturing and assembly facilities Foxconn will incorporate Omniverse into its manufacturing process Announced two new Omniverse Cloud services on Microsoft Azure ― for virtual factory simulation and autonomous vehicle simulation Q3 FY23 Q4 FY23 Q1 FY24 Q2 FY24 Q3 FY24 Revenue ($M) NVIDIA#10$251 Automotive 4% Y/Y and 3% Q/Q $294 $296 $261 $253 Q3 FY23 Q4 FY23 Q1 FY24 Q2 FY24 Q3 FY24 Revenue ($M) Highlights Growth primarily driven by continued growth in self-driving platforms based on NVIDIA DRIVE Orin SoC, and the ramp of Al cockpit solutions with global OEM customers Extended automotive partnership with Foxconn to include NVIDIA DRIVE Thor, next-generation automotive SoC NVIDIA#11$392 $2,911 $2,249 $6,348 Sources & Uses of Cash 1,771% Y/Y and 16% Q/Q $7,333 Highlights • Y/Y and Q/Q growth primarily driven by higher revenue partially offset by higher cash tax payments Utilized cash of $3.9 billion towards shareholder returns, including $3.8 billion in share repurchases and $99 million in cash dividends Invested $291M in capex (includes principal payments on PP&E) Ended the quarter with $18.3B in gross cash and $9.8B in debt; $8.5B in net cash Q3 FY23 Q4 FY23 Q1 FY24 Q2 FY24 Q3 FY24 Cash Flow from Operations ($M) Gross cash is defined as cash/cash equivalents & marketable securities. Debt is defined as principal value of debt. Net cash is defined as gross cash less debt. NVIDIA#12Revenue Q4 FY24 Outlook $20.0 billion, plus or minus 2% Expect strong Q/Q growth to be driven by Data Center, with continued strong demand for both compute and networking. Gaming will likely decline Q/Q, as it is now more aligned with notebook seasonality Gross Margins 74.5% GAAP and 75.5% non-GAAP, plus or minus 50 basis points Operating Expense Other Income & Expense Tax Rate Approximately $3.17 billion GAAP and $2.20 billion non-GAAP Income of approximately $200 million for GAAP and non-GAAP Excluding gains and losses on non-affiliated investments 15.0% GAAP and non-GAAP, plus or minus 1%, excluding discrete items Refer to Appendix for reconciliation of Non-GAAP measures. NVIDIA#13Key Announcements This Quarter#14NVIDIA® New TensorRT-LLM Software More Than Doubles Inference Performance TensorRT-LLM Supercharges Hopper Performance Software optimizations double leading performance 8X Increase in GPT-J 6B Inference Performance 4.6X Higher Llama2 Inference Performance NVIDIA developed TensorRT-LLM, an open-source software library that enables customers to more than double the inference performance of their GPUs TensorRT-LLM on H100 GPUs provides up to an 8X performance speedup compared to prior generation A100 GPUs running GPT-J 6B without the software 5.3X reduction in TCO and 5.6X reduction in energy costs With TensorRT-LLM for Windows, LLMs and generative Al applications can run up to 4x faster locally on PCs and Workstations powered by NVIDIA GeForce RTX and NVIDIA RTX GPUs TensorRT-LLM for data centers now publicly available; TensorRT-LLM for Windows in beta 8X 7X 6X 5X 2X 5X 8x 4X 3X 4X 4x 2X 3X 1X 1X 2.6X 4.6X 1X 1x OX OX A100 H100 August H100 Tensor RT- A100 LLM H100 August H100 TensorRT- LLM Text summarization, variable input/output length, CNN / DailyMail dataset | A100 FP 16 PyTorch eager mode/H100 FP8 | H100 FP8, TensorRT-LLM, in-flight batching#15NVIDIA Partners With Foxconn to Build Factories and Systems for the Al Industrial Revolution Foxconn, the world's largest manufacturer, will integrate NVIDIA technology to develop "Al factories", a new class of data centers Based on the NVIDIA accelerated computing platform, including NVIDIA GH200 and NVIDIA AI Enterprise software, these Al factories will power a wide range of applications, including: Digitalization of manufacturing and inspection workflows Development of Al-powered EVs and robotics platforms A growing number of language-based generative Al services In addition: Foxconn Smart EV will be built on NVIDIA DRIVE Hyperion 9, next-gen platform for autonomous automotive fleets, powered by NVIDIA DRIVE Thor, our future automotive SoC Foxconn Smart Manufacturing robotic systems will be built on the NVIDIA Isaac autonomous mobile robot platform. Foxconn Smart City will incorporate the NVIDIA Metropolis intelligent video analytics platform NVIDIA® Al Factory Data DASRAH NVIDIA DRIVE NVIDIA AI NVIDIA Orin AV Fleet Al factories are a new class of data centers, optimized for refining data and training, inferencing, and generating Al#16NVIDIA® NVIDIA Partners With India Tech Giants to Advance Al Across World's Most Populous Nation NVIDIA announced collaborations with Reliance Industries, Tata Group and Infosys to bring Al technology and skills to India • With Reliance, the companies will work together to develop India's own foundation LLM trained on India's diverse languages and tailored for generative Al applications; build supercomputing infrastructure to support the exponential computational demands of Al With Tata, the collaboration will bring a state-of-the-art Al supercomputer to provide infrastructure-as-a-service and platform for Al services in India With Infosys, the partnership will bring the NVIDIA AI Enterprise ecosystem of models, tools, runtimes and GPU systems to drive productivity gains with generative Al applications and solutions Infosys plans to set up an NVIDIA Center of Excellence where it will train and certify 50,000 of its employees on NVIDIA AI technology Infosys (R Reliance Industries Limited TATA#17NVIDIA Sets New LLM Training Record