Q3 2024 NVIDIA Corp Earnings Call

Speaker 1: Good afternoon. My name is Jael and I will be your conference operator.

Good afternoon. My name is Gail and I will be your conference operator today at this time I would like to welcome everyone to <unk> third 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.

Speaker 1: At this time, I would like to welcome everyone to NVIDIA's third quarter earnings call. All lines have been placed on mute to prevent any background noise. After the speakers are...

Speaker 1: 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, again, press the star one. Thank you. Simone Judith J fried

If you'd 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 again press the star one thank you.

Simona Jankowski you May now begin your conference.

Thank you good afternoon, everyone and welcome to video conference call for the third quarter of fiscal 2024 with me today from Nvidia are Jensen, Huang President and Chief Executive Officer, and Colette, Kress Executive Vice President and Chief Financial Officer.

Speaker 2: Thank you. Good afternoon, everyone, and welcome to NVIDIA's conference call for the third quarter of fiscal 2024. With me today from NVIDIA are Jensen Huang, President and Chief Executive Officer, and Colette Kress, Executive Vice President and Chief Financial Officer.

Speaker 2: I'd like to remind you that our call is being webcast live on NVIDIA's investor relations website. The webcast will be available through B-Play until the conference call to discuss our financial results for the fourth quarter and fiscal 2024. The content of today's call is NVIDIA's property. It can be reproduced or transcribed without our prior written consent.

I'd like to remind you that our call is being webcast live on <unk> Investor Relations website.

Webcast will be available for replay until the conference call to discuss our financial results for the fourth quarter and fiscal 2024. The contents of today's call is <unk> property, it can't be reproduced or transcribed without our prior written consent.

Speaker 2: 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 in 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.

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

Speaker 2: All statements are made as of today, November 21, 2023, based on information currently available to us. Except it's required by law, we assume no obligation to update any singleokrama on our website.

All statements are made as of today November 21, 2023 based on information currently available to us.

Sept as required by law, we assume no obligation to update any.

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 with that let me turn the call over to Colette.

Speaker 2: 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. With that, let me turn the call over to Collette.

Speaker 3: Thanks, Simona. Q3 was another record quarter. Revenue of $18.1 billion was up 34% sequentially and up more than 200% year on year and well above our outlook of $16 billion.

Thanks Simona Q.

Q3 was another record quarter revenue of $18 1 billion was up 34% sequentially and up more than 200% year on year.

Well above our outlook of $16 billion.

Speaker 3: Starting with Data Center, the continued ramp of the NVIDIA HDX platform based on our Hopper Tensor Core GPU architecture, along with InfiniBand and networking drove record revenue of $14.5 billion, up 41% sequentially and up 279% year-on-year.

Starting with data center.

<unk> realm of the Nvidia HD X platform based on our Hopper tensor core GPU architecture, along with Infiniband and networking drove record revenue of $14 5 billion up 41% sequentially and up 279% year on year.

Speaker 3: NVIDIA HDX with InfiniBand together are essentially the reference architecture for AI supercomputers and data center infrastructures. Some of the most exciting generative AI applications are built and run on NVIDIA, including Adobe Firefly, ChatGPT, Microsoft 365 CoPilot, CoAssist,

Video HTS with Infiniband together are essentially the reference architecture for AI supercomputers and data center infrastructures.

Some of the most exciting generative AI applications are built and run on Nvidia, including Adobe Firefly Chuck <unk>.

Microsoft 365 co pilot co exist now.

Speaker 3: Now assist with ServiceNow and Zoom AI companion.

Now assist with service now zoom aortic companion.

Speaker 3: Our data center compute revenue quadrupled from last year, and networking revenue nearly tripled.

Our data center compute revenue quadrupled from last year and networking revenue nearly tripled.

Speaker 3: Investment in infrastructure for training and inferencing large language models, deep learning recommender systems, and generative AI applications is fueling strong broad-based demand for NVIDIA accelerated computing.

Investment in infrastructure for training and Inferencing large language models.

Learning recommendation systems and generative AI applications.

<unk> strong.

Strong broad based demand for Nvidia accelerated computing.

Speaker 3: Inprinting is now a major workload for NVIDIA AI computing.

