Q1 2025 NVIDIA Corp Earnings Call

Regina: Good afternoon. My name is Regina, and I will be your conference operator today. At this time, I would like to welcome everyone to NVIDIA's first quarter earnings call. All lines have been placed on mute to prevent any background noise.

Good afternoon. My name is Regina and I will be your conference operator today at this time I would like to welcome everyone doing videos first quarter earnings call. All lines have been placed on mute to prevent any background noise. After the speakers' remarks, there will be a question and answer session. If you would like to ask a question. During this time.

Regina: After the speaker's remarks, there will be a question and answer session. If you would like to ask a question during this time, simply press star, followed by the number one on your telephone keypad. If you would like to withdraw your question, press star one again. Thank you. Simona Jankowski, you may begin your conference.

Simply press Star followed by the number one on your telephone keypad. If you would like to withdraw your question Press Star. One again. Thank you Simona Jankovskis you may begin your conference.

Simona Jankowski: Thank you. Good afternoon, everyone, and welcome to NVIDIA's conference call for the first quarter of fiscal 2025. With me today from NVIDIA are Jensen Huang, President and Chief Executive Officer, and Colette Kress, Executive Vice President and Chief Financial Officer. I'd like to remind you that our call is being webcast live on NVIDIA's Investor Relations website. The webcast will be available for replay until the conference call to discuss our financial results for the second quarter of fiscal 2025. The content of today's call is NVIDIA's property. It can be reproduced or transcribed without prior written consent.

Simona Jankowski: Thank you good afternoon, everyone and welcome to video conference call for the first quarter of fiscal 2025 with me today from Nvidia are Jensen, Huang President and Chief Executive Officer, and Colette, Kress Executive Vice President and Chief Financial Officer.

Speaker Change: Like to remind you that our call is being webcast live on <unk> Investor Relations website. The webcast will be available for replay until the conference call to discuss our financial results for the second quarter of fiscal 2025, the contents of today's call isn't <unk> property, it can't be reproduced or transcribed without our prior written consent.

Simona Jankowski: During this call, we may make forward-looking statements based on current expectations. However, these are subject to a number of significant risks and uncertainties, and our actual results may differ materially. For a discussion of factors that could affect our future financial results and business, please refer to the disclosure in today's earnings release, our most recent Forms 10-K and 10-Q, and the reports that we may file on Form 8-K with the Securities and Exchange Commission.

Speaker Change: During this call we may make forward looking statements based on current expectations. These are subject to a number of significant risks and uncertainties and our actual results may differ materially for a discussion of factors that could affect our future financial results and business. Please refer to the disclosure in today's earnings release, our most recent forms 10-K and 10-Q and.

Speaker Change: The reports that we may file on form 8-K, with the Securities and Exchange Commission.

Simona Jankowski: All our statements are made as of today, May 22, 2024, based on information currently available to us. Except as required by law, we assume no obligation to update any such statements. During this call, we will discuss non-GAAP financial measures. You can find a reconciliation of these non-GAAP financial measures to GAAP financial measures in our CFO commentary, which is posted on our website.

Speaker Change: All our statements are made as of today may 22nd 2024 based on information currently available to us.

Speaker Change: Except as required by law, we assume no obligation to update any such statements. During this call. We will discuss non-GAAP financial measures you can find a reconciliation of these non-GAAP financial measures to GAAP financial measures in our CFO commentary, which is posted on our website.

Simona Jankowski: Let me highlight some upcoming events. On Sunday, June 2nd, ahead of the Computex Technology Trade Show in Taiwan, Jensen will deliver a keynote which will be held in person in Taipei, as well as streamed live. And on June 5th, we will present at the Bank of America Technology Conference in San Francisco. With that, I will turn the call over to Collette.

Speaker Change: Let me highlight some upcoming events on Sunday June 2nd ahead of the Computex technology trade show in Taiwan Jensen will deliver a keynote which will be held in person in Taipei as well as streamed life and on June 5th we will present at the Bank of America Technology Conference in San Francisco with that let me turn the call over to Colette.

Colette M. Kress: Thanks, Simona Q1 was another record quarter revenue of 26 billion was up 18% sequentially and up 262% year on year and well above our outlook of 24 billion.

Colette M. Kress: Q1 was another record quarter. Revenue of $26 billion was up 18% sequentially and up 262% year-on-year and well above our outlook of $24 billion. Starting with Datacenter. Datacenter revenue of $22.6 billion was a record, up 23% sequentially and up 427% year-on-year, driven by continued strong demand for the NVIDIA Hopper GPU computing platform. Compute revenue grew more than 5x, and networking revenue more than 3x from last year. Strong sequential data center growth was driven by all customer types, led by enterprise and consumer internet companies.

Starting with data Center data center revenue of $22 6 billion was a record up 23% sequentially and up 427% year on year driven by continued strong demand for the Nvidia Hopper GPU computing platform.

Colette M. Kress: <unk> revenue grew more than five bucks and networking revenue more than three X from last year.

Colette M. Kress: <unk> sequential data center growth was driven by all customer types led by enterprise and consumer Internet company large cloud providers continue to drive strong growth as they deploy and rep Nvidia AI infrastructure at scale and represented the mid Forty's.

Colette M. Kress: Large cloud providers continue to drive strong growth as they deploy and ramp NVIDIA AI infrastructure at scale, and represent the mid-40s as a percentage of our data center revenue. Training and inferencing AI on NVIDIA CUDA is driving meaningful acceleration in cloud rental revenue growth, delivering an immediate and strong return on cloud providers' investments. For every $1 spent on NVIDIA AI infrastructure, cloud providers have an opportunity to earn $5 in GPU Instant Hosting revenue over four years.

Colette M. Kress: As a percentage of our data center revenue.

Colette M. Kress: Training and inferencing AI on Nvidia Cuda is driving meaningful acceleration in cloud rental revenue growth delivering an immediate and strong return on cloud providers investment.

For every $1 spent on Nvidia AI infrastructure cloud providers have an opportunity to earn $5 in GPU instant hosting revenue over four years.

Colette M. Kress: NVIDIA's rich software stack and ecosystem and tight integration with cloud providers make it easy for end customers to get up and running on NVIDIA GPU instances in the public cloud. For cloud rental customers, NVIDIA GPUs offer the best time-to-train models, the lowest cost-to-train models, and the lowest cost-to-inference large-language models.

Colette M. Kress: And video software stock and ecosystem and tight integration with cloud providers. It makes it easy for end customers up and running on Nvidia GPU instances in the public cloud.

Colette M. Kress: For cloud rental customers Nvidia Gpus offer the best time to train models, the lowest cost to train model and the lowest cost to infringe large language models.

Colette M. Kress: For public cloud providers, NVIDIA brings customers to their cloud, driving revenue growth and returns on their infrastructure investment. Leading LLM companies such as OpenAI, ADEPT, Anthropic, Character AI, Cohere, Databricks, DeepMind, Meta, Mistral, XAI, and many others are building on NVIDIA AI in the cloud. Enterprises drove strong sequential growth in data centers this quarter.

Colette M. Kress: For public cloud providers and video brings customers to their cloud driving revenue growth and returns on their infrastructure investments.

Colette M. Kress: Leading all of them companies such as opening I adopt anthropic character AI co. Her data breaks deep mine Nutter Mistral X AI and many others are building on Nvidia AI in the cloud.

Colette M. Kress: Enterprises drove strong sequential growth in data center. This quarter. We supported tussle is expansion of their training AI cluster to 35000, H 100 G. P is their use of Nvidia AI infrastructure paved the way for the breakthrough performance of FSD version.

Colette M. Kress: We supported Tesla's expansion of their training AI cluster to 35,000 H100 GPUs. Their use of NVIDIA AI infrastructure paved the way for the breakthrough performance of FSD version 12, their latest autonomous driving software based on vision. Video transformers, while consuming significantly more computing, are enabling dramatically better autonomous driving capabilities and propelling significant growth for NVIDIA AI infrastructure across the automotive industry. We expect automotive to be our largest enterprise vertical within the data center this year, driving a multi-billion revenue opportunity across on-premises and cloud consumption. Consumer internet companies are also a strong growth vertical.

Colette M. Kress: 12, their latest autonomous driving software based on vision.

Colette M. Kress: Transformers Wow consuming significantly more computing are able dramatically better autonomous driving capability and propelled significant growth for Nvidia AI infrastructure across the automotive industry.

Colette M. Kress: We expect automotive to be our largest enterprise vertical within data center. This year, driving a multibillion revenue opportunity across on Prem and cloud consumption.

Colette M. Kress: Consumer Internet companies are also a strong growth vertical the big highlight this quarter was met us announcement of Lama Threep.

Colette M. Kress: A big highlight this quarter was Meta's announcement of LLAMA3, their latest large language model, which was trained on a cluster of 24,000 H100 GPUs. LLAMA3 Powers, Meta AI, a new AI assistant available on Facebook, Instagram, WhatsApp, and Messenger. Lama3 is openly available and has kicked-started a wave of AI development across industries. As generative AI makes its way into more consumer internet applications, we expect to see continued growth opportunities as inference scales both with model complexity, as well as with the number of users and the number of queries per user, driving much more demand for AI compute. In the trailing four quarters, we estimate that inference drove about 40% of our data center revenue. Both training and inference are growing significantly.

Colette M. Kress: This large language model, which was trained.

Colette M. Kress: Of 24800 Gpus.

Colette M. Kress: Of the three powers meta AI, a new AI assistant available on Facebook, Instagram Whatsapp and messenger llama.

<unk> openly available and has kick started a wave of AI development across industries.

Colette M. Kress: Turning to my eye it makes its way into more consumer Internet applications, we expect to see continued growth opportunities as infringe scales, both with model complexity as well as with the number of users and number of queries per user driving much more demand for AI compute.

Colette M. Kress: And our trailing four quarters, we estimate that infringe drove about 40% of our data center revenue both training and inference are growing significantly.

Colette M. Kress: Large clusters, like the ones built by Meta and Tesla, are examples of the essential infrastructure for AI production, what we refer to as AI factories. These next-generation data centers host advanced, full-stack, accelerated computing platforms where data comes in, and intelligence comes out. In Q1, we worked with over 100 customers building AI factories ranging in size from hundreds to tens of thousands of GPUs, with some reaching 100,000 GPUs. From a geographic perspective, data center revenue continues to diversify as countries around the world invest in sovereign AI. Sovereign AI refers to a nation's ability to produce artificial intelligence using its own infrastructure, data, workforce, and business network.

Colette M. Kress: Large clusters like the one built by meta and Tesla are examples of the essential infrastructure for AI production, what we referred to as AI factories.

Colette M. Kress: These next generation data centers hosted advanced Colestock accelerated computing platforms, where the data comes in and intelligence comes out.

Colette M. Kress: In Q1, we worked with over 100 customers building AI factories ranging in size.

<unk> from hundreds to tens of thousands of Gpus with some reaching 100000 Gpus.

Colette M. Kress: From a geographic perspective data center revenue continues to diversify as countries around the world invest in sovereign AI.

Colette M. Kress: Sovereign AI refers to a nation's capabilities to produce artificial intelligence using its own infrastructure data workforce and business networks Nations are building up domestic computing capacity through various models some are procuring and operating sovereign AI cloud.