With Largest MLPerf Submission Ever ILF 111 1997 Six New Performance Records The fastest gets even faster NVIDIA set six new performance records in this round, with the performance increase stemming from a combination of advances in software and scaled-up hardware 2.8x faster on generative AI - completing a training benchmark based on a GPT-3 model with 175 billion parameters trained on 1 billion tokens in just 3.9 minutes 1.6x faster on training recommender models 1.8x faster on training computer vision models The GPT-3 benchmark ran on NVIDIA Eos - a new Al supercomputer powered by 10,752 H100 GPUs and NVIDIA Quantum-2 InfiniBand networking • The 10,752 H100 GPUs far surpassed the scaling in Al training in June, when NVIDIA used 3,584 Hopper GPUs The 3x scaling in GPU numbers delivered a 2.8x scaling in performance, a 93% efficiency rate thanks in part to software optimizations Microsoft Azure achieved similar results on a nearly identical cluster, demonstrating the efficiency of NVIDIA AI in public cloud deployments GPT-3 175B (1B Tokens) 3.9 Minutes 2.8X Faster Stable Diffusion 2.5 Minutes New Workload DLRM-dcnv2 1 Minute 1.6X Faster BERT-Large 7.2 Seconds 1.1X Faster 5 RetinaNet 55.2 Seconds 1.8X Faster 3D U-Net 46 Seconds 1.07X Faster MLPerf™ Training v3.1. Results retrieved from www.mlperf.org on November 8, 2023. Format: Chip Count, MLPerf ID | GPT-3: 3584x 3.0-2003, 10752x 3.1-2007 | Stable Diffusion: 1024x 3.1-2050 | DLRMv2: 128x 3.0-2065, 128x 3.1-2051 | BERT-Large: 3072x 3.0-2001, 3472x 3.1-2053| RetinaNet: 768x 3.0-2077, 2048x 3.1-2052 | 3D U-Net: 432x 3.0-2067, 768x 3.1-2064. The MLPerf™ name and logo are trademarks of MLCommons Association in the United States and other countries. All rights reserved. Unauthorized use strictly prohibited. See www.mlcommons.org for more information. NVIDIA®#18New NVIDIA HGX H200 Supercharges Hopper NVIDIA NVIDIA H200 is the first GPU to offer HBM3e — faster, larger memory to fuel the acceleration of generative Al and large language models, while advancing scientific computing for HPC workloads H200 delivers 141GB of memory at 4.8 terabytes per second, nearly double the capacity and 2.4X more bandwidth compared with its predecessor, NVIDIA A100 Boosts inference speed by up to 2X compared to H100 GPUs when handling LLMs such as Llama2 Microsoft announced plans to add the H200 to Azure next year for larger model inference with no increase in latency H200-powered systems from the world's leading server manufacturers and cloud service providers are expected to begin shipping in the second quarter of 2024 NVIDIA® NVIDIA NVIDIA NVIDIA#19NVIDIA® Grace Hopper Gains Significant Traction with Supercomputing Customers Cumulative Al Performance (ExaFLOPS of AI) Isambard AI Initial shipments to Los Alamos National Lab and the Swiss National Supercomputing Centre took place in the third quarter The U.K. government announced it will build one of the world's fastest Al supercomputers with almost 5.5K Grace Hopper Superchips German supercomputing center Jülich will build its next- gen Al supercomputer, with close to 24K Grace Hopper Superchips and Quantum-2 InfiniBand Will be the world's most powerful Al system with over 90 exaflops of Al performance Marks the debut of a quad NVIDIA GH200 Grace Hopper Superchip node configuration Combined Al compute capacity of all the supercomputers built on Grace Hopper across the U.S., EMEA and Japan next year estimated to exceed 200 exaflops 400 350 300 250 200 150 100 50 VENADO JCAHPC orm C JUPITER ALPS CSCS ETH Zürich OFP-II Isambard Al Delta Jupiter Venado Vista Alps Eos Azure ND H100 v5 Mare Nostrum-5 Perlmutter Selene Leonardo Piz Daint Tsubame3 Summit Pangea Polaris 0 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 Cumulative AI FLOPS#20NVIDIA® NVIDIA AI Foundry Service for Enterprises on Microsoft Azure Create from Foundation Model Microsoft Azure Introduced new NVIDIA AI foundry service for the development and tuning of custom generative Al enterprise applications, running on Microsoft Azure Customers can bring their domain knowledge and proprietary data, and we help them build their Al models using our Al expertise and software stack in DGX Cloud Al factory - all with enterprise-grade security and support Businesses can deploy their customized models with the NVIDIA AI Enterprise software runtime to power generative Al applications such as intelligent search, summarization, and content generation Industry leaders SAP SE, Amdocs and Getty Images are among the first customers of NVIDIA AI foundry service Your Enterprise Model Run Anywhere Al Foundations NeMo DGX Cloud Running on NVIDIA AI Enterprise Microsoft Azure RAG LLM B Prompts Agent LLM Vector Store#21NVIDIA® NVIDIA Spectrum-X Ethernet networking platform for Al Available Soon from Dell, HPE and Lenovo Purpose-built for gen Al, Spectrum-X offers enterprises a new class of Ethernet networking that can achieve 1.6x higher networking performance for Al communication versus traditional Ethernet offerings Dell, Hewlett Packard Enterprise and Lenovo will be the first to integrate NVIDIA Spectrum-X Ethernet networking technologies for Al into their server lineups New systems bring together Spectrum-X with NVIDIA GPUs, NVIDIA AI Enterprise software and NVIDIA AI Workbench software to provide enterprises the building blocks to transform their businesses with generative Al Available in the first quarter of next year о о 10 O:#22NVIDIA® NVIDIA Collaborates With