In printing.

Now a major workload for Nvidia AI computing.

Consumer Internet companies and enterprises drove exceptional sequential growth in Q3.

Speaker 3: Consumer internet companies and enterprises drove exceptional sequential growth in Q3, comprising approximately half of our data center revenue and outpacing total growth.

Comprising approximately half of our data center revenue and outpacing total growth.

Speaker 3: Companies like Meta are in full production with learning recommender systems and also investing in innovative AI to help advertisers optimize images and text.

Companies like meta are in full production with learning recommendation systems and also investing in centers of AI to help advertiser optimize images.

Yes.

Speaker 3: Most major consumer internet companies are racing to ramp up generative AI deployment. The enterprise wave of AI adoption is now beginning. Enterprise software companies such as Adobe, Databricks, Snowflake, and ServiceNow are adding AI co-pilots and assistants to their platforms.

Most major consumer Internet companies are racing to ramp up generally they are deploying.

Enterprise wave of AI adoption is now beginning.

Enterprise software companies, such as Adobe data Brooks Snowflake and service now are adding AI copilot and assistance to those platforms.

Speaker 3: And broader enterprises are developing custom AI for vertical industry applications such as Tesla and autonomous driving.

And broader enterprises are developing custom AI for vertical industry applications, such as Tesla and autonomous driving.

Cloud service providers drove roughly the other half of our data center revenue in the quarter.

Speaker 3: cloud service providers drove roughly the other half of our data center revenue in the quarter.

Speaker 3: Demand was strong from all hyperscale CSPs, as well as from a broad new set of GPU specialized CSPs globally that are rapidly grown to address the new market opportunities in AI.

Demand was strong from all Hyperscale TSP as well as from a broadening set of GPU specialized.

Globally that are rapidly grown to address the new market opportunity and a yard.

Nvidia H 100 tensor core GPU instances are now generally available in virtually every cloud with instances.

Speaker 3: NVIDIA H100 Tensor Core GPU instances are now generally available in virtually every cloud with instances in high demand.

Mt.

We have significantly increased supply every quarter this year to meet strong demand and expect to continue to do so next year.

Speaker 3: We have significantly increased supply every quarter this year to meet strong demand and expect to continue to do so next year. We will also have a broader and faster product launch cadence to meet a growing and diverse set of AI opportunities.

We'll also have a broader and faster product launch cadence to meet a growing and diverse set of AI opportunities.

Towards the end of the quarter. The U S government announced a new set of export control regulations for China, and other markets, including Vietnam and certain countries in the middle East.

Speaker 3: Toward the end of the quarter, the U.S. government announced a new set of export control regulations for China and other markets, including Vietnam and certain countries in the Middle East.

Speaker 3: These regulations require licenses for the export of a number of our products, including our Hopper and Ampere 100 and 800 series, and several others.

These regulations require licenses for the export of a number of our products, including our Hopper and am pure 108 hundred series and several others.

Speaker 3: Our sales to China and other affected destinations derived from products that are now subject to licensing requirements have consistently contributed approximately 20 to 25 percent of data center revenue over the past few quarters.

Our sales to China and other affected destination derived from products that are now subject to licensing requirements has consistently contributed approximately 20% to 25% of data center revenue over the past few quarters.

Speaker 3: We expect that our sales to these destinations will decline significantly in the fourth quarter, though we believe they'll be more than offset by strong growth in other regions.

We expect that our sales to do to these destinations will decline significantly in the fourth quarter, but we believe there'll be more than offset by strong growth in other regions.

The U S government defined the regulations to allow the U S industry to provide data center compute products to markets worldwide.

Speaker 3: The U.S. government designed the regulation to allow the U.S. industry to provide data center compute products to markets worldwide, including China.

<unk> China.

Speaker 3: Continuing to compete worldwide as the regulations encourage, promote U.S. technology leadership, spurs economic growth, and supports U.S. jobs.

Continuing to compete worldwide as the regulations encourage promote U S technology leadership.

<unk> economic growth.

And supports U S jobs.

For the highest performance levels the government requires licenses.

Speaker 3: For the highest performance levels, the government requires licenses.