Colette M. Kress: Nations are building up domestic computing capacity through various models. Some are procuring and operating sovereign AI clouds in collaboration with state-owned telecommunication providers or utilities. Others are sponsoring local cloud partners to provide a shared AI computing platform for public and private sector use. For example, Japan plans to invest more than $740 million in key digital infrastructure providers, including KDDI, Sakura, Internet, and SoftBank, to build out the nation's sovereign AI infrastructure.

Colette M. Kress: <unk> in collaboration with state owned telecommunication providers or utilities.

Colette M. Kress: Others are sponsoring local cloud partners to provide a shared AI computing platform for public and private sector use.

Colette M. Kress: For example, Japan plans to invest more than $740 million in key digital infrastructure providers, including <unk>, <unk> Internet and Softbank to build out the nations sovereign AI infrastructure.

Colette M. Kress: France-based Scaleway, a subsidiary of the Iliad Group, is building Europe's most powerful cloud-native AI supercomputer. In Italy, Swisscom Group will build the nation's first and most powerful NVIDIA DTX-powered supercomputer to develop the first LLM natively trained in the Italian language. And in Singapore, the National Supercomputer Center is getting upgraded with NVIDIA Opera GPUs, while Singtel is building NVIDIA's Accelerated AI factories across Southeast Asia. NVIDIA's ability to offer end-to-end compute to networking technologies, full-stack software, AI expertise, and a rich ecosystem of partners and customers allows sovereign AI companies and regional cloud providers to jumpstart their country's AI ambitions.

Colette M. Kress: France based scale way a subsidiary of the Elia Group is building Europe's most powerful cloud native AI supercomputer.

Colette M. Kress: In Italy Swiss Com group, we will build the nations first and most powerful and video Gtx powered supercomputer to develop the first of all natively trained in the Italian language.

Colette M. Kress: And then Singapore the National Supercomputer centre is getting upgraded with Nvidia offer Gpus, one thing Tal is building and videos accelerated AI factories across South East Asia.

Colette M. Kress: And videos ability to offer end to end compute networking technologies post tax software AI expertise and rich ecosystem of partners and customers allows sovereign AI and regional cloud providers to jumpstart third countries AI ambitions.

Colette M. Kress: From nothing the previous year, we believe sovereign AI revenue can approach the high single-digit billions this year. The importance of AI has caught the attention of every nation. We've ramped new products designed specifically for China that don't require an export control license. However, our data center revenue in China is down significantly from the level prior to the imposition of the new export control restrictions in October.

Colette M. Kress: For nothing.

Colette M. Kress: This year, we believe sovereign AI revenue can approach the high single digit billions this year.

Colette M. Kress: Important today I has caught the attention of every nation.

Colette M. Kress: We ramped new products designed specifically for China that don't require court control license our data center revenue in China is down significantly from the level prior to the imposition of the new export control restrictions in October.

Colette M. Kress: We expect the market in China to remain very competitive going forward. From a product perspective, the vast majority of compute revenue was driven by our Hopper GPU architecture. Demand for Hopper during the quarter continues to increase.

Colette M. Kress: We expect the market in China to remain very competitive going forward.

Colette M. Kress: From a product perspective, the vast majority of compute revenue was driven by our hopper.

Colette M. Kress: Architecture <unk>.

Colette M. Kress: A man for Hopper during the quarter continues to increase thanks to Cuda algorithm innovation, we've been able to accelerate.

Colette M. Kress: Thanks to CUDA algorithm innovations, we've been able to accelerate LLM inference on H-100 by up to 3x, which can translate to a 3x cost reduction for serving popular models like LLAMA3. We started sampling the H200 in Q1 and are currently in production with shipments on track for Q2. The first H200 system was delivered by Jensen to Sam Altman and the team at OpenAI, and Powered, their amazing GPT 4.0 demos last week. H-200 nearly doubles the inference performance of H-100, delivering significant value for production deployment.

Colette M. Kress: In France on H, one hundreds by up to three X, which can translate to a three X cost reduction for serving popular models like Lam with Reed.

Colette M. Kress: We started sampling the H 200 in a in Q1 and are currently in production with shipments on track for Q2. The first H 200 system was delivered by Johnson to Sam Hoffman and the team at open AI and.

Colette M. Kress: Empowered their amazing GPT forero demos last week.

Colette M. Kress: H 200, nearly doubles the inference performance of H 100 <unk>.

Colette M. Kress: <unk> significant value for production deployments for.

Colette M. Kress: For example, using Vama3 with 700 billion parameters, a single NVIDIA HGX H200 server can deliver 24,000 tokens per second, supporting more than 2,400 users at the same time. That means for every $1 spent on NVIDIA HTX H200 servers at current prices per token, an API provider serving LLAMA-3 tokens can generate $7 in revenue over four years. With ongoing software optimizations, we continue to improve the performance of NVIDIA AI infrastructure for serving AI models, while supply for H-100 to prove we are still constrained on H200. At the same time, Blackwell is in full production.

Colette M. Kress: For example, using bomber III, which with 700 billion parameters a single Nvidia <unk> H 200 server can deliver 24000 tokens per second supporting more than 2400 users at the same time.

Colette M. Kress: That means for every $1 spent on Nvidia <unk> H 200 servers at current prices per token.

Colette M. Kress: And API provider, serving Lama, three tokens can generate $7 and revenue over four years with.

Colette M. Kress: With ongoing software optimizations, we continue to improve the performance of Nvidia AI infrastructure for serving AI models.

Colette M. Kress: While supply for <unk> 100.

Colette M. Kress: We are still constrained on H 200 at the same time Blackwell is in full production.

Colette M. Kress: We are working to bring our system and cloud partners for global availability later this year. Demand for H200 and Blackwell is well ahead of supply, and we expect demand may exceed supply well into next year. The Grace Hopper Superchip is shipping in volume.

Colette M. Kress: We're working to bring up our system and cloud partners for global availability later this year demand for H 200, and Blackwell is well ahead of supply and we expect demand may exceed supply well into next year.

Colette M. Kress: Grace Hopper Super Chip is shipping in volume last week at the International Supercomputing Conference, we announced that nine new supercomputers worldwide are using Grace Hopper for combined 200 extra plops of energy efficient AI processing power.

Colette M. Kress: Last week at the International Supercomputing Conference, we announced that nine new supercomputers worldwide are using Grace Hopper for a combined 200 exaplots of energy-efficient AI processing power delivered this year. These include the ALPS supercomputer at the Swiss National Supercomputing Center, the fastest AI supercomputer in Europe. Zombard AI at the University of Bristol in the UK and Jupyter at the Julek Supercomputing Center in Germany. We are seeing an 80% attach rate of grace to hopper in supercomputing due to its high energy efficiency and performance. We are also proud to see supercomputers powered by Grace Hopper take the number one, the number two, and the number three spots of the most energy efficient supercomputers in the world.

Colette M. Kress: Delivered this year.

Colette M. Kress: These include the Alps supercomputer at the Swiss National Supercomputing Center, the fastest AI supercomputer in Europe.

Colette M. Kress: Hassan Bart AI at the University of Bristol in the UK and.

Colette M. Kress: In Jupiter and the ULA Supercomputing center in Germany.

Colette M. Kress: We are seeing an 80% attach rate of Grace Hopper and supercomputing due to its high energy efficiency and performance.

Colette M. Kress: We are also proud to see supercomputers powered with Grace Hopper take the number one.

Colette M. Kress: Number two and the number three spots of the most energy efficient supercomputers in the world.

Colette M. Kress: Strong networking year on year growth was driven by Infiniband, we experienced a modest sequential decline, which was largely due to the timing of supply with demand well ahead of what we were able to shop.

Colette M. Kress: Strong networking year-on-year growth was driven by InfiniBand. However, we experienced a modest sequential decline, which was largely due to the timing of supply, with demand well ahead of what we were able to ship. We expect networking to return to sequential growth in Q2. In the first quarter, we started shipping our new Spectrum X Ethernet networking solution, Optimized for AI, from the ground up. It includes our Spectrum 4 switch, BlueField-3 DPU, and new software technologies to overcome the challenges of AI on Ethernet to deliver 1.6x higher networking performance for AI processing compared with traditional Ethernet.

Colette M. Kress: We expect networking to return to sequential growth in Q2.

Colette M. Kress: In the first quarter, we started shipping our new spectrum ex Ethernet networking solution optimized for AI from the ground up that includes our spectrum for switch Bluefield, three GPU and new software technologies to overcome the challenges of AI on Ethernet to deliver one.

Colette M. Kress: Six X higher networking performance for AI processing compared with traditional Ethernet.

Colette M. Kress: Spectrum X is ramping in volume with multiple customers, including a massive 100,000 GPU plus. It opens a brand new market for NVIDIA networking and enables Ethernet-only data centers to accommodate large-scale AI. We expect Spectrum X to jump to a multi-billion dollar product line within a year. At GTC in March, we launched our next generation AI factory platform, Blackwell. The Blackwell GPU architecture delivers up to 4x faster training and 30x faster inference than the H100 and enables real-time generative AI on trillion-parameter large-language models.

Colette M. Kress: Spectrum <unk> ramping in volume with multiple customers, including a massive 100000 GPU cluster.

Colette M. Kress: Spectrum X opens a brand new market to Nvidia networking and enables Ethernet only data centers to accommodate large scale AI.

Colette M. Kress: We expect spectrum X two jumped to a multibillion dollar product line within a year.

Colette M. Kress: Blackwell is a giant leap with up to 25x lower TCO and energy consumption than Hopper. The Blackwell platform includes the fifth-generation NVLink with a multi-GPU spine and new InfiniBand and Ethernet switches, the X800 series, designed for trillion-parameter scale AI. Blackwell is designed to support data centers universally from hyperscale to enterprise, training to inference, x86 to great CPUs, Ethernet to InfiniBand networking, and air cooling to liquid cooling.

Colette M. Kress: At GTC in March we launched our next generation AI factory platform Blackwell.

Colette M. Kress: The Blackwater GPU architecture delivers up to four X faster training and 30 X faster inference than the H 100, and enable real time generative AI on Trillium parameter of barge language models.

Colette M. Kress: Blackwell is a giant leap with up to 25 X lower Tcl and energy consumption than hopper.

Colette M. Kress: The Blackwell platform includes the fifth generation and dealing with a multi GPU spine and new Infiniband and Ethernet switches. The <unk> 800 series designed for a trillion parameter scale AI.

Colette M. Kress: Blackwell is designed to support Datacenters universally from Hyperscale to enterprise training to infringe X 86 degrees Cpus Ethernet to Infiniband networking and air cooling the liquid cooling.

Colette M. Kress: Blackwell will be available in over 100 OEM and ODM systems at launch, more than double the number of hoppers launched, and representing every major computer maker in the world. This will support fast and broad adoption across customer types, workloads, and data center environments in the first year of shipment. Blackwell time to market customers include Amazon, Google, Meta, Microsoft, OpenAI, Oracle, Tesla, and XAI. We announced a new software product with the introduction of NVIDIA Inference Microservices, or NIMS.

Colette M. Kress: Blackwell will be available in over 100, OEM and ODM systems at large more than double the number of hawkers launch and representing every major computer maker in the world.

Colette M. Kress: This will support fast and broad adoption across the customer types workloads and data center environments in the first year of shipments.