Genentech to Accelerate Drug Discovery Using Generative Al Genentech is pioneering the use of generative Al to discover and develop new therapeutics and deliver treatments to patients more efficiently NVIDIA will work with Genentech to accelerate Genentech's proprietary algorithms on NVIDIA DGX Cloud Genentech plans to use NVIDIA BioNeMo to help accelerate and optimize their Al drug discovery platform NVIDIA plans to use insights learned from this collaboration to improve its BioNeMo platform BioNeMo is now generally available as a training service#23NVIDIA Overview#24Headquarters: Santa Clara, CA NVIDIA pioneered accelerated computing to help solve impactful challenges classical computers cannot. A quarter of a century in the making, NVIDIA accelerated computing is broadly recognized as the way to advance computing as Moore's law ends and Al lifts off. NVIDIA's platform is installed in several hundred million computers, is available in every cloud and from every server maker, powers 76% of the TOP500 supercomputers, and boasts 4.5 million developers. KBB ENDEAVOR#25AI APPLICATION FRAMEWORK VOICE MODULUS ΜΟΝΑΙ RIVA MAXINE PLATFORMS H NVIDIA's Accelerated Computing Platform Full-stack innovation across silicon, systems and software NEMO MERLIN CUOPT MORPHEUS TOKKIO AVATAR DRIVE ISAAC NVIDIA HPC cuNumeric CV-CUDA cuQuantum NVIDIA AI Parabricks Sionna Jetpack ACCELERATION LIBRARIES RAPIDS Spark cuDNN TensorRT cuGraph Triton Deepstream Flare CLOUD-TO-EDGE DATACENTER-TO- ROBOTIC SYSTEMS 3-CHIPS DOCA Mag 10 Aerial RTX DGX HGX EGX OVX Super POD AGX GPU CPU DPU BUS VAN AN IGX NVIDIA Omniverse MOTORBIKE PEDESTRI BUS CYCLIST METROPOLIS HOLOSCAN With nearly three decades of singular focus, NVIDIA is expert at accelerating software and scaling compute by a Million-X, going well beyond Moore's law Accelerated computing requires full-stack innovation - optimizing across every layer of computing - from silicon and systems to software and algorithms, demanding deep understanding of the problem domain Our full-stack platforms - NVIDIA HPC, NVIDIA AI, and NVIDIA Omniverse - accelerate high performance computing, Al and industrial digitalization workloads We accelerate workloads at data center scale, across thousands of compute nodes, treating the network and storage as part of the computing fabric Our platform extends from the cloud and enterprise data centers to supercomputing centers, edge computing and PCs NVIDIA#26NVIDIA® What Is Accelerated Computing? A full-stack approach: silicon, systems, software Amdahl's law: The overall system speed-up (S) gained by optimizing a single part of a system by a factor (s) is limited by the proportion of execution time of that part (p). S = || 1 (1 − p) + Р S Not just a superfast chip - accelerated computing is a full-stack combination of: Chip(s) with specialized processors Algorithms in acceleration libraries Domain experts to refactor applications To speed-up compute-intensive parts of an application For example: If 90% of the runtime can be accelerated by 100X, the application is sped up 9X If 99% of the runtime can be accelerated by 100X, the application is sped up 50X If 80% of the runtime can be accelerated by 500X, or even 1000X, the application is sped up 5X OOOOOOOO#27NVIDIA® Why Accelerated Computing? Advancing computing in the post-Moore's Law era Accelerated computing is needed to tackle the most impactful opportunities of our time—like Al, climate simulation, drug discovery, ray tracing, and robotics NVIDIA is uniquely dedicated to accelerated computing ―working top-to-bottom, refactoring applications and creating new algorithms, and bottom-to-top-inventing new specialized processors, like RT Core and Tensor Core "It's the end of Moore's Law as we know it." - John Hennessy Oct 23, 2018 "Moore's Law is dead." - Jensen Huang, GTC 2013 Trillions of Operations per Second (TOPS) 109 108 107 GPU-Computing perf 2X per year 106 105 104 103 102 Single-threaded CPU perf 1980 1990 1.5X perf per year 1.1X per year 1000X In 10 years 2000 2010 2020 2030#28Waves of Adoption of Accelerated Computing A generational computing platform shift Enterprise Cloud Service Providers & Consumer Internet Industrial Digitalization Autonomous Vehicles & Robotics A new computing era has begun Accelerated computing enabled the rise of Al, which is driving a platform shift from general purpose to accelerated computing, and enabling new, never-before-possible applications The trillion dollars of installed global data center infrastructure will transition to accelerated computing to achieve better performance, energy- efficiency and cost by an order of magnitude Hyperscale cloud service providers and consumer internet companies have been the early adopters of Al and accelerated computing, with broader enterprise adoption now under way Al and accelerated computing will also make possible the next big waves - autonomous machines and industrial digitalization NVIDIA#29NVIDIA Accelerated Computing for Every Wave Enterprise Cloud Service Providers & Consumer Internet Industrial Digitalization Autonomous Vehicles & Robotics NVIDIA Omniverse is a software platform for designing, building, and operating 3D and virtual world simulations. It harnesses the power of NVIDIA graphics and Al technologies and runs on NVIDIA-powered data centers and workstations NVIDIA DRIVE is a full-stack platform for autonomous vehicles (AV) that includes hardware for in-car compute, such as the Orin system-on-chip, and