Speaker 3: For lower performance levels, the government requires a streamlined prior notification process. And for products, even lower performance levels, the government does not require any notice at all.

For lower performance levels, the government requires a streamlined prior notification process and.

And for products, even lower performance levels. The government does not require any notice at all.

Speaker 3: Following the government's clear guidelines, we are working to expand our data center product portfolio to offer compliant solutions for each regulatory category, including products for which the U.S. government does not wish to have advance notice before each shipment.

Following the government's clear guidelines, we are working to expand our data center product portfolio to offer compliance solution for each regulatory category.

<unk> products for which the U S government does not wish to have advance notice before each shipment.

Speaker 3: We are working with some customers in China and the Middle East to pursue licenses from the U.S. government.

We are working with some customers in China, and the middle East to pursue pursue licenses from the U S government.

It is too early to know whether these will be granted for any significant amount of revenue.

Speaker 3: It is too early to know whether these will be granted for any significant amount of revenue.

Speaker 3: Many countries are awakening to the need to invest in sovereign AI infrastructure to support economic growth and industrial innovation.

Many countries are awakening to the need to invest in sovereign AI infrastructure to support economic growth and industrial innovation.

Speaker 3: With investments in domestic compute capacity, nations can use their own data to train LLMs and support their local generative AI ecosystem.

With investments in domestic compute capacity nations can use their own data to train and support their local general today. Our ecosystems. For example, we are working with India's government and largest tech companies, including Infosys reliance and Tata.

Speaker 3: For example, we are working with India's government and largest tech companies, including Infosys, Reliance, and Tata, to boost their sovereign AI infrastructure.

To boost their sovereign AI infrastructure.

Speaker 3: and French private cloud provider, Scaleway, is building a regional AI cloud based on NVIDIA H100, InfiniBand, and NVIDIA AI enterprise software to fuel advancement across France and Europe .

And French private cloud provider scale way is building a regional AI cloud based on Nvidia H 100, Infiniband and Nvidia AI enterprise software to fuel advancement across France and Europe.

Speaker 3: National investment in compute capacity is a new economic imperative and serving the sovereign AI infrastructure market represents a multi-billion dollar opportunity over the next few years.

National investment and compute capacity is a new economic imperative.

Serving the sovereign AI infrastructure market represents a multibillion dollar opportunity over the next few years.

Speaker 3: From a product perspective, the vast majority of revenue in Q3 was driven by the NVIDIA HGX platform based on our Hopper GPU architecture with lower contribution from the prior generation Ampere GPU architecture.

From a product perspective, the vast majority of revenue in Q3 was driven by the Nvidia <unk> platform based on our Hopper GPU architecture with lower contribution from the prior generation and peer GPU architecture.

Speaker 3: The new L40S GPU built for industry standard servers began to ship, supporting training and inference workloads across a variety of customers.

The new <unk> 40 S. GPU built for industry standard servers began to ship supporting training and inference workloads across a variety of customers.

This was also the first revenue quarter of our G. H 200, Grasshopper Super Chip, which combines our arm based CPU with a hopper GPU.

Speaker 3: This is also the first revenue quarter of our GH200 Grace Hopper Superchip, which combines our ARM-based Grace CPU with a Hopper GPU.

Speaker 3: Grace and Grace Hopper are ramping into a new multi-billion dollar product line.

Grace and Grace Hopper are ramping into a new multibillion dollar product line.

Speaker 3: Grace Hopper instances are now available at GPU specialized cloud providers and coming soon to Oracle Cloud.

Hopper instances are now available at GPU specialized cloud providers and coming soon to Oracle cloud.

Speaker 3: Grace Hopper is also getting significant traction with supercomputing customers. Initial shipments to Los Alamos National Lab and the Swiss National Supercomputing Center took place in the third quarter.

Grace Hopper is also getting significant traction with supercomputing customers.

Initial system shipments to Lasalle, Los Alamos National Lab, and the Swiss National Supercomputing Center took place in the third quarter.

The UK government announced it will build one of the worlds fastest AI supercomputers.

Speaker 3: The UK government announced it will build one of the world's fastest AI supercomputers called eSambar AI with almost 5,500 Grace Hopper superchips.