Colette M. Kress: Blackwell time to market customers include Amazon Google Microsoft.

Colette M. Kress: Microsoft opening our Oracle Tesla and <unk>.

Colette M. Kress: We announced a new software product with the introduction of Nvidia and friends Micro services core NIM NIM.

Colette M. Kress: NIMS provides secure and performance-optimized containers powered by NVIDIA CUDA acceleration for network computing and inference software, including Triton Inference Server and TensorRT LLM with industry-standard APIs for a broad range of use cases, including large language models for text, speech, imaging, vision, robotics, genomics, and digital biology. They enable developers to quickly build and deploy generative AI applications using leading models from NVIDIA, AI21, Adept, Cohere, Steady Images, and Shutterstock and open models from Google, Hugging Face, Meta, Microsoft, Mistral AI, Snowflake, and Stability AI. NIMS will be offered as part of our NVIDIA AI enterprise software platform for production deployment in the cloud or on-prem, Moving to gaming and Gaming revenue of $2.65 billion was down 8% sequentially and up 18% year-on-year, consistent with our outlook for a seasonal decline.

Colette M. Kress: <unk> provides secure and performance optimized containers powered by Nvidia Cuda acceleration and network computing and inference software, including Triton and print server and tensor RT LLM with industry standard Apis for broad range of use cases, including large language.

Colette M. Kress: Model for text speech imaging vision robotics genomics and digital biology.

Colette M. Kress: They involve enable developers to quickly build and deploy generally AI applications using leading models from Nvidia AI 'twenty, one adapt cohere Getty images and Shutterstock.

Colette M. Kress: And open models from Google Hugging face better Microsoft Mr. All AI snowflake and stability AI.

Colette M. Kress: Nims will be offered as part of our Nvidia AI enterprise software platform for production deployment in the cloud or on Prem.

Colette M. Kress: Moving to gaming and AI Pcs gaming revenue of $2 six 5 billion was down 8% sequentially and up 18% year on year consistent with our outlook for a seasonal decline the G force our checks supers Gpus market reception is strong.

Colette M. Kress: <unk> and end demand and channel inventory remained healthy across the product range.

Colette M. Kress: The GeForce RTX Super's GPU market reception is strong, and end-demand and channel inventory remain healthy across the product range. From the very start of our AI journey, we equipped GeForce RTX GPUs with CUDA Tensor Cores. Now with over 100 million installed base, GeForce RTX GPUs are perfect for gamers, creators, AI enthusiasts, and offer unmatched performance for running generative AI applications on PCs. NVIDIA has a full technology stack for deploying and running fast and efficient generative AI inference on GeForce RTX-powered PCs.

Colette M. Kress: From the very start of our AI journey, we equipped G Force <unk> Gpus with Cuda tensor cores now with over 100 million of an installed base G Force RPX Gpus are perfect for gamers creators AI enthusiasts and offer unknown.

Colette M. Kress: <unk> performance, we're running generally today our applications on Pcs.

Colette M. Kress: Nvidia has full technology stack for deploying and running fast and efficient generative AI infringe on G Force Rtr's Pcs.

Colette M. Kress: TensorRT LLM now accelerates Microsoft's PHY3 mini model and Google's Gemma 2B and 7B models, as well as popular AI frameworks, including LangChain and Llama Index. Yesterday, NVIDIA and Microsoft announced AI performance optimizations for Windows to help run LLMs up to 3x faster on NVIDIA, GeForce, RTX, and AI PCs. And top game developers, including NetEase Games, Tencent, and Ubisoft, are embracing NVIDIA Avatar Character Engine to create lifelike avatars to transform interactions between gamers and non-playable characters. Moving to ProViz.

Colette M. Kress: Tensor RT LLM now accelerates, Microsoft Phase III, many model and Google's Jemma, <unk> and 70 models as well as popular AI frameworks, including Lane change and Lama Index.

Colette M. Kress: Yesterday, Nvidia and Microsoft announced AI performance optimizations for Windows to help run all of them up to three X faster on Nvidia G Force RPX AI T seats.

Colette M. Kress: And top game developers, including <unk> games, Tencent and UV software embracing Nvidia Avatar character engine to create lifelike avatars to transform interactions between gamers and non playable characters.

Colette M. Kress: Moving to progress.

Colette M. Kress: Revenue of $427 million was down 8% sequentially and up 45% year-on-year. We believe generative AI and Omniverse industrial digitalization will drive the next wave of professional visualization growth. At GTC, we announced new Omniverse cloud APIs to enable developers to integrate Omniverse industrial digital twin and simulation technologies into their applications. Some of the world's largest industrial software makers are adopting these APIs, including Ansys, Cadence, 3DXcite, Dassault Systems, Brand, and Siemens, and developers can use them to stream industrial digital twins to spatial computing devices such as Apple Vision Pro. Omniverse Cloud APIs will be available on Microsoft Azure later this year.

Colette M. Kress: Revenue of $427 million was down 8% sequentially and up 45% year on year, we believe generative AI and omni versus industrial Digitalization will drive the next wave of professional visualization growth.

Colette M. Kress: At GTC, we announced new omni versus cloud Apis to enable developers to integrate omni versus industrial digital twin and simulation technologies into their applications. Some of the world's largest industrial software makers are adopting these API, including answers cadence really excite.

Colette M. Kress: <unk> systems brand and Siemens and.

Colette M. Kress: And developers can use them to stream industrial digital twins spatial computing devices, such as Apple vision CRO.

Colette M. Kress: <unk> cloud <unk> will be available on Microsoft Azure later this year.

Colette M. Kress: Companies are using Omniverse to digitalize their workflows. Omniverse-powered digital twins enable Wistron, one of our manufacturing partners, to reduce end-to-end production cycle times by 50% and defect rates by 40%, and BYD, the world's largest electric vehicle maker, is adopting Omniverse for virtual factory planning and retail configuration. Moving to automotive, revenue was $329 million, up 17% sequentially and up 11% year-on-year. Sequential growth was driven by the ramp of AI cockpit solutions with global OEM customers and strength in our self-driving platforms.

Colette M. Kress: Companies are using omni versus to digitalize, the workflows omni versus powered digital twins enable with strong one of our manufacturing partners to reduce end to end production cycle times by 50% and defect rates by 40%.

Colette M. Kress: <unk> the world's largest electric vehicle maker is adopting omni one for virtual factory planning and retail configurations.

Colette M. Kress: Moving to automotive revenue was $329 million up 17% sequentially and up 11% year on year sequential growth was driven by the ramp of AI cockpit solutions with global OEM customers and strength in our self driving platforms.

Colette M. Kress: Year-on-year growth was driven primarily by self-driving. We supported Xiaomi in the successful launch of its first electric vehicle, the SU-7 sedan built on the NVIDIA DRIVE O-RAN, our AI car computer for software-defined AV blades. We also announced a number of new design wins on NVIDIA DRIVE Thor, the successor to O-RAN, powered by the new NVIDIA Blackwell architecture, with several leading EV makers, including BYD, Xpeng, DACs, Ion Hyper, and Neural. DriveThough is slated for production vehicles starting next year.

Colette M. Kress: Year on year growth was driven primarily by self driving.

We supported Xiaomi and the successful launch of its first electric vehicle. The <unk> sedan built on the Nvidia drive Oren, our AI car computer for software defined.

Colette M. Kress: We also announced a number of new design wins on Nvidia drive Thor the successor to own powered by the new in video Blackwell architecture, with several leading leading EV makers, including BYD X 10, tacs ion hyper anoro.

Colette M. Kress: Drive thorough is slated for production vehicles, starting next year.

Colette M. Kress: Okay, moving to the rest of the P&L. GAAP gross margin expanded sequentially to 78.4% and non-GAAP gross margins to 78.9% on lower inventory charts. As noted last quarter, both Q4 and Q1 benefited from favorable component costs. However, sequentially, GAAP operating expenses were up 10%, and non-GAAP operating expenses were up 13%, primarily reflecting higher compensation-related costs and increased compute and infrastructure investment. In Q1, we returned $7.8 billion to shareholders in the form of share repurchases and cash dividends. Today, we announced a 10-for-1 split of our shares, with June 10th as the first day of trading on a split-adjusted basis.

Speaker Change: Okay moving to the rest of the P&L GAAP gross margin expanded sequentially to 78, 4% and non-GAAP gross margins to 78, 9% on lower inventory charges as noted last quarter, both Q4, and Q1 benefited from favorable component cost.

Speaker Change: Sequentially GAAP operating expenses were up 10% and non-GAAP operating expenses were up 13%, primarily reflecting higher compensation related costs and increased compute and infrastructure investments.

Speaker Change: In Q1, we returned $7 8 billion to shareholders in the form of share repurchases and cash dividends today, we announced a 10 for one split of our shares with June 10th as the first day of trading on a split adjusted basis. We are also increasing our dividend by 150%.

Colette M. Kress: We are also increasing our dividend by 150%. Now, let me turn to the outlook for the second quarter. Total revenue is expected to be $28 billion, plus or minus 2%. We expect sequential growth in all markets. GAAP and non-GAAP gross margins are expected to be 74.8% and 75.5%, respectively, plus or minus 50 basis points, consistent with our discussion last quarter. For the full year, we expect gross margins to be in the mid-70s percent range.

Speaker Change: Let me turn to the outlook for.

Speaker Change: For the second quarter.

Speaker Change: Total revenue is expected to be $28 billion, plus or minus 2%, we expect sequential growth in all market platforms.

Speaker Change: GAAP and non-GAAP gross margins are expected to be 74, 8% and 75, 5%, respectively, plus or minus 50 basis points consistent with our discussion last quarter for the full year, we expect gross margins to be in the mid 70% range.

Colette M. Kress: GAAP and non-GAAP operating expenses are expected to be approximately $4 billion and $2.8 billion, respectively. Full-year OpEx is expected to grow in the low 40% range. GAAP and non-GAAP other income and expenses are expected to be an income of approximately 700,000 excuse me, of approximately 300 million excluding gains and losses from non-affiliated investment. GAAP and non-GAAP tax rates are expected to be 17% plus or minus 1% excluding any discrete items. Further financial details are included in the CFO commentary and other information available on our IR website. I would now turn it over to Jensen, if he would like to make a comment.

GAAP and non-GAAP operating expenses are expected to be approximately $4 billion and $2 8 billion respectively.

Speaker Change: Full year Opex is expected to grow in the low 40% range.

Speaker Change: non-GAAP other income and expenses are expected to be an income of approximately $700.

Speaker Change: Excuse me of approximately $300 million, excluding gains and losses from non affiliated investments.

Speaker Change: GAAP and non-GAAP tax rates are expected to be 17% plus or minus 1%, excluding any discrete items.

Speaker Change: Further financial details are included in the CFO commentary and other information available on our IR website.

Speaker Change: I would like to now turn it over to Jensen as he would like to make a few comments.

Jensen Huang: Thanks, Colette. The industry is going through a major change. Before we start Q&A, let me give you some perspective on the importance of the transformation. The next industrial revolution has begun. Companies and countries are partnering with NVIDIA to shift the trillion-dollar installed base of traditional data centers to Accelerated Computing and build a new type of data center, the AI Factor, to produce a new commodity. AI will bring significant productivity gains to nearly every industry and help companies be more cost and energy efficient while expanding revenue opportunities. CSPs were the first generative AI movers. With NVIDIA, they accelerated workloads to save money and power.