the full AV and Al cockpit software stack NVIDIA DGX Cloud is a cloud service that allows enterprises immediate access to the infrastructure and software needed to train advanced models for generative Al and other groundbreaking applications NVIDIA AI Enterprise is the operating system of Al, with enterprise-grade security, stability, manageability and support. It is available on all major CSPs and server OEMs and supports enterprise deployment of Al in production NVIDIA HGX is an Al supercomputing platform purpose-built for Al. It includes 8 NVIDIA GPUs, as well as interconnect and networking technologies, delivering order-of-magnitude performance speed-ups for Al over CPU servers. It is broadly available from all major server OEMs/ODMs. NVIDIA DGX, an Al server based on the same architecture, along with NVIDIA AI software and support, is also available NVIDIA#30NVIDIA's Accelerated Computing Ecosystem Developers CUDA Downloads* 1.8M 4.5M 2020 48M 20M 2023 2020 2023 Al Startups GPU-Accelerated Applications 6K 15K 700 3,200 2020 2023 2020 2023 *Cumulative The NVIDIA accelerated computing platform has attracted the largest ecosystem of developers, supporting a rapidly growing universe of applications and industry innovation Developers can engage with NVIDIA through CUDA — our parallel computing programming model introduced in 2006 — or at higher layers of the stack, including libraries, pre-trained Al models, SDKs and other development tools 300 Libraries 600 AI Models 100 Updated in the Last Year NVIDIA#31NVIDIA's Multi-Sided Platform and Flywheel Scale R&D $ Installed Base Speed-Up Developers End-Users Cloud & OEMS NVIDIA Accelerated Computing Virtuous Cycle The virtuous cycle of NVIDIA's accelerated computing starts with an installed base of several hundred million GPUs, all compatible with the CUDA programming model For developers - NVIDIA's one architecture and large installed base give developer's software the best performance and greatest reach For end users - NVIDIA is offered by virtually every computing provider and accelerates the most impactful applications from cloud to edge For cloud providers and OEMS - NVIDIA's rich suite of Acceleration Platforms lets partners build one offering to address large markets including media & entertainment, healthcare, transportation, energy, financial services, manufacturing, retail, and more For NVIDIA - Deep engagement with developers, computing providers, and customers in diverse industries enables unmatched expertise, scale, and speed of innovation across the entire accelerated computing stack - propelling the flywheel NVIDIA#32Huge ROI from Al Driving a Powerful New Investment Cycle Al can augment creativity and productivity by orders of magnitude across industries Knowledge workers will use copilots based on large language models to generate documents, answer questions, or summarize missed meetings, emails and chats - adding hours of productivity per week Copilots specialized for fields such as software development, legal services or education can boost productivity by as much as 50% Social media, search and e-commerce apps are using deep recommenders to offer more relevant content and ads to their customers, increasing engagement and monetization Creators can generate stunning, photorealistic images with a single text prompt - compressing workflows that take days or weeks into minutes in industries from advertising to game development Call center agents augmented with Al chatbots can dramatically increase productivity and customer satisfaction Drug discovery, financial services, agriculture and food services and climate forecasting are seeing order-of-magnitude workflow acceleration from Al Maps ON LINE-SHOP Maps Microsoft Edge SANDALS Media Player Microsoft Edge Microsoft Endpoint Manager Microsoft Intune Management Ext... Weather Play and explore Microsoft Office Tools Database Compare Office Language Preferences Spreadsheet Compare Telemetry Log for Office Xbox Console... Movies & TV Office Al Copilots Over 1B knowledge workers Legal Services, Education 1M legal professionals in the US 9M educators in the US Catoner afumaten 0 Charles Witteam Recommended Rate Pans www hondson Carton Ohe Conversation Transcription AXAA Dos Aca Customer Service with Al 15M call center agents globally 5 YOU MAY ALSO LIKE Shopping Bag ADD TO SHOPPING BAG DREAM BOR a beautiful desert sky late in the evening Search & Social Media $700B in digital advertising annually Al Content Creation 50M creators globally HIILI Al Software Development 30M software developers globally 800 CI F F CI HS OH CI [ t ] Z Na HUN HC OH Drug Discovery 1018 molecules in chemical space 40 exabytes of genome data Bds Offers 14.40 Volume Bids Offers Volume 89,200 62.25 62.50 87,400 243,700 62.00 144 2000 62.00 62.75 49,300 118,300 143,500 61.75 63.00 227,500 13.50 High 13.50 130,900 61.75 61.50 62.50 242,40 62.75 142,60 72.75 High -1.50 (-2.02%) Low 7 Offers Volum 565.30 SENIN LOW 14.30 14.60 14.40 14.70 14.30 14.80 Volume 144,200 602.200 570,800 13.50 High 13.50 S 4475 Low 13.30 Volume 11.20 13.70 2.229,100 46.50 High 47.00 4,665,200 13.30 6,754,900 13.20 13.70 2,229,100 Bds Offers 13.40 13.50 2.898,500 1130 13.60 2.232,800 +0.10 (+0.75%) Low 13.30 Volume Bids Volume 315,900 72.75 73.00 377,000 72.50 73.25 423,40 Volume Bids Offers 633,900 13.40 13.50 2,898,500 13.60 232,800 351,600 72.25 73.50 563,00 2.10 High 212 4000.00 Low 46.25 +0.04 (+1.94%) Low 08 Offers Volume Volume Bids