<unk> AI with almost 5500 Grace Hopper Super chips.

German Supercomputing Center Buluk also announced that it will build its next generation AI supercomputer with close to 24000, Grace Hopper Super chips and content to Infiniband, making it the world's most powerful AI supercomputer went over 90 extra flops.

Speaker 3: German Supercomputing Center, EULEC, also announced that it will build its next generation AI supercomputer with close to 24,000 Grace Hopper superchips and Quantum II InfiniBand, making it the world's most powerful AI supercomputer with over 90 exaflops of AI performance.

AI performance.

Speaker 3: All in, we estimate that the combined AI compute capacity of all the supercomputers built on Great Hopper across the US, Europe , and Japan next year will exceed 200 exaflops, with more wins to come.

All in we estimate that the combined AI compute capacity of all the supercomputers build on Grace Hopper.

The U S Europe, and Japan next year will exceed 200 extra flops with more wins to come.

And France is contributing significantly to our data center demand as AI is now in full production for deep learning recommend first chat.

Speaker 3: inference is contributing significantly to our data center demand as AI is now in full production for deep learning, recommend, search.

Speaker 3: chatbots, copilots, and text-to-image generation. And this is just the beginning. NVIDIA AI offers the best inference performance and versatility, and thus, the lower power and cost of ownership.

<unk> bought co pilots and tests to English and French generation and this is just the beginning.

Nvidia AI offers the best in class performance, and Brussels holiday and thus the lower power and cost of ownership.

We are also driving a fast cost reduction curve with the release of Nvidia tensor RT LLM, we now achieved more than two X the inference performance for <unk>.

Speaker 3: We are also driving a fast cost reduction curve. With the release of NVIDIA Tensor RT LLM, we now achieve more than 2x the inference performance or half the cost of inferencing LLMs on NVIDIA GPUs.

Half the cost of inferencing LLM on Nvidia Gpus.

Speaker 3: We also announced the latest member of the Hopper family, VH200, which will be the first GPU to offer HBM3E, faster, larger memory, to further accelerate generative AI and LLM.

We also announced the latest member of the Hopper family Th 200.

Which would be the first GPU to offer HBM three.

Faster larger memory to further accelerate Jenner today I M O M.

Speaker 3: It boosts inference speed up to another 2x compared to H100 GPUs for running LLMs like Lama 2.

It boost infringed speed up to another two X compared to <unk> hundred Gpus for running <unk> like Lama too.

Speaker 3: Combined, TensorRT LLM and H200 increased performance or reduced cost by 4x in just one year without customers changing their stock. This is a benefit of CUDA and our architecture compatibility.

Combined tensor RT <unk> and H 200 increase performance reduce cost by Forex in just one year with our customers changing their stock. This is a benefit of cuda and our architecture compatibility.

Speaker 3: Compared to the A100, H200 delivers an 18x performance increase for infancy models like GPT-3, allowing customers to move to larger models and with no increase in latency.

Compared to the E 100, H 200 delivers an 18 ex performance increase for infantry models like JBT three allow customers to customers to move to larger models and with no increase in latency.

Speaker 3: Amazon Web Services, Google Cloud, Microsoft Azure, and Oracle Cloud will be among the first CSPs to offer H200-based instances starting next year.

Amazon Web services, Google Cloud, Microsoft Azure, and Oracle cloud will be among the first CSP to offer H 200 based instances starting next year.

At last week's Microsoft Ignite, we deepened and expanded our collaboration with Microsoft across the entire stock reentered.

Speaker 3: We introduced an AI foundry service for the development and tuning of custom generative AI enterprise applications running on Azure.

We introduced an AI foundry service for the development and tuning of custom generative AI enterprise applications running on Azure.

Speaker 3: Customers can bring their domain knowledge and proprietary data, and we help them build their AI models using our AI expertise and software stack in our DGX Cloud, all with enterprise-grade security and support. SAP and Amdocs are the first customers of the NVIDIA AI Foundry service on Microsoft Azure.

Customers can bring their domain knowledge and proprietary data and we help them build their AI models, using our AI expertise and software stock and our <unk> cloud.

All with enterprise grade security and support.

SAP.