Speaker Change: Thanks Colette.

Jensen Huang: The industry is going through a major change.

Jensen Huang: Before we start Q&A, let me give you some perspective on the importance of the transformation.

Jensen Huang: The next industrial Revolution has begun.

Jensen Huang: Companies and countries are partnering with Nvidia to shift.

Jensen Huang: The trillion dollar installed base of traditional data centers.

Jensen Huang: Two accelerated computing.

Jensen Huang: And build a new type of data center.

Jensen Huang: Factories.

Jensen Huang: To produce a new commodity.

Jensen Huang: Artificial intelligence.

Jensen Huang: Hey, I will bring significant productivity gains to nearly every industry and.

Jensen Huang: And help companies be more cost and energy efficient.

Jensen Huang: While expanding revenue opportunities.

Jensen Huang: Csp's, where their first generative AI movers.

Jensen Huang: With Nvidia Csp's accelerated workloads to save money empower.

Jensen Huang: The tokens generated by NVIDIA Hopper drive revenues for their AI services, and NVIDIA Cloud instances attract rental customers from our rich ecosystem of developers. Strong and accelerating demand for generative AI training and inference on the Hopper platform propels our data center growth. Training continues to scale as models learn to be multi-modal.

Jensen Huang: The tokens generated by Nvidia Hopper drive revenues for their AI services.

Jensen Huang: And video cloud instances attract rental customers from a rich ecosystem of developers.

Jensen Huang: Strong and accelerated demand accelerating demand for generative AI training and inference on the Hopper platform propels our data center growth.

Jensen Huang: Training continues to scale as models learned to be multi modal.

Jensen Huang: Understanding Tech. Speech, Images, Video, and 3D, and learn to reason and plan. Our inference workloads are growing incredibly fast, with generative AI and friends, which is now about fast. Token generation at massive scale has become incredibly complex. Generative AI is driving a Full Stack Computing Platform Shift that will transform every computer interaction from today's information retrieval model.

Jensen Huang: Understanding text.

Jensen Huang: Beach images video and three D.

Jensen Huang: And learned to reason and plan.

Jensen Huang: Our inference workloads are growing incredibly.

Jensen Huang: With generative AI.

Jensen Huang: Inference.

Jensen Huang: Which is now about fast.

Jensen Huang: Token generation at massive scale has become.

Jensen Huang: Incredibly complex.

Jensen Huang: Generative AI is driving from foundation up.

Jensen Huang: Full stack computing platform shift that will transform every computer interaction.

Jensen Huang: From today's information retrieval model.

Jensen Huang: We are shifting to an answers and skills generation model of computing. AI will understand context and our intentions; be knowledgeable, reason, plan, and perform tasks.

Jensen Huang: We are shifting to an answers and skills generation model of computing.

Jensen Huang: AI will understand context, and our intentions.

Jensen Huang: Knowledgeable.

Speaker Change: Isn't plan and perform tests.

Jensen Huang: We are fundamentally changing how computing works and what computers can do, from general purpose CPUs to GPU-accelerated computing, from instruction-driven software to intention understanding models, from retrieving information to performing skills. And at the industrial level, from producing software to generating tokens. Manufacturing Digital Intelligence. Token generation will drive a multi-year build-out of AI factories. Beyond cloud service providers, generative AI has expanded to consumer internet companies and Enterprise, such as Sovereign AI, Animo, and HealthcareCustomerService, creating multiple multi-billion dollar vertical markets.

We are fundamentally changing how computing works.

Speaker Change: Computers can do.

Speaker Change: From general purpose CPU to GPU accelerated computing.

Speaker Change: From instruction driven software.

Two intention understanding models.

Speaker Change: From retrieving information to performing skills.

Speaker Change: And at the industrial level.

Speaker Change: From producing software.

Speaker Change: Two generating tokens manufacturing.

Manufacturing digital intelligence.

Token generation will drive a multi year buildout of AI factories.

Speaker Change: Beyond cloud service providers generative AI has expanded to consumer internet companies and.

Speaker Change: In enterprise <unk>.

Speaker Change: Sovereign AI <unk>.

Speaker Change: Automotive and healthcare customers.

Creating multiple multibillion dollar vertical markets.

Jensen Huang: The Blackwell platform is in full production, and forms the foundation for trillion-parameter-scale generative AI. The combination of great CPUs. Blackwell GPU, NVLink Quantum, Spectrum, Nixon Switch, High-Speed Interconnect, and a rich ecosystem of software and partners. These enable us to expand and offer a richer and more complete solution for AI factories than previous generations. Spectrum X opens a brand new market for us to bring large-scale AI to Ethernet-only data centers, and NVIDIA NIMS is our new software offering that delivers enterprise-grade, optimized, generative AI to run on CUDA everywhere, from the cloud to on-prem data centers to RTX AI PCs, through our expansive network of ecosystem partners. From Blackwell to Spectrum X to NIMS, we are poised for the next

Speaker Change: The Blackwell platform is in full production.

Speaker Change: And forms the foundation for Trillium parameters scaled generative AI.

Speaker Change: The combination of Grace CPU black.

Speaker Change: Blackwell Gpus.

Speaker Change: <unk> link.

Speaker Change: Quantum.

Spectrum mixed and switches.

Speaker Change: High speed Interconnects.

Speaker Change: Rich ecosystem of software and partners.

Speaker Change: Let us expand and offer a richer and more complete solution for AI factories than previous generations.

Speaker Change: Spectrum X opens a brand new market for us to bring large scale AI to Ethernet only data centers.

Speaker Change: And Nvidia Nims is our new software offering that delivers enterprise grade optimized generative AI to run on Cuda everywhere.

Speaker Change: From the cloud to on Prem data centers to our TX <unk>.

Speaker Change: Through our expansive network of ecosystem partners.

Speaker Change: From Blackwell to spectrum X two nims, we are poised for the next wave of growth.

Speaker Change: Thank you.

Simona Jankowski: Thank you, Jensen. We will now open the call for questions. Operator, could you please poll for questions? At this time, I would like to remind everyone, in order to ask a question, press star, then number one on your telephone keypad. We'll pause for just a moment to compile the Q&A list. As a reminder, please limit your questions to...

Thank you gentlemen, we will now open the call for questions. Operator could you. Please poll for questions.

Regina: At this time, I would like to remind everyone in order to ask a question, press star then the number one on your telephone keypad. We'll pause for just a moment to compile the Q&A roster. As a reminder, please limit yourself to one question. Your first question comes from the line of Stacy Raskon with Bernstein. Please go ahead. Hi guys!

Speaker Change: At this time I would like to remind everyone in order to ask a question Press Star then the number one on your telephone keypad well pause for just a moment to compile the Q&A roster. As a reminder, please limit yourself to one question.

Speaker Change: Your first question comes from the line of Stacy <unk> with Bernstein. Please go ahead.

Speaker Change: Okay.

Speaker Change: Hey, guys. Thanks for taking my questions.

Stacy Aaron Rasgon: My first one I wanted to drill a little bit into the Blackwell comments that it's in full production now.

Stacy Aaron Rasgon: What does that suggest with regard to shipments and delivery timing to get that product. It doesn't sound like it's sampling anymore.

Stacy Aaron Rasgon: What does it mean, when that's actually customers and if it's in production now.

Unknown Speaker: We will be shipping, well, we've been in production for a little bit of time, but our production shipments will start in Q2 and ramp up in Q3, and customers should have data centers stood up in Q4. Got it.

Speaker Change: We will be shipping, while we have been in production.

Speaker Change: For for a little bit of time.

Speaker Change: But.

Speaker Change: Our production shipments will start in Q2.

And ramped in Q3 and customers should have data centers stood up in Q4.

Speaker Change: Got it so this year, we will see Blackwell revenue it sounds like.

Speaker Change: We will see a lot of black while revenue this year.

Unknown Speaker: Our next question will come from the line of Timothy R. Curry with CBS. Please go ahead. Thanks a lot.

Unknown Speaker: So this year we will see Blackwell revenue. We will see a lot of Blackwell revenue this year. Our next question will come from the line of Timothy R. Curry with CBS. Please go ahead. Thanks a lot. I wanted to ask Jensen about the deployment of, you know, Blackwell versus Hopper, just given the system nature and

Speaker Change: Our next question will come from the line of Timothy Arcuri with UBS. Please go ahead.

Thanks, a lot I wanted to ask Jensen about the deployment of Blackwell versus Hopper, just given the system's nature and all of the demand for <unk> that you have how does the deployment of this stuff differ from Hopper I guess I asked because liquid cooling at scale hasn't been done before and there's some engineering challenges both at the node level and was.

Speaker Change: The data center. So do these complexities sorta elongate the transition and how do you sort of.

Speaker Change: Think about how that's all going thanks.

Jensen Huang: Yep, Blackwell comes in many configurations. Blackwell is a platform, not a graphics card. And the platform includes support for air-cooled, liquid-cooled x86, and Grace. InfiniBand, Now Spectrum X, and a very large NVLink domain that I demonstrated at GTC that I showed at GTC. And so, for some customers, they will ramp into their existing installation base of data centers that are already shipping hoppers. They will easily transition from H-100 to H-200 to B100.

Speaker Change: Yes.

Speaker Change: Blackwell comes in many configurations Blackwell is a platform.

Not a GPU and the platform includes support for air cooled liquid cooled.

Speaker Change: X 86 and Grace.

Speaker Change: Infiniband.

Speaker Change: Now spectrum X.

Speaker Change: And <unk>.

Speaker Change: Very large envy linked domain.

Speaker Change: That is demonstrated at GTC that show that GTC.

Speaker Change: And so for some customers.

Speaker Change: They will ramp.

Speaker Change: Into their existing installed base.

Speaker Change: Data centers that are already shipping hoppers, they will easily transitioned from each 100 to H 200.

Speaker Change: To be 100.

Jensen Huang: And so Blackwell systems have been designed to be backwards compatible, if you will, electrically, and mechanically, and of course, the software stack that runs on Hopper will run fantastically on Blackwell. We have also been priming the pump, if you will, with the entire ecosystem, getting them ready for liquid cooling. We've been talking to the ecosystem about Blackwell for quite some time, and the CSPs, the data centers, the ODMs, the system makers. And our supply chain beyond them, the cooling supply chain base, liquid cooling supply chain base, data center supply chain base.

Speaker Change: And so Blackwell systems have been designed to be backwards compatible if you will electrically mechanically.

Speaker Change: And of course, the software stack that runs on Hopper will run fantastically on on Blackwell.

Speaker Change: We also have been priming the pump if you will.

Speaker Change: With the entire ecosystem getting them ready for liquid cooling.

Speaker Change: We've been talking to the ecosystem of block about Blackwell for quite some time.

Speaker Change: And the <unk> the data centers.

Speaker Change: The ODM the system makers.

Speaker Change: Our supply chain beyond them.

Speaker Change: <unk> <unk>.

Speaker Change: Cooling supply chain base.

Speaker Change: Local calling supply chain based data center supply chain base.

Speaker Change: No one is going to be surprised with Blackwell coming and the capabilities that we would like to deliver with Grace Blackwell 200 GB to 100 is going to be exceptional.