Offers Volume 575 3.489.300 1,319,000 210 2.12 2,372,500 7004433.300 Volume 196,000 Bids 74.50 1,707,400 2.08 2.14 7 Volun 75 748,7 1801401 697,100 1578.700 653,800 4.25 216 8020 1.224.9 744500 00 6.70 Financial Services 74.50 High 7 0(+0.68%) Low Offers 678B annual credit card transactions Agri-Food | Climate 1B people in agri-food worldwide Earth-2 for km-scale simulation Source: Goldman Sachs, Cowen, Statista, Capital One, Wall Street Journal, Resource Watch, NVIDIA internal analysis NVIDIA#33Quantum co TEXT AUDIO IMAGE 3D VIDEO DNA PROTEIN MOLECULE ANIMATION mechanics to Generative Al The most important computing platform of our generation HD+GH |||||||| CO Quantum co mechanics to ANIMATION MOLECULE PROTEIN DNA VIDEO 3D IMAGE AUDIO The era of generative Al has arrived, unlocking new opportunities for Al across many different applications Generative Al is trained on large amounts of data to find patterns and relationships, learning the representation of almost anything with structure It can then be prompted to generate text, images, video, code, or even proteins For the very first time, computers can augment the human ability to generate information and create 1,600+ Generative Al companies are building on NVIDIA TEXT NVIDIA#34Training Compute (petaFLOPs) 1010 109 108 107 106 105 104 103 Modern Al is a Data Center Scale Computing Workload Data centers are becoming Al factories: Data as input, intelligence as output Before Transformers = 8X/2yrs Transformers = 215X/2yrs 102 2011 AlexNet 2012 PALM MT NLG 530B • Chinchilla GPT-3⚫ • BLOOM Microsoft T-NLG GPT-2⚫ Megatron-NLG. XLNet Wav2Vec 2.0 MoCo ResNet50 Xception BERT Large Inception V3 • Seq2Seq Resnet VGG-19 • GPT-1 Transformer • ResNext ELMO DenseNet201 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 Al Training Computational Requirements Large Language Models, based on the Transformer architecture, are one of today's most important advanced Al technologies, involving up to trillions of parameters that learn from text. 111 Developing them is an expensive, time-consuming process that demands deep technical expertise, distributed data center-scale infrastructure, and a full-stack accelerated computing approach. 8898 NVIDIA Τηνια NVIDIA NVIDIA Jaen Fueling Giant-Scale Al Infrastructure NVIDIA compute & networking GPU | DPU | CPU NVIDIA#35Full-Stack & Data Center Scale Acceleration Drive significant cost savings and workload scaling Classical Computing―960 CPU-only servers Application CPU server racks LLM Workload: Bert-Large Training and Inference | CPU Server: Dual-EYPC 7763 | GPU Server: Dual-EPYC 7763 + 8X H100 PCle GPUs Accelerated Computing-2 GPU servers Application Re-Engineered for Acceleration CUDA-X Acceleration Libraries Magnum 10 25X lower cost 84X better energy-efficiency NVIDIA#36The High ROI of High Compute Performance $1 upfront investment in NVIDIA compute and networking can translate to $5 in CSP revenue over 4 years 4-Year Cost of Al Infrastructure ~$1B 16K GPU DC Facility Build & Operate GPU Compute Networking 15% Utilization Increase Worth $350M+ 4-Year Rental Opportunity @$4 per GPU-HR ~$2.5B Illustrative example of NVIDIA GPU cost vs Al infrastructure total cost of ownership (TCO) 25% Performance Increase Worth $600M+ NVIDIA#37NVIDIA DGX | HGX H100 NVIDIA L40S Training & Inference - One Architecture Cloud | On-Prem | Edge EIEEEEEEE Training IN THE DATA CENTER AT THE EDGE NVIDIA L40 Image Generation NVIDIA L4 Al Video NVIDIA H100 | L40S Universal GPUs NVIDIA Grace Hopper RecSys, Gen Al IGX Industrial-Grade System for Healthcare, Logistics, Manufacturing AGX Functionally-Safe System for Autonomous Vehicles Inference NVIDIA#38NVIDIA AI Foundation Pre-Trained LLMs NVIDIA DGX Cloud Microsoft Azure ORACLE Google Cloud CLOUD Infrastructure I Powering the Al Industrial Revolution Building and Running Enterprise Gen Al Applications NVIDIA AI Enterprise 1 aws Cloud Google Cloud DGX Cloud Microsoft Azure ORACLE CLOUD Infrastructure Enterprise On-Prem DELL Hewlett Packard Enterprise Lenovo vmware Enterprise SaaS & Al Platforms ΑΙ Foundation Model Tech Custom LLM Model Container DGX Cloud Factory NVPS Experts A Adobe gettyimages servicenow databricks Hugging Face snowflake WPP NVIDIA AI foundry service for building Enterprise Al applications JIL NVIDIA AI enterprise ecosystem for running Enterprise Al applications Enterprise Al Chatbot with "RAG" Vector Database Cloud Al APIS Enterprise Al chatbots Are built with Retrieval Augmented Generation (RAG), which augments the knowledge in the LLM with Enterprise data mapped to a Vector Database, thus reducing "hallucinations". Developers can connect additional or 3rd party services to the Al chatbot via cloud Al APIs. NVIDIA#39NVIDIA® The NVIDIA Al Foundry Model on DGX Cloud For building enterprise Al applications Pre-trained LLMs | | NVIDIA'S "Al foundry" service leverages our Al infrastructure and expertise to build custom Al models for enterprise customers — analogous to a semiconductor foundry that uses its infrastructure and expertise to build custom chips for fabless customers. An enterprise customer starts with an NVIDIA or 3rd party pre-trained Al model, available in NVIDIA AI Foundations. This model making service includes frameworks such as NVIDIA NeMo for custom LLMs and NVIDIA Picasso for custom generative Al for visual design. With help from NVIDIA experts, the enterprise customer fine-tunes the model on their proprietary enterprise data and adds guardrails, using