In Amdocs are the first customers of the Nvidia AI foundry service on Microsoft Azure.

Speaker 3: In addition, Microsoft will launch new confidential computing instances based on the H-100.

In addition, Microsoft will launch new confidential computing instances based on the H 100.

Th 100 remains the top performing and most versatile platform for AI training and by a wide margin as shown in the latest MLB curve industry benchmark results.

Speaker 3: Our training cluster included more than 10,000 H100 GPUs or 3X more than in June , reflecting very efficient scaling. Efficient scaling is a key requirement in generative AI because LLMs are growing by an order of magnitude every year. Microsoft Azure achieved similar results on a nearly identical cluster, demonstrating the efficiency of NVIDIA AI in public cloud deployment.

Our training cluster included more than 10800, Gpus or three X more than in June reflecting very efficient scaling.

Efficient scaling is a key requirement in general today odd because our loans are growing by an order of magnitude every year.

Microsoft Azure achieved similar results on a nearly identical cluster demonstrating the efficiency of Nvidia AI and public cloud deployment.

Speaker 3: Networking now exceeds a 10 billion dollar annualized revenue run.

Networking now exceed a $10 billion annualized revenue run rate.

Speaker 3: Strong growth was driven by exceptional demand for InfiniBand, which grew five-fold year-on-year. InfiniBand is critical to gaining the scale and performance needed for training LLM.

Strong growth was driven by exceptional demand for Infiniband, which grew five fold year on year.

Infiniband is critical to gaining the scale and performance needed for training.

Yes.

Speaker 3: Microsoft made this very point last week highlighting that Azure uses over 29,000 miles of in Cinnabon tabling. Enough to circle the globe, we are expanding and video.

Microsoft made this very point last week, highlighting the azure users over 29000 miles of Infiniband table it.

Enough to circle the globe.

We are expanding in video networking into the Ethernet space, our new spectrum ex end to end Ethernet offering with technologies purpose built for AI will be available in Q1 next year.

Speaker 3: Our new Spectrum X and N Ethernet offering with technologies purpose built for AI will be available in Q1 next year with support from leading OEMs including Dell HPE and Lenovo.

Support from leading Oems, including Dell HP and Lenovo.

Spectrum X 10 achieved one six X higher networking performance for AI communication compared to traditional Ethernet offerings.

Speaker 3: Spectrum X can achieve 1.6x higher networking performance for AI communication compared to traditional Ethernet.

Speaker 3: Let me also provide an update on our software and services offerings, where we are starting to see excellent adoption. We are on track to exit the year at an annualized revenue run rate of 1 billion for recurring software support and services offerings. We see 2 primary opportunities for growth over the intermediate term. With our DGX cloud service and with our NVIDIA AI enterprise software.

Let me also provide an update on our software and services offerings, where we are starting to see excellent adoption. We are on track to exit the year at an annualized revenue run rate of $1 billion for our recurring software support and services offerings.

We see two primary opportunities for growth over the intermediate term.

With our <unk> cloud service and with our Nvidia AI Enterprise software.

Speaker 3: Each reflects the growth of enterprise AI training and enterprise AI inference, respectively.

Each reflects the growth of enterprise AI training and enterprise AI inference, respectively.

Our latest <unk> cloud customer announcement was this morning as part of an AI research collaboration with Genentech. The biotechnology pioneer also plans to use our buyer Nemo LLM framework to help accelerate and optimize their AI drug discovery platform.

Speaker 3: Our latest DGX Cloud customer announcement was this morning as part of an AI research collaboration with Genentech, the biotechnology pioneer also plans to use our BioNemo LLM framework to help accelerate and optimize their AI drug discovery platform.

Speaker 3: We now have enterprise AI partnerships with Adobe, Dropbox, Getty, SAP, ServiceNow, Snowflake, and others to come.

We now have enterprise AI partnerships with Adobe Dropbox Getty.

Service, now snowflake and others to come.

Okay moving to gaming.

Gaming revenue.

Speaker 3: Gaming revenue of $2.86 billion was up 15% sequentially and up more than 80% year-on-year, with strong demand in the important back-to-school shopping season, with NVIDIA RTX ray tracing and AI technologies now available at price points as low as $299. We enter the holidays with the best ever lineup for gamers and creators.