Jensen Huang: No one is going to be surprised with Blackwell coming and the capabilities that we would like to deliver with Grace Blackwell 200. GB200 is going to be exceptional. Our next question will come from the line of Vivek Arya with Bank of America Securities. Please go ahead. Thanks for taking my question. Jensen, how are you ensuring that there is

Unknown Speaker: Our next question will come from the line of Vivek Arya with Bank of America Securities. Please go ahead.

Speaker Change: Our next question will come from the line of Vivek Arya with Bank of America Securities. Please go ahead.

Thanks for taking my question gentlemen, how are you ensuring that that is enough utilization off the order our products and that there isn't a pull ahead are holding.

Speaker Change: Behavior because of tight supply competition or other factors.

Speaker Change: Really what checks have you built in the system to give us confidence that our monetization is keeping pace with you already.

Speaker Change: Very strong shipment growth.

Speaker Change: Well.

Jensen Huang: Well... There's the big picture view that I'll come to, but I'll answer your question directly. The demand for GPUs in all the data centers is incredible. We're racing every single day, and the reason for that is because applications like ChatGPT and GPT-4.0 and now there's going to be multi-modality and Gemini and DitzRAMP and Anthropic, and you know all of the work that's being done at all the CSPs is consuming every GPU that's out there.

Speaker Change: I guess I guess.

Speaker Change: There is the there is the big picture view that'll come to and then but.

Speaker Change: I'll answer your question directly.

Speaker Change: The demand for for our Gpus and all the data centers is incredible.

Speaker Change: We're racing every single day and the reason for that is because.

Speaker Change: Applications like chat GBT and GPT for Owen.

Speaker Change: Now, it's going to be multi modality.

Jim: Jim and I and it's ramping in <unk>.

Jim: All of the work that's been done and all the Csp's are consuming.

Jim: Every GPU that's out there.

Jensen Huang: There's also a long line of generative AI startups, some 15,000, 20,000 startups, in all different fields from multimedia to digital characters. Of course, all kinds of design tool applications, productivity applications, digital biology, the movement, the moving of the AV industry to video so that they can train end-to-end models to expand the operating domain of self-driving cars. The list is just quite extraordinary.

Jim: There's also a long line of generative AI startups.

Jim: 15000.

Jim: 20000 startups.

Jim: That in all different fields from from multi media.

Jim: Two digital characters.

Jim: Of course, all kinds of design tool application productivity applications.

Jim: Digital biology.

Jim: The movement, the moving of the AAV industry to video so that they can train.

Jim: End to end models.

Jim: To expand the operating domain of self driving cars are the list is just quite extraordinary we're racing actually.

Jensen Huang: We're racing, actually. [inaudible] is putting a lot of pressure on us to deliver the systems and stand them up as quickly as possible. And of course, I haven't even mentioned all of the sovereign AIs who would like to train all of their regional natural resources of their country, which is their data, to train their regional models. And there's a lot of pressure to stand up those. So anyway, the demand, I think, is really, really high, and it outstrips our supply. Longer term, that's what, You know, that's the reason why I jumped in to make a few comments. Longer term.

Our customers are.

Jim: Our.

Jim: Putting a lot of pressure on us to deliver.

Jim: The delivered a systems and stand it up as quickly as possible and of course I haven't even mentioned all of the sovereignty is who would like to train all of their regional natural natural resource of their country, which is their data.

To train their original models and there is a lot of pressure to stand those systems up.

Jim: So so anyhow the demand I think as is.

Jim: Really really high and has outstripped our supply.

Jim: Longer term.

Jim: That's.

Speaker Change: The reason why I jumped into to make a few comments.

Speaker Change: Longer term.

Speaker Change: Alright.

Jensen Huang: You know, we're completely redesigning how computers work. This is a platform shift. Of course, it's been compared to other platform shifts in the past, but time will clearly tell that this is much, much more profound than previous platform shifts.

Speaker Change: We're completely redesigning how computers work.

Speaker Change: And it's this is this is a platform shift of course has been.

Speaker Change: And compare to other platform shifts in the past.

Speaker Change: Time will clearly tell that this is much much more profound than previous platform shifts and the reason for that is because the computer is no longer an instruction driven only computer it's an intention understanding computer.

Unknown Speaker: And the reason for that is that the computer is no longer an instruction-driven-only computer. It's an intention-understanding computer. And it understands, it understands, of course, the way we interact with it, but it also understands our meaning, what we intend that we have asked it to do, and it has the ability to reason, by inference iteratively process a plan and come back with a solution. And so every aspect of the computer is changing in such a way that, instead of retrieving pre-recorded files, it is now generating contextually relevant, intelligent answers.

Speaker Change: And it understand it understand of course, the way, we interact with it but it also understands our meaning what we intend that we asked it to do and has the ability to reason.

Edwards: Inference Edwards has lead to process a plan and come.

Edwards: Come back with a solution.

Edwards: And so so every aspect of the computer is changing in such a way that instead of retrieving prerecorded files. It is now generating contextually relevant intelligent.

Edwards: Answers.

Unknown Speaker: And so that's going to change computing stacks all over the world. And you saw a build that, in fact, even the PC computing stack is going to be revolutionized, and this is just the beginning of all the things that you know what people see today are just the beginning of the things that we're working on in our labs and the things that we're doing with all the startups and large companies and developers all over the world. It's going to be, it's going to be quite extraordinary. Our next question will come Please go ahead. Great, thank you. Understanding what you just said about how strong demand is.

Edwards: And so that that's going to change computing stacks.

Edwards: All over the World and you saw a built.

Edwards: In fact, given the PC computing stack is going to get revolutionized.

Edwards: This is just the beginning of all the things that.

People see today are the beginning of the things that we're working in our labs and the things that we're doing with all the startups and large companies.

Edwards: Developers all over the world, that's going to be it's going to be quite quite extraordinary.

Unknown Speaker: Our next question will come from the line of Joe Moore with Morgan Stanley. Please go ahead.

Speaker Change: Our next question will come from the line of Joe Moore with Morgan Stanley. Please go ahead.

Joseph Lawrence Moore: Great. Thank you I understand what you just said about how strong demand is.

Speaker Change: You have a lot of demand for <unk> 200, and for black oil products.

Do you anticipate any kind of pause with hopper.

Speaker Change: And each 100 as you sort of migrate to those price where people wait for those new products with speaker product to have or do you think there is enough demand for each 100 to sustained growth.

Unknown Speaker: We see increasing demand for Hopper through this quarter, and we expect demand to outstrip supply for some time, as we now transition to H200 as we transition to Blackwell. Everybody is anxious to get their infrastructure online, and the reason for that is that they're saving money and making money, and they would like to do that as soon as possible.

Speaker Change: We see increasing demand of hopper through this quarter.

Speaker Change: And we expect to be.

We expect demand to outstrip supply.

Speaker Change: For some time.

Speaker Change: As we now transition to H 200, as we transition to Blackwell.

Speaker Change:

Everybody is anxious to get their infrastructure online and the reason for that is because they're saving money and making money.

Speaker Change: They would like to do that as soon as possible.

Unknown Speaker: Our next question will come from the line of Toshiha Hari with Goldman Sachs. Please go ahead.

Unknown Speaker: Our next question will come from the line of Toshia Hari with Goldman Sachs. Please go ahead. Hi, thank you so much for taking the question. Jensen, I wanted to ask about competition. I think many of your customers are cloud.

Speaker Change: Our next question will come from the line of Toshi Hari with Goldman Sachs. Please go ahead.

Speaker Change: Hi, Thank you so much for taking the question Jensen I wanted to ask about competition.

Speaker Change: I think many of your cloud customers have announced new or updates to their existing internal programs in parallel to what they are working on with you guys.

Speaker Change: What extent did you consider them as competitors medium to long term and in your view do you think they are limited to addressing mostly internal workloads or could they be broader and what the address going forward.

Speaker Change: <unk>.

Speaker Change: Okay.

Speaker Change: Yeah, we are different in several ways.

Unknown Speaker: We're different in several ways. First,

Speaker Change: First.

Jensen Huang: NVIDIA's accelerated computing architecture allows customers to process every aspect of their pipeline from unstructured data processing to prepare for training to structured data processing, data frame processing like SQL to prepare for training, to training to inference. And as I was mentioning in my remarks, inference has really fundamentally changed. It's now generation. It's not trying to just detect a cat, which was plenty hard in itself, but it has to generate every pixel of a cat.

Speaker Change: NVIDIA's to accelerated computing architecture.

Speaker Change: Allows.

Speaker Change: Customers to process.

Speaker Change: Every aspect of their pipeline from <unk>.

Speaker Change: On structured data processing to prepare for training to structured data processing data frame processing like sequel to prepare for training.

Speaker Change: Yeah.

Speaker Change: Two training to inference.

Speaker Change: And as I was mentioning in my remarks that inference is really fundamentally changed its now generation.

Speaker Change: It's not trying to just detect the cat, which is which was plenty hard.

Speaker Change: In itself.

Speaker Change: But it has to generate every pixel of a cap.

Jensen Huang: And so, the generation process is a fundamentally different processing architecture, and it's one of the reasons why TensorRT LLM was so well received. We improved the performance by using the same chips on our architecture by a factor of three. That kind of tells you something about the richness of our architecture and the richness of our software. So one, you could use NVIDIA for everything from computer vision, to image processing, to computer graphics, to all modalities of computing. And, as the world is now,

Speaker Change: And so so the generation process is a fundamentally different processing architecture and it's one of the reasons why <unk> was so so well received.

Speaker Change: We improved our performance.

Speaker Change: And using the same chips on our architecture by a factor of three that kind of tells you something about the richness of our architecture and the richness of our software.

Speaker Change: So once you can use nvidia for for everything from computer vision to image processing for computer graphics too.

Speaker Change: All modalities of computing and as the World is now.

Jensen Huang: They are suffering from computing cost and computing energy inflation because general purpose computing has run its course. Accelerated computing is really the sustainable way of going forward. So accelerated computing is how you're gonna save money in computing, it's how you're gonna save energy in computing. And so the versatility of our platform results in the lowest TCO for their data center. Second,

Speaker Change: Suffering from <unk>.

Speaker Change: Computing cost in computing energy inflation, because general purpose computing has run its course accelerated computing is really the sustainable way of going forward.

Speaker Change: Accelerated computing is how you're going to save money and computing is how you're going to save energy and computing.

Speaker Change: The versatility of our of our platform our results in the lowest tcl for their data centers.

Jensen Huang: We're in every cloud. And so for developers that are looking for a platform to develop on, starting with NVIDIA is always a great choice, and we're on-premises, we're in the cloud, you know, we're in computers of any size and shape; we're practically everywhere. And so that's the second reason. The third reason has to do with the fact that we build AI factories. And this is becoming more and more apparent to people that AI is not just a chip problem. It starts, of course, with very good chips.

Speaker Change: We're in every cloud and.

Speaker Change: And so for developers.

Speaker Change: Or are looking for a platform to develop on starting with the video is always a great choice.

Speaker Change: And we were on prem or in the cloud.

Speaker Change: Yes.

Speaker Change: And computers are of any size and shape.

Speaker Change: We're practically everywhere and so.

Speaker Change: The second reason.