tools available in NVIDIA AI Foundations. The fine-tuning and optimization is done on NVIDIA DGX Cloud, a cloud service that allows enterprises immediate access to NVIDIA Al infrastructure and software, hosted at partner cloud providers. The enterprise customer ends up with a fully-trained and optimized Al model, fine-tuned on their proprietary enterprise data, that can be deployed anywhere in the cloud or on-prem. The NVIDIA AI Foundry model generates revenue based on per-node, per-month consumption of NVIDIA DGX Cloud. | | NVIDIA AI Foundations NeMo | Picasso NVIDIA DGX Cloud Microsoft Azure Google Cloud ORACLE CLOUD Infrastructure NVIDIA AI foundry#40NVIDIA® Al Factories - A New Class of Data Centers For running enterprise Al applications | | "Al factories" are a new class of data centers specially built for processing, refining and transforming vast amounts of data into valuable Al models and tokens. Unlike traditional data centers built for IT workloads, Al factories are built to deliver automated, professional skills. Al factories are not multi-workload or multi-tenant. They run one workload – an Al model - and have just one customer or owner – analogous to a traditional factory. - Al factories can be built on-prem, in the cloud, or in the data centers of SaaS and Al platform vendors. We believe that in the future, every important company will run its own Al factories in order to securely process its valuable proprietary data and turn it into monetizable tokens, encapsulating its knowledge, intelligence, and creativity. In addition to the up-front revenue opportunity from data center systems, NVIDIA can generate recurring revenue from Al factories for their use of NVIDIA AI Enterprise, the operating system for enterprise Al. DATA | NVIDIA AI Enterprise Cloud ४४७४४ Al Factory Enterprise On-Prem Enterprise SaaS & Al Platforms I TOKENS#41NVIDIA AI Enterprise The operating system for enterprise Al NVIDIA AI Enterprise NVIDIA AI Enterprise is software for deploying and running Al with enterprise-grade security, API stability, manageability and support. Cloud-native and available in every major cloud marketplace. Certified to run on servers and workstations from all major OEMs. Supported by all major global system integrators. Integrated with and distributed by VMware. 三三三三 Hello LLM Medical Imaging Al Use Cases and Workflows վիլի Speech Al Γ ד E Recommenders Cybersecurity L Video Analytics Route Optimization More Run Anywhere NVIDIA AI Enterprise Cloud Azure | GCP | OCI | AWS Consumption pricing per GPU-hour NVIDIA Certified Server Dell | HPE | Lenovo Subscription pricing per GPU/year (included with H100 PCIe/DGX) NVIDIA#42GSI & Service Delivery accenture NVIDIA AI Enterprise Broad and deep ecosystem and distribution to reach every enterprise + III (ap)) Al Platforms Software Platforms Booz | Allen | Hamilton databricks Hugging Face snowflake gettyimages servicenow shutterstock A Adobe WPP Capgemini Deloitte. Infosys tcs ΤΑΤΑ CONSULTANCY SERVICES wipro Public Cloud Marketplaces aws Private Cloud Google Cloud Microsoft Azure vmware® BOXX 1|1.1|1 CISCO DELL Technologies ORACLE Server OEMs Cloud Infrastructure HPE GreenLake hp Lenovo SUPERMICRO NVIDIA#43NVIDIA Partners NVIDIA Go-to-Market Across Cloud and On-Premises Reaching customers everywhere DGX Cloud Google Cloud Microsoft Azure ORACLE CLOUD Infrastructure NVIDIA AI Foundations - Cloud services for customizing and operating generative Al models DGX aws Google Cloud Microsoft Azure ORACLE CLOUD Infrastructure DELL Technologies Hewlett Packard Enterprise Lenovo == HGX Cloud INFERENCE MGX AGX IGX On-Prem NVIDIA#44$11,716 $10,918 Driving Strong & Profitable Growth Operating Income (Non-GAAP, $M) -Operating Margin (Non-GAAP) Revenue ($M) $24,000 $22,385 70% $38,819 $20,000 58% 60% $16,000 $12,000 $12,690 47% 50% $9,040 $6,803 $8,000 $4,407 34% 40% $3,735 $26,914 $26,974 41% $4,000 38% 34% $0 FY19 FY20 FY21 FY22 FY23 30% YTD FY24 $16,675 FY19 FY20 FY21 FY22 FY23 YTD FY24 Fiscal year ends in January. Refer to Appendix for reconciliation of Non-GAAP measures. Operating margins rounded to the nearest percent. 41 YTD FY21 34 7 45 YTD FY24 75 321 19 Gaming Data Center ProViz Auto OEM & Other FY23 financial metrics reflect a $2.2B charge for inventory and related reserves primarily related to Data Center and Gaming. NVIDIA#45NVIDIA Gross Margins Reflect Value of Acceleration Accelerated computing requires full-stack and data center-scale innovation across silicon, systems, algorithms and applications. Significant expertise and effort are required, but application speed-ups can be incredible, resulting in dramatic cost and time-to-solution savings. For example, 2 NVIDIA HGX nodes with 16 NVIDIA H100 GPUs that cost $400K can replace 960 nodes of CPU servers that cost $10M for the same LLM workload. NVIDIA chips carry the value of the full-stack, not just the chip. Gross Profit (Non-GAAP, $M) —Gross Margin (Non-GAAP) $28,000 $17,969 $15,965 67% $10,947 $6,821 $7,233 66% 62% 63% 59% 72% FY19 FY20 FY21 FY22 FY23 YTD FY24 Cost comparison example based on latest available NVIDIA A100 GPU and Intel CPU inference results in the commercially available category of the MLPerf industry benchmark; includes related infrastructure costs such as networking. FY23 financial metrics reflect a $2.2B charge for inventory and related reserves primarily related to Data Center and Gaming. Fiscal year ends in January. Refer to Appendix for reconciliation of Non-GAAP measures. Gross margins are rounded to the nearest percent. NVIDIA#46Strong Cash Flow Generation Free Cash Flow (Non-GAAP) $15.7B $4.7B $4.3B $3.1B $8.0B $3.8B FY19 FY20 FY21 FY22 FY23 YTD FY24 Fiscal year ends in January. Refer to Appendix for reconciliation of Non-GAAP measures. 