2.86 billion was up 15% sequentially and up more than 80% year on year.

With strong demand in the important back to school shopping season with.

With Nvidia <unk> Ray tracing and AI technologies now available at price points as low as $299, we enter the holidays with the best ever lineup for gamers and creators.

Speaker 3: Gaming has doubled relative to pre-COVID levels, even against the backdrop of lackluster PC market performance.

Gaming has doubled relative to pre COVID-19 levels, even against the backdrop of lack luster PC market performance.

Speaker 3: This reflects the significant value we've brought to the gaming ecosystem with innovations like RTX and DLSS.

This reflects the significant value we brought to the gaming ecosystem with innovations like <unk> and Elisa.

Speaker 3: The number of games and applications supporting these technologies has exploded in that period, driving upgrades and attracting new buyers.

The number of games and applications supporting these technologies has exploded in up period, driving upgrades and attracting new buyers.

Speaker 3: The RTX ecosystem continues to grow. There are now over 475 RTX-enabled games and applications.

<unk> ecosystem continues to grow.

Now over 475, RPX enabled games and applications.

Generative AI is quickly emerging as the new killer App for high performance Pcs, Nvidia RPX Gpus to find the most performance AI Pcs and workstations.

Speaker 3: Generative AI is quickly emerging as the new pillar app for high-performance PCs. NVIDIA RTX GPUs define the most performance AI PCs and workstations.

Speaker 3: We just released TensorRT LLM for Windows, which speeds on-device LLM inference up by 4x. With an installed base of over 100 million, NVIDIA RTX is the natural platform for AI application developers.

We just released tensor RT LLM for Windows, which speeds on device LLM inference up by four X with an installed base of over 100 million Nvidia RPX has been natural platform for AI application developers.

Finally, our G. First now cloud gaming service continues to build momentum its library of PC games surpassed 1700 titles, including the launches of our wave two.

Speaker 3: Finally, our GeForce Now cloud gaming service continues to build momentum. Its library of PC games surpassed 1,700 titles, including the launches of Alan Wake 2, Baldur's Gate 3, Cyberpunk 2077, Phantom Liberty, and Starfield.

K three cyberpunk 2077 <unk>.

Liberty and Star field.

Moving to programs.

Speaker 3: Revenue of $416 million was up 10% sequentially and up 108% year-on-year. NVIDIA RTX is the workstation platform of choice for professional design, engineering, and simulation use cases, and AI is emerging as a powerful demand driver.

Revenue of $416 million was up 10% sequentially and up 108% year on year end.

Nvidia RPX is the workstation platform of choice for professional design engineering and simulation use cases, and AI is emerging as a powerful demand driver.

Speaker 3: Early applications include inference for AI imaging in healthcare and edge AI in smart spaces and the public sector.

Early applications include <unk> for AI imaging, and healthcare and edge, AI and smart spaces and the public sector.

We launched a new mine of desktop workstations based on Nvidia RPX Ada Lovelace generation Gpus and connect X Smart next offering up to two X. The AI processing Ray tracing and graphics performance of the previous generations.

Speaker 3: We launched a new line of desktop workstations based on NVIDIA RTX Ada Lovelace generation GPU and ConnectX SmartNEXT, offering up to 2x the AI processing, ray tracing, and graphics performance of the previous generation.

Speaker 3: These powerful new workstations are optimized for AI workloads, such as fine-tuning AI models, training smaller models, and running inference locally.

These powerful new workstations are optimized for AI workloads, such as fine tuning AI models transpire model and running inference locally.

Speaker 3: We continue to make progress on Omniverse, our software platform for designing, building, and operating 3D virtual worlds.

We continue to make progress on our omni versus our software platform for designing building and operating three D virtual worlds.

Speaker 3: Mercedes Benz is using Omniverse-powered digital twins to plan, design, build, and operate its manufacturing and assembly facilities, helping it increase efficiency and reduce.

Mercedes Benz is using omni versus power digital twins to plan design build and operate its manufacturing and assembly facilities, helping it increase efficiency and reduce defects.

Speaker 3: Oxon is also incorporating Omniverse into its