Speaker Change: The third reason has to do with the fact that that.

Speaker Change: We build AI factories.

Jensen Huang: And we build a whole bunch of chips for our AI factories, but it's a systems problem. In fact, even AI is now a systems problem. It's not just one large language model. It's a complex system of a whole bunch of large language models that are working together.

Speaker Change: And this is this is becoming more an apparent to people that.

Speaker Change: But AI is not a chip problem.

Speaker Change: Only it starts of course with very good ships and we build a whole bunch of chips for our factories, but it's a systems problem.

Speaker Change: In fact, even even AI is nellix systems problem. It's not just one large language model is a complicated complex system of whole bunch of large language models.

Speaker Change: Working together.

Jensen Huang: And so the fact that NVIDIA builds this system causes us to optimize all of our chips to work together as a system, to be able to have software that operates as a system, and to be able to optimize across the system. Just to put it in perspective in simple numbers, you know, if you had a $5 billion infrastructure, and you improved the performance by a factor of two, which we routinely do.

Speaker Change: So the fact that Nvidia builds the system.

Speaker Change: Causes us to optimize all of our chips to work together as a system.

Speaker Change: To be able to have software that operates as a system and to be able to optimize across the system.

Speaker Change: And just to put in perspective and simple numbers.

Speaker Change: If you had if you had a $5 billion infrastructure.

Speaker Change: And you improve the performance by a factor of two which we routinely do.

Jensen Huang: When you improve the infrastructure by a factor of two, the value to you is $5 billion. But all the chips in that data center don't pay for it. And so the value of it is really quite extraordinary. And this is the reason why, today, performance matters everything. This is at a time when the highest performance is also the lowest cost because the infrastructure cost of carrying all of these chips costs a lot of money.

Speaker Change: When you improve the infrastructure by factor two the value to use $5 billion all the chips in that data center it doesn't pay for it and so the value of it is really quite extraordinary and that's the reason why today performance matters everything.

Speaker Change: This is at a time when when.

Speaker Change: The highest performance is also the lowest cost because the infrastructure cost of <unk>.

Speaker Change: Carrying all of these chips cost a lot of money and it takes a lot of money to two.

Jensen Huang: And it takes a lot of money to fund the data center, to operate the data center, the people that go along with it, the hardware that goes along with it, the real estate that goes along with it. All of it adds up. And so the highest performance is also the lowest.

Speaker Change: Fund the data center to operate the datacenter to assemble that goes along with it the power that goes along with it the real estate that goes along with it and all of it all of it adds up and so the highest performance is also the lowest tcl.

Unknown Speaker: Our next question will come from the line of Matt Ramsey with TD Cowan. Please go ahead.

Speaker Change: Our next question will come from the line of Matt Ramsay with TD Cowen. Please go ahead.

Unknown Speaker: Thank you very much. Good afternoon, everyone.

Speaker Change: Thank you very much good afternoon, everyone.

Unknown Speaker: Jensen, I've been in the data center industry my whole career, but I've never seen the velocity that you guys are introducing new platforms at the same combination of the performance jumps that you're getting, I mean 5x in training. Some of the stuff you talked about a GTC up to 30x in, and Inference.

Jensen Huang: Jensen I.

Speaker Change: Internet data center industry, my whole career I've never seen.

Speaker Change: <unk> that you guys are introducing new platforms.

Speaker Change: At the same combination of the performance jumps that youre getting buybacks.

Speaker Change: And training.

Speaker Change: The stuff you talked about at GTC up to 30 acts in.

Jensen Huang: And it's an amazing thing to watch, but it also creates an interesting juxtaposition between the current generation of products that your customers are spending, and so on. So, I'd like you to, if you wouldn't mind, speak a little bit about how you're seeing that situation evolve with customers. As you move to Blackwell, they're going to have very large installed bases, obviously software compatible, but large installed bases not nearly as performant as your new generation stuff. And it'd be interesting to hear what you see happening with customers along that path. Thank you.

Speaker Change: And in France.

And that it is an amazing thing to watch, but it also creates an interesting juxtaposition ware.

The current generation of product that your customers are spending.

Speaker Change: Billions of dollars on.

Speaker Change: It's going to be not as competitive with your new stuff very very much more quickly than the depreciation cycles of that product. So I'd like you did.

Wouldn't mind speak a little bit about how youre seeing that situation evolve itself with customers as you move to Blackwell.

Speaker Change: Going to have very large installed bases.

Speaker Change: Obviously software compatible.

Speaker Change: But large installed bases of product.

Not nearly as performance as your new generation stuff and it'd be interesting to hear what you see happening with customers along that path. Thank you.

Jensen Huang: Yeah, I really appreciate it. There are three points that I'd like to make. If you're five percent into the build out, versus if you're 95% into the build-out, you're gonna feel very different. And because you're only 5% into the build-out, anyway, you build as fast as you can. And, you know, when Blackwell comes, it's going to be terrific. And then after Blackwell, as you mentioned, we have other Blackwells coming.

Speaker Change: Yes, it really appreciated three three points that I'd like to make.

Speaker Change: If you are 5%.

Speaker Change: Into the build out.

Speaker Change: Versus over 95% into the build out.

Speaker Change: Youre going to feel very differently.

Speaker Change: And because you are only 5% into the build out anyhow.

Speaker Change: You built your build as fast as you can.

Speaker Change: And.

Speaker Change: When Blackwell comps, that's going to be terrific and then after Blackwell as you mentioned, we have we have.

Speaker Change: Other black loss coming and then there's the short.

Jensen Huang: And then there's a short, you know; we're on a one-year rhythm, as we've explained to the world. And we want our customers to see our roadmap for as far as they like, but they're early in their build-out anyway. And so they have to just keep on building, and so there's going to be a whole bunch of chips coming at them, and they just have to keep on building, and you know if you performance average your way into it. So that's the smart thing to do. They need to make money today, they want to save money today, and time is really, really valuable to them. Let me give you an example of time being really valuable.

Speaker Change: We're on a one year rhythm as we've explained to the world.

Speaker Change: And we want our customers to see our roadmap for as far as they like.

Speaker Change: They're they're.

Speaker Change: They're early in their build out anyways and so they are just keep on building.

Speaker Change: Okay, and so there's going to be a whole bunch of chips coming at them and they just got to keep on building and just you know if you will our performance average your way into it.

So that's that's the smart things smart thing to do they need to make month today, they want to save money today.

Speaker Change: And.

Speaker Change: And time is really really valuable to them.

Speaker Change: Let me give you. An example of time being really valuable why is this idea of standing up a datacenter instantaneously is so valuable and getting this thing called time to training is so valuable the reason for that is because the.

Jensen Huang: Why this idea of standing up a data center instantaneously is so valuable, and getting this thing called time to train is so valuable. The reason for that is because the next company who reaches the next major plateau gets to announce a groundbreaking AI, and the second one after that gets to announce something that's, you know... 0.3% better. And so the question is, do you want to be the company delivering groundbreaking AI again and again?

Speaker Change: The next the next company who reaches the next major plateau.

Speaker Change: It gets to announce a groundbreaking AI.

Speaker Change: And then the second one after that gets to announce something that's you know.

Speaker Change: 0.3% better.

Speaker Change: And so the question is do you Wanna be repeatedly the company delivering groundbreaking AI.

Speaker Change: For the company.

Speaker Change: Delivering 3% better.

Jensen Huang: or the company, you know, delivering 0.3% better. And that's the reason why this race, as in all technology races, the race is so important, and and and you're seeing this race across multiple companies because it is so vital to have technology leadership for companies to trust the leadership and want to build on your platform and know that the platform that they're building on is going to get better and better. And so leadership matters a great deal. Time to train matters a great deal. The difference between them.

Speaker Change: And that's the reason why this race and all technology races.

Speaker Change: The race is so important.

Speaker Change: And you're seeing this race across multiple companies because this is so vital to have technology leadership for companies too.

Speaker Change: Trust, the leadership and want to build on your platform and know that the platform that they're building on it is going to get better and better.

Speaker Change: And so leadership matters, a great deal of time to train matters, a great deal the difference between.

Jensen Huang: Time to train that is, you know, three months earlier, just to get it done, in order to get time to train on a three months' project. You know, getting started three months earlier is everything. And so it's the reason why we're setting up hopper systems like mad right now, because the next plateau is just around the corner. And so that's the second reason that the first comment that you made is really a great comment, which is, you know, how is it that we're doing? We're moving so fast and advancing so quickly because we have all the stacks here.

Speaker Change: Hi, I'm disclaimed that is.

Speaker Change: Three months earlier just to get it done.

Speaker Change: In order to get time to train on three months project.

Speaker Change: Getting started three months earlier is everything and so what's the reason why we're standing up Hopper systems like Mad right now because the next plateau is just around the corner.

Speaker Change: And so so that's the second reason that the first the first comment that you made is really a great comment which is how is it that we're doing we're moving so fast.

Speaker Change: Dancing, so quickly because we have all the stats here.

Jensen Huang: We literally built the entire data center, and we can monitor everything, measure everything, optimize across everything. We know where all the bottlenecks are. We're not guessing about it. We're not putting up PowerPoint slides that look good.

Speaker Change: We literally build the entire data center and we can monitor everything measure everything optimized across everything we know where all the bottlenecks are we're not guessing about it.

Speaker Change: We're not putting a powerpoint slides that look good we're actually you know.

Jensen Huang: We're actually, you know, we also like our PowerPoint slides to look good, but we're delivering systems that perform at scale. And the reason why we know they perform at scale is because we built them all. Now, one of the things that we do that's a bit of a miracle is that we build entire AI infrastructures here, but then we disaggregate it and integrate it into our customers' data centers however they like.

Speaker Change: We also like our Powerpoint slides look good, but but we're delivering systems that perform at scale.

Speaker Change: And the reason why we noted the performance scale is because we built it all here.

Speaker Change: Now one of the things that we do that as a bit of a miracle is that we build entire AI infrastructure here.

Speaker Change: But then we disaggregated and integrated into our customers' data centers. However, daylight.

Jensen Huang: But we know how it's going to perform, and we know where the bottlenecks are. We know where we need to optimize with them, and we know where we have to help them improve their infrastructure to achieve the highest performance. This deep, intimate knowledge at the entire data center scale is fundamentally what sets us apart today; we build every single chip from the ground up. We know exactly how processing is done across the entire system, and so we understand exactly how it's going to perform and how to get the most out of it with every single generation. I appreciate it. Those are the three points.

Speaker Change: We know how it's going to perform and we know where the bottlenecks are we know where we need to optimize with them and we know where we have to help them improve their infrastructure to achieve the most performance. This deep intimate knowledge at the entire data center scale is fundamentally what sets us apart today.

Speaker Change: <unk>.

Speaker Change: Build every single chip from the ground up.

No exactly how processing is done across the entire system.

Speaker Change: And so we understand exactly how it is going to perform and how to get the most out of it with every single generation. So I. Appreciate those are the three points.

Unknown Speaker: Your next question will come from the line of Marc Lopakis with Evercore ISI. Please go ahead.

Speaker Change: Your next question will come from the line of Mark <unk> with Evercore ISI. Please go ahead.