1 Subject to continuing determination by our Board of Directors. Capital Allocation Share Repurchase $10B repurchased in FY23 $25.2B Remaining Authorization as of end of Q3 Dividend $398M in FY 2023 Plan to Maintain¹ Strategic Investments Growing Our Talent Platform Reach & Ecosystem NVIDIA#47NVIDIA NVIDIA 150 NVIDIA Our Market Platforms at a Glance RETURNAL SACKBOY A BIG ADVENTURE WITCHER WILD HUNT FORT APLAGUETA REQUIEM 138 DYINGLIGHT2 STAY HUMAN 4437 g MINECRAFT OR WINDOWS 10 PORTAL WITH BTX RTX 4090 GH WO a 10 Gaming 33% of FY23 Revenue Professional Visualization 6% of FY23 Revenue Automotive 3% of FY23 Revenue FY23 Revenue $15.0B 5-YR CAGR 51% DGX/HGX/MGX/IGX systems GPU | CPU | DPU | Networking NVIDIA AI software FY23 Revenue $9.1B 5-YR CAGR 10% GeForce GPUs for PC gaming GeForce NOW cloud gaming FY23 Revenue $1.5B 5-YR CAGR 11% NVIDIA RTX GPUs for workstations Omniverse software FY23 Revenue $0.9B 5-YR CAGR 10% DRIVE Hyperion sensor architecture with AGX compute DRIVE AV & IX full stack software for ADAS, AV & Al cockpit Data Center 56% of FY23 Revenue NVIDIA#4851% 5-YR CAGR Through FY23 Revenue ($M) $2,932 $2,983 $6,696 $10,613 Data Center The leading accelerated computing platform $15,005 $29,121 Leader in AI & HPC #1 in Al training and inference Used by all hyperscale & major cloud computing providers and 40,000 enterprises Powers 76% of the TOP500 supercomputers FY19 FY20 FY21 FY22 FY23 YTD FY24 Growth Drivers Broad data center platform transition from general- purpose to accelerated computing Emergence of "Al factory" - optimized for refining data and training, inferencing, and generating Al Broader and faster product launch cadence to meet a growing and diverse set of Al opportunities DGX Cloud services and NVIDIA AI Enterprise software for building and running enterprise Al applications NVIDIA#49NVIDIA AI - One Architecture | Train and Deploy Everywhere One-Year Rhythm GPU CPU + GPU H100 L40S BlueField-3 DPU 400G Quantum 400G Spectrum-X x86 Training & Inference X100 B100 x86 Enterprise & Inference H200 X40 B40 Arm Training & Inference GX200NVL Arm Inference GB200NVL GX200 GH200NVL GB200 GH200 Enterprise & Hyperscale Infrastructure Computing BlueField-4 InfiniBand Al Infrastructure 1,600G Ethernet Enterprise & 800G Hyperscale Al Infrastructure 1,600G 800G 2023 2024 2025 NVIDIA#5010% 5-YR CAGR Through FY23 Revenue ($M) $12,462 $6,246 $5,518 Gaming GeForce ― the world's largest gaming platform $9,067 $7,759 $7,582 FY19 FY20 FY21 FY22 FY23 YTD FY24 Leader in PC Gaming Strong #1 market position 15 of the top 15 most popular GPUs on Steam Leading performance & innovation 200M+ gamers on GeForce Growth Drivers Rising adoption of NVIDIA RTX in games Expanding universe of gamers & creators Gaming laptops & Gen Al on PCs GeForce NOW Cloud gaming NVIDIA#51GeForce Extends Growth, Large Upgrade Opportunity GeForce Gaming Revenue 20% CAGR FY20 FY23 3YR CAGR ASP 10% Units 9% Installed Base $699+ Cumulative Sell-Through $ 47% RTX RTX 3060+ 20% RTX3060+ Performance More Gamers, Richer Mix Installed Base Needs Upgrade Source: NVIDIA estimates NVIDIA Ada NVIDIA Ampere NVIDIA Turing 0 1 2 3456789 10 11 12 13 14 15 16 17 18 19 20 Weeks After Launch Ada: 3× Turing Ramp at $699+ NVIDIA#5211% 5-YR CAGR Through FY23 Revenue ($M) $2,111 $1,212 $1,130 " $1,053 Professional Visualization Workstation graphics $1,544 $1,090 FY19 FY20 FY21 FY22 FY23 YTD FY24 Leader in Workstation Graphics 95%+ market share in graphics for workstations 45M Designers and Creators Strong software ecosystem with over 100 RTX accelerated and supported applications Growth Drivers Ray Tracing and generative Al revolutionizing design and content creation Expanding universe of designers and creators Collaborative 3D design / Omniverse Hybrid work environments NVIDIA#5310% 5-YR CAGR Through FY23 Revenue ($M) $700 $641 $536 $566 Automotive Autonomous Vehicles (AV) & Al Cockpit $903 $810 FY 19 FY20 FY21 FY22 FY23 YTD FY24 Leader in Autonomous Driving NVIDIA DRIVE is our end-to-end Autonomous Vehicle (AV) and Al Cockpit platform featuring a full software stack and is powered by NVIDIA (systems-on-a-chip) SoCs in the vehicle DRIVE Orin SoC ramp began in FY23 Next-generation DRIVE Thor SoC ramp to begin in FY26 Over 40 customers including 20 of top 30 EV makers, 7 of top 10 truck makers, 8 of top 10 robotaxi makers Growth Drivers Adoption of centralized car computing and software-defined vehicle architectures AV software and services: Mercedes-Benz Jaguar Land Rover NVIDIA#54$1 Trillion Long-Term Available Market Opportunity Enterprise Industrial Digitalization Autonomous Vehicles & Robotics Omniverse Enterprise $150B Autonomous Machines $300B NVIDIA AI Enterprise & DGX Cloud $150B Cloud Service Providers & Consumer Internet Gaming $100B Data Center Systems $300B NVIDIA#55Financials#56Annual Cash & Cash Flow Metrics Operating Income (Non-GAAP) — $M 6,803 4,407 3,735 12,690 9,040 Operating Cash Flow - $M 5,822 4,761 3,743 9,108 5,641 FY 19 FY20 FY21 FY22 FY23 FY19 FY20 FY21 FY22 FY23 Free Cash Flow (Non-GAAP) — $M 4,677 4,272 3,143 8,049 Cash Balance - $M 21,208 3,750 10,897 11.561 7,422 FY19 FY20 FY21 FY22 FY23 FY19 FY20 FY21 FY22 13,296 FY23 Cash balance is defined as cash and cash equivalents plus marketable securities Refer to Appendix for reconciliation of non-GAAP measures NVIDIA#57Corporate Responsibility Environmentally Conscious By FY26, aim to engage manufacturing suppliers comprising at least 67% of NVIDIA's scope 3 category 1 GHG emissions with goal of effecting supplier adoption of science-based targets NVIDIA GPUs are typically 20X more energy efficient for certain Al and HPC workloads than traditional CPUs 典 Plan to achieve & maintain 100% renewable electricity for our operations and data centers by FY25 and annually thereafter A Place For People To Do Their Life's Work glassdoor BEST PLACES TO WORK 203 "100 Best Companies to Work For" FORTUNE "America's Most Just Companies" CNBC "Most Responsible Companies" NEWSWEEK "Best Places to Work for LGBT Equality" HUMAN RIGHTS CAMPAIGN Management Time Magazine's 100 Most Influential Companies Fast Company's Best Workplaces for Innovators Fortune's World's Most Admired Companies Wall Street Journal's Management Top 250 All-Stars Corporate Governance 43% of Board is Gender, Racially, or Ethnically Diverse 93% of Directors are independent NVIDIA#58Reconciliation of Non-GAAP to GAAP Financial Measures#59Q3 FY24 Reconciliation of Non-GAAP to GAAP Financial Measures Non-GAAP Acquisition-Related and Other Costs (A) Stock-Based Compensation (B) IP-Related Costs Other (C) Tax Impact of Adjustments GAAP $13,583 (119) (38) (26) Gross margin ($ in million) 75.0% (0.7) (0.2) (0.1) $13,400 74.0% Operating income $11,557 (135) (979) (26) $10,417 ($ in million) Net income $10,020 (135) (979) (26) (70) 433 $9,243 ($ in million) Shares used in diluted per share calculation 2,494 (millions) Diluted EPS $4.02 A. Consists of amortization of intangible assets and transaction costs. B. Stock-based compensation charge was allocated to cost of goods sold, research and development expense, and sales, general and administrative expense. C. Other represents net losses from non-affiliated investments and interest expense related to amortization of debt discount 2,494 $3.71 NVIDIA#60Reconciliation of Non-GAAP to GAAP Financial Measures (contd.) Acquisition-Related Gross Margin Non-GAAP and Other Costs (A) Stock-Based Compensation (B) IP-Related Costs GAAP Q3 FY 2023 56.1% (2.0) (0.5) 53.6% Q4 FY 2023 66.1% (2.0) (0.5) (0.3) 63.3% Q1 FY 2024 66.8% (1.7) (0.4) (0.1) 64.6% Q2 FY 2024 71.2% (0.9) (0.2) A. Consists of amortization of intangible assets B. Stock-based compensation charge was allocated to cost of goods sold 70.1% NVIDIA#61Reconciliation of Non-GAAP to GAAP Financial Measures (contd.) Gross Margin ($ in Millions & Margin Percentage) Non-GAAP Acquisition-Related and Other Costs (A) Stock-Based Compensation (B) IP-Related Costs GAAP $7,233 (27) (35) $7,171 FY 2019 61.7% (0.2) (0.3) 61.2% $6,821 (39) (14) $6,768 FY 2020 62.5% (0.4) (0.1) 62.0% $10,947 (425) (88) (38) $10,396 FY 2021 65.6% (2.6) (0.5) (0.2) 62.3% $17,969 (344) (141) (9) $17,475 FY 2022 66.8% (1.4) (0.5) 64.9% $15,965 (455) (138) (16) $15,356 FY 2023 59.2% (1.7) (0.5) (0.1) 56.9% $11,966 (335) (108) YTD Q3 2023 57.2% (1.6) (0.5) $11,523 55.1% $28,000 (358) (96) (36) $27,510 YTD Q3 2024 72.1% (0.9) (0.2) (0.1) 70.9% A. Consists of amortization of intangible assets and inventory step-up B. Stock-based compensation charge was allocated to cost of goods sold NVIDIA#62Acquisition-Related Non-GAAP Acquisition Termination Cost (A) Reconciliation of Non-GAAP to GAAP Financial Measures (contd.) Operating Income and Margin ($ in Millions & Margin Percentage) Stock-Based IP-Related Other and Other Costs Compensation GAAP Costs (C) (B) $4,407 (2) (557) (35) (9) $3,804 FY 2019 37.6% (4.7) (0.3) (0.1) 32.5% $3,735 (31) (844) (14) $2,846 FY 2020 34.2% (0.3) (7.7) (0.1) 26.1% $6,803 (836) (1,397) (38) $4,532 FY 2021 40.8% (5.0) (8.4) (0.2) 27.2% $12,690 (636) (2,004) (9) $10,041 FY 2022 47.2% (2.5) (7.4) 37.3% $9,040 (1,353) (674) (2,710) (16) (63) $4,224 FY 2023 33.5% (5.0) (2.5) (10.0) (0.1) (0.2) 15.7% $6,816 (1,353) (499) (1,971) (25) $2,968 YTD Q3 2023 32.6% (6.5) (2.4) (9.4) (0.1) 14.2% $22,385 (446) (2,555) (36) 10 $19,358 YTD Q3 2024 57.7% (1.1) (6.6) (0.1) 49.9% A. Consists of amortization of acquisition-related intangible assets, inventory step-up, transaction costs, compensation charges, and other costs B. Stock-based compensation charge was allocated to cost of goods sold, research and development expense, and sales, general and administrative expense C. Comprises of legal settlement cost, contributions, restructuring costs and assets held for sale related adjustments NVIDIA#63Reconciliation of Non-GAAP to GAAP Financial Measures (contd.) ($ in Millions) Free Cash Flow FY 2019 $3,143 Purchases Related to Property and Equipment and Intangible Assets Principal Payments on Property and Equipment and Intangible Assets Net Cash Provided by Operating Activities 600 FY 2020 $4,272 489 $3,743 $4,761 FY 2021 $4,677 1,128 17 $5,822 FY 2022 $8,049 976 83 $9,108 FY 2023 $3,750 1,833 58 $5,641 YTD Q3 2023 $2,015 1,324 54 $3,393 YTD Q3 2024 $15,732 815 44 $16,591 NVIDIA#64($ in Millions) Reconciliation of Non-GAAP to GAAP Financial Measures Non-GAAP gross margin Impact of stock-based compensation expense, acquisition-related costs, and other costs GAAP gross margin Non-GAAP operating expenses Q4 FY24 Outlook 75.5% (1.0%) 74.5% $2,200 Impact of stock-based compensation expense, acquisition-related costs, and other costs 965 GAAP operating expenses $3,165 NVIDIA#65NVIDIA