Unknown Speaker: Hi, thanks for taking my question. Jensen, in the past, you've made the observation that general purpose computing ecosystems typically dominated each computing era. And I believe the argument was that they could adapt to different workloads, get higher utilization, and drive the cost of the compute cycle down. And this was a motivation for why you were driving to a general purpose GPU CUDA ecosystem for accelerated computing. And if I mischaracterized that observation, please do let me know. So the question is:

Hi, Thanks for taking my question.

Speaker Change: <unk> in the past you've made the observation that general purpose computing ecosystem is typically dominated each competing era and I believe the argument was that they could adapt to different workloads get higher utilization driving cost compute cycle down.

Speaker Change: As a motivation for why you were driving to a general purpose GPU cuda ecosystem for accelerated computing.

Speaker Change: And if I mischaracterized that.

Speaker Change: <unk>. Please do let me know so the question is <unk>.

Unknown Speaker: Given that the workloads that are driving demand for your solutions are being driven by neural network training and inferencing, which on the surface seems like a limited number of workloads, then it might also seem to lend themselves to custom solutions. And so the question is, does the general purpose computing framework become more at risk? Or is there enough variability or a rapid enough evolution of these workloads to support that historical general purpose framework?

Speaker Change: Given that the workloads that are driving demand for your solutions are being driven by neural network training and inferencing, which on the surface seem like a limited.

Speaker Change: Number of workloads.

Speaker Change: And then Ben might also seem to lend themselves to custom solutions and so then the question is does that does the general purpose computing framework become more at risk or is there enough variability or.

Speaker Change: Rapid enough evolution on these workloads that that support that historical and general purpose framework. Thank you.

Unknown Speaker: Thank you.

Jensen Huang: Yeah, NVIDIA's accelerated computing is versatile, but I wouldn't call it general purpose. Like, for example, we wouldn't be very good at running a spreadsheet. You know, that was really designed for general purpose computing. And so the control loop of an operating system code probably isn't fantastic for general purpose computing, not for accelerated computing. And so I would say that we're versatile. And that's usually the way I describe it.

Yes, and these accelerated computing is versatile, but I wouldn't call. It a general purpose like for example, we wouldn't be very good at running the spreadsheet.

Speaker Change: That was really designed for general purpose computing.

And so there's a there's a.

The control loop of an operating system code, probably isn't isn't fantastic for general purpose computing not for not for our for accelerated computing.

Jensen Huang: There's a rich domain of applications that we've been able to accelerate over the years. But they all have a lot of commonalities. Maybe maybe some deep differences, but commonalities, you know, they're all things that I can run in parallel, they're all highly heavily threaded. 5% of the code represents 99% of the runtime, for example.

So I would say that we're versatile.

Speaker Change: And that's usually the way I'd describe it theres a theres a rich domain of applications that we're able to accelerate over the years, but they all have a lot of commonalities.

Speaker Change: Maybe maybe some deep differences, but commonalities. They are all things that I can run in parallel they are all heavily shred it.

Speaker Change: 5% of the code represents 99% of the run time for example, those are our properties of accelerated computing.

Jensen Huang: You know, those are all properties of accelerated computing. The versatility of our platform and the fact that we design entire systems is the reason why, over the course of the last 10 years or so, the number of startups that you guys have asked me about in these conference calls have been fairly large. And every single one of them, because of the brittleness of their architectures, the moment the generative AI came along, or the moment the fusion models came along, the moment the next models, you know, the next models are coming along now. And now, all of a sudden, look at this.

Yeah.

Speaker Change: The versatility of our platform and the fact that we design entire systems is the reason why.

Speaker Change: All right.

Speaker Change: Over the course of.

Speaker Change: The last.

Speaker Change: 10 years or so the number of startups that you guys have asked me about in these conference calls.

Speaker Change: It's fairly large.

Speaker Change: And.

Speaker Change: Every single one of them because of the brittleness of their architecture. The moment the moment generative AI came along.

Speaker Change: The moment diffusion models came along the moment. The next models. The next models are coming along now.

Speaker Change: And now all of a sudden look at this.

Jensen Huang: Large language models with memory. Because the large language models need that memory so they can carry on a conversation with you, understand the context, all of a sudden, the versatility of the grace memory became super important. And so each one of these advances in generative AI and the advancement of AI really begs for not having a widget that's designed for one model but to have something that is really good for this entire domain, has the properties of this entire domain, but obeys the first principles of software.

Speaker Change: Large languished models with memory.

Speaker Change: Because the larger language modeled niestat memory. So they can carry on a conversation with you understand the context.

Speaker Change: All of a sudden the versatility of the Grace memory became Super important.

Speaker Change: And so each one of these advances.

Speaker Change: And.

Speaker Change: Generative AI and the advancement of AI.

Speaker Change: Really begs for not having a.

Speaker Change: A widget that's designed for one model but.

To have something that is really good for this entire domain properties of this entire domain, but obeys the the first principles of software.

Jensen Huang: That software is going to continue to evolve. That software is going to keep getting better and bigger. We believe in the scaling of these models. There are a lot of reasons why we're going to scale by easily a million times in the coming few years for good reasons. And we're looking forward to it, and we're ready for it. And so the versatility of our platform is really quite key. And if you're too brittle and too specific, you might as well just build an FPGA, or you build an ASIC, or something like that. But that's hardly a computer.

Speaker Change: That software is going to continue to evolve that software is going to keep getting better and bigger we.

Speaker Change: We believe in the scaling.

Speaker Change: Of these models.

Speaker Change: A lot of reasons, why we're going to scale.

Speaker Change: By easily 1 million times.

Speaker Change: In the coming few years for good reasons, and we're looking forward to it and we're ready for it.

Speaker Change: And so the versatility of our platform is really quite key and it's not if you were to if you're too brittle and two specific you might as well just build an FPGA or you build an ASIC or something like that but that's hardly a computer.

Unknown Speaker: Our next question will come from the line of Blaine Curtis with Jefferies. Please go ahead.

Unknown Speaker: Our next question will come from the line of Blaine Curtis with Jefferies. Please go ahead. Hey, thanks for taking my question. I'm actually kind of curious. I mean, being

Speaker Change: Our next question will come from the line of Blayne Curtis with Jefferies. Please go ahead.

Blayne Curtis: Hey, Thanks for taking my question I actually kind of curious I mean are you.

Blayne Curtis: Being supply constrained how do you think about I mean, you came out with a product for China, It's 'twenty I'm, assuming there'll be upon a demand for it but obviously, we're trying to serve your customers with.

Blayne Curtis: The other hopper products. It was kind of curious how you're thinking about that in the second half you could elaborate any impact what are you thinking for sales as well as gross margin.

Unknown Speaker: I didn't hear the questions; something bleeped out.

Speaker Change: I didn't hear the questions on <unk>.

Unknown Speaker: H20 and how you're thinking about allocating supply between the different hopper products.

Speaker Change: H 'twenty and how youre thinking about allocating supply between the different products.

Jensen Huang: Well, you know, we have customers that we honor, and we do our best for every customer. But it is the case that our business in China is substantially lower than the levels of the past, and it's a lot more competitive in China now because of the limitations on our technology, and so those matters are two. However, you know, we continue to do our best to serve the customers and the markets there, and to the best of our ability, we'll do our best, you know?

Speaker Change: Well.

Speaker Change: We have customers that we honor.

Speaker Change: <unk>.

Speaker Change: And we do our best for every customer.

Speaker Change: It is the case that.

Speaker Change: But.

Speaker Change: Our business in China.

Speaker Change: It is substantially lower.

Speaker Change: And then the levels of the past.

Speaker Change: And and.

Speaker Change: It's a lot more competitive in China now.

Speaker Change: Because of the limitations on our technology.

Speaker Change: And and.

So those those those matters are true however.

Speaker Change: We continue to to do our best to serve the customers in the markets, there and and to divest our ability, we'll we'll do our best.

Jensen Huang: And so, but I think overall, the comments that we made about Demand outstripping supply are for the entire market and particularly so for H-200 and Blackwell towards the end of the year. Our next question will come from the line of Sreeni Pazhuri with Raymond James. Please go ahead. Thank you. Jensen, actually, more of a clarification on what you said. GP200

Speaker Change: And so but I think overall the comments that we made.

Speaker Change: About.

Speaker Change: Demand outstripping supply.

Speaker Change: As is for the entire.

Speaker Change: The entire market.

Speaker Change: And.

Speaker Change: Particularly so for <unk> hundred and Blackwell.

Speaker Change: Towards the end of the year.

Unknown Speaker: Our next question will come from the line of Srini Pazhuri with Raymond James. Please go ahead.

Speaker Change: Our next question will come from the line of <unk> with Raymond James. Please go ahead.

Speaker Change: Thank you Jensen actually more of a clarification on what you said.

GBP 200 systems. It looks like there is significant demand for systems. Historically, I think you've sold a lot of eight gx Butch and from Gpus and the systems business was relatively small so I'm. Just curious why is it that now you are seeing such a strong demand for systems going forward is it just a <unk>.

Speaker Change: Or is there something else or is it just the architecture. Thank you.

Jensen Huang: Yeah, I appreciate that. In fact, the way we sell GB200 is the same. We disaggregate all of the components that make sense, and we integrate them into Computer Makers. We have a hundred different computer SIS configurations that are coming this year for Blackwell, and that is off the chart. Hopper, Hopper, Frankly, had only half, but that's at its peak, you know it started out with way less than that even, and so you're going to see liquid-cooled versions, air-cooled versions, x86 versions, GRACE versions, and so on and so forth. There's a whole bunch of systems that are being designed, and they're offered from all of our ecosystem of great partners. Nothing has really changed. The Blackwell platform has expanded our offering. It is tremendous.

Speaker Change: I appreciate that.

Speaker Change: <unk>.

Speaker Change: The way we sell GBP 200 is the same.

Speaker Change: We disaggregate.

Speaker Change: All of the components that makes sense.

And we integrated into computer makers.

We have 100 different computer system configurations that are Columbia coming this year for Blackwell.

Speaker Change: And that is that is off the charts.

Speaker Change: Hopper Hopper.

Speaker Change: Frankly has.

Speaker Change: Only half, but that's at its peak.

Speaker Change: It started out with way less than that even.

Speaker Change: And so youre going to see liquid cooled version air cooled version X 86 versions Grace versions.

Speaker Change: So and so forth and there's a whole bunch of systems that are being designed and they're offered from all of our ecosystem of great partners. Nothing nothing has really changed now of course.

Speaker Change: The Blackberry platform is has expanded our offering.

Speaker Change: Tremendously.

Jensen Huang: The integration of CPUs and the much more compressed density of computing, liquid cooling is going to save data centers a lot of money in provisioning power, and not to mention being more energy efficient. And so it's a much better solution; it's more expansive, meaning that we offer a lot more components of a data center. And everybody wins. You know, the data center gets much higher performance networking, from networking switches, networking. Of course, next, we have Ethernet now, so that we can bring NVIDIA AI to, large-scale NVIDIA AI to customers who only operate, only know how to operate Ethernet because of the ecosystem that they have, and, And so, so Blackwell is much more expansive. We have a lot more to offer our customers this time, this generation.

Speaker Change: Hi.

Speaker Change: The.

Speaker Change: Integration of Cpus and.

Speaker Change: The much more compressed density of computing.

Speaker Change: Cooling is going to save data centers, a lot of money and provisioning power.

Speaker Change: And not to mentioned to be more energy efficient.

Speaker Change: And so so.

Speaker Change: It's a much better solution.

Speaker Change: More expensive, meaning that we offer a lot more components of our.

Speaker Change: Data center and everybody wins, the datacenter gets much higher performance networking from networking switches networking of.

Speaker Change: Of course next we have Ethernet now.

Speaker Change: So that we can bring Nvidia AI too.

Speaker Change: And large scale in video AI to customers who.

Speaker Change: Only operate on not only know how to operate Ethernet because of the.

Speaker Change: Yeah.

Speaker Change: The ecosystem that they have.

Speaker Change: <unk>.

Speaker Change: And so so blackwell is much more expensive we have a lot more to offer our customers system.

Speaker Change: This generation around.

Speaker Change: Okay.

Unknown Speaker: Our next question will come from the line of William Stein with Truist Securities. Please go ahead.

Speaker Change: Our next question will come from the line of William Stein with Truth Securities. Please go ahead.

Jensen Huang: Great, thanks for taking my question. Jensen, at some point, NVIDIA decided that while there were reasonably good CPUs available for data center operations, your ARM-based grace CPU provided some real advantage that made that technology worth delivering to customers, perhaps related to cost or power consumption or technical synergies between Grace and Hopper, Grace and Blackwell. Can you address whether there could be a similar dynamic that might emerge on the client side whereby, while there are very good solutions, you've highlighted that Intel and AMD are very good partners and deliver great products in x86. But there might be some, especially in emerging AI workloads, some advantages that NVIDIA can deliver that others have more of a challenge.

William Wong: Great. Thanks for taking my question Jensen at some point Nvidia decided that.

Speaker Change: When there were.

Speaker Change: While they were reasonably good Cpus available for data Center operations. Your arm based CPU provided some real advantage that made that technology, we're delivering to customers.

Speaker Change: Press related to cost of power consumption or.

Speaker Change: Technical synergies between Greece, and Hopper Blackwell can you address whether there could be a similar dynamic that might emerge on the client side, whereby while they are very good solutions you've highlighted that.

Speaker Change: Tony <unk>.

Speaker Change: Partners and deliver great products and exceed six but there might be some especially in emerging AI workloads. Some advantage that nvidia can deliver that others have more of a challenge.

Speaker Change: Well.

Jensen Huang: So, you mentioned some really good reasons. It is true that for many applications, our partnership with x86, our x86 partners are really terrific, and we build excellent systems together. But GRACE allows us to do something that isn't possible with the current configuration, the system configuration today. The memory system between Grace and Hopper is coherent and connected. The interconnect between the two chips, You know, calling it two chips is almost weird because it's like a superchip. The two of them are connected by this interface.

Speaker Change: You mentioned you mentioned some really good reasons.

Speaker Change: It is true that for many of the applications.

Speaker Change: Our partnership with X 86, or <unk> 80.

Speaker Change: Six partners or are a really terrific and we build excellent systems together.

Speaker Change: But grace allows us to do something that is impossible with the configuration of the system configuration today.

Speaker Change: The memory system between Grace and Hopper, our coherent and connected.

Speaker Change: The interconnect between the two chips.

Speaker Change: Calling it two chips is almost weird because it's like a super chip. The two of them are connected with this with this interface that's like a terabytes per second.

Jensen Huang: That's like a terabyte per second. Yeah, it's off the charts, and the memory that's used by Grace is LPDDR. It's the first data center grade low-power memory.

Speaker Change: After charts.

Speaker Change: And the memory that's used by Grace is LP DDR.

Speaker Change: It's the first data center grade low power memory.

Jensen Huang: And then, finally, because of the architecture, because we can create our own architecture with the entire system now, we could create something that has a really large NVLink domain, which is vitally important to the next generation of large language models for inferencing. And so you saw that GB200 has a 72 node NVLink domain. That's like 72 Blackwells connected together into one giant GPU.

Speaker Change: And so we save a lot of power on every single node and then finally, because because of the architecture because we can create our own architecture with the entire system now we can.

Speaker Change: Could create something that has.

Speaker Change: Really large NV link domain, which is vitally important to the next generation large language models for inferencing.

And so.

Speaker Change: You saw that GBP 200 has a.

Speaker Change: 72 node M D link domain.

Speaker Change: 72, Black wells connected together into one giant GPU and so we needed we needed a great platform to be able to do that and so there is there are architectural reasons their software programming reasons.

Jensen Huang: And so we needed, we needed Grace Blackwells to be able to do that. And so there are architectural reasons, there are software programming reasons, and then there are system reasons that are essential for us to build them that way. So if we see opportunities like that, we'll explore them. And today, as you saw at the build yesterday, which I thought was really excellent, as Satya announced, the next-generation PC is CoPilot Plus PC, which runs fantastically on NVIDIA's RTX GPUs that are shipping in laptops, but it also supports ARM beautifully, and so it opens up opportunities for system innovation, even for PCs.

Speaker Change: And then there are system reasons.

Speaker Change: Are essential for us to build them that way.

Speaker Change: And so if we see opportunities like that.

Speaker Change: We will explore it.

Speaker Change: Today as you saw as you saw it.

Speaker Change: The Bill yesterday, which I thought was really excellent.

Speaker Change:

Speaker Change: As such you announced.

Speaker Change: The next generation, Dcs, copilot, plus PC, which which.

Speaker Change: Which runs fantastically on NVIDIA's RPX Gpus that are that are shipping in laptops.

Speaker Change: But it also supports arm beautifully.

Speaker Change: So it opens up it opens up opportunities for system innovation.

Speaker Change: Even for even for Pcs.

Jensen Huang: Our last question comes from the line of CJ Mews with Kandra Fitzgerald. Please go ahead. Yeah, good afternoon. Thank you for taking the question. I guess, Jensen, a bit of a longer term view. I know Blackwell hasn't even launched yet, but obviously investors are forward-looking.

Unknown Speaker: Our last question comes from the line by CJ Mews with Kandra Fitzgerald. Please go ahead. Yeah, get out.

Speaker Change: Our last question comes from the line of C. J Muse with Cantor Fitzgerald. Please go ahead.

Speaker Change: Good afternoon, and thank you for taking the question I guess Jensen a bit of a longer term question I know Blackwell hasn't even launched yet, but obviously investors are forward looking and amidst rising potential competition from Gpus and estimate.

Speaker Change: How are you thinking about in video spaces innovation Ed.

Speaker Change: Bold scaling over the last decade.

Speaker Change: Truly impressive kudos parts of the precision great career and connectivity, we look forward, what frictions need to be solved in the coming decade, and I guess, maybe more importantly, what are you what are you willing to share with us today.

Unknown Speaker: Well, I can announce that after Blackwell, there's another chip, and we are on a one-year rhythm, and so I, and, you can also count on us having new networking technology on a very fast rhythm. We're announcing Spectrum X for Ethernet, but we're all in on Ethernet, and we have a really exciting roadmap coming for Ethernet. We have a rich, rich, rich ecosystem of partners.

Speaker Change: Well I can announce that.

Speaker Change: After Blackwell Theres another chip.

Speaker Change: And and we are on a one year with them.

Speaker Change: And so.

Speaker Change: Hi.

Speaker Change: And and.

Speaker Change: You can also count that count on us, having new networking technology on a very fast rhythm.

Speaker Change: We're announcing spectrum extra Ethernet.

Speaker Change: But we're all in on Ethernet.

Speaker Change: And we have a really exciting roadmap coming for Ethernet.

We have a rich rich rich ecosystem of partners.

Jensen Huang: Dell announced that they're taking Spectrum X to market. We have a rich ecosystem of customers and partners who are going to bring our entire AI factory architecture to market, and so for companies that want the ultimate performance. We have the InfiniBand Computing Fabric. InfiniBand is a computing fabric. Ethernet's a network, and InfiniBand, over the years, started out as a competing fabric, became a better and better network. Ethernet is a network, and with Spectrum X, we're going to make it a much better computing fabric, and we're committed, fully committed to all three. Lynx, NVLink Computing Fabric, for a single computing domain, to InfiniBand computing fabric, to Ethernet networking computing fabric. And so we're gonna take all three of them forward at a very fast clip.

Speaker Change: Dell announced that they were taking spectrum extra market.

Speaker Change: We have a rich ecosystem of customers and partners, who are going to announce taking our entire AI factory architecture to market.

Speaker Change: And so.

Speaker Change: For companies that want the ultimate performance.

Speaker Change: We have infiniband computing fabric infiniband as a computing fabric Ethernet as a network.

Speaker Change: Infiniband over the years started out as a computing fabric became a better and better network.

Internet is a network and which spectrum X, we're going to make it a much better computing fabric.

Speaker Change: And we're committed fully committed to all three.

Speaker Change: Links NV linked commute.

Speaker Change: Our computing fabric.

Speaker Change: Sure.

Speaker Change: Single single computing domain to Infiniband computing fabric to Ethernet networking computing fabric and so so we're going to take all three of them.

Speaker Change: Forward at a very fast clip.

Jensen Huang: And so you're going to see new switches coming, new NICs coming, new capability, new software stacks that run on all three of them, new CPUs, new GPUs, new networking NICs, new switches, a mountain of chips that are coming. And all of it, the beautiful thing is, all of it runs CUDA. And all of it runs our entire software stack. So if you invest in our software stack today without doing anything at all, it's just gonna get faster and faster and faster and faster.

Speaker Change: And so you're going to see new switches coming new next coming.

Speaker Change: New capability, new software stacks that run on all three of them, New Cpus, New Gpus, new networking next new switches.

Speaker Change: A mountain of chips that are that are coming.

Speaker Change: And all of it the beautiful thing is all of it runs cuda.

Speaker Change: And all of it runs our entire software stack. So if you invest today on our software stack.

Speaker Change: Without doing anything at all it's just going to get faster and faster and faster and faster and if you invest in art are our architecture today without doing anything.

Jensen Huang: And if you invest in our architecture today without doing anything, it will go to more and more clouds and more and more data centers, and everything just works. And so I think the pace of innovation that we're bringing will drive up the capability on the one hand and drive down the TCO on the other. And so we should be able to scale out with the NVIDIA architecture for this new era of computing and start this new industrial revolution where we manufacture not just software anymore, but we manufacture artificial intelligence tokens, and we're going to do that at scale.

Speaker Change: Go to more and more clouds, and more and more data centers and everything just months.

Speaker Change: And so so I think the.

Speaker Change: The pace of innovation that we're bringing.

Speaker Change: Yeah.

Speaker Change: It will drive up the capability on the one hand and drive down the Tcl on the other hand, and so we should be able to scale out with the Nvidia architecture for this new era of computing and <unk>.

Speaker Change: Start this new industrial Revolution, where we manufacture and not just software anymore, but we manufacturer artificial intelligence tokens and we're going to do that at scale.

Speaker Change: Thank you.

Unknown Executive: That will conclude our question and answer session and our call for today. We thank you all for joining us, and you may now disconnect.

Speaker Change: That will conclude our question and answer session and our call for today. We thank you all for joining and you may now disconnect.

Speaker Change: Yeah.

Speaker Change: Okay.

Speaker Change: Okay.

Speaker Change: Okay.

Speaker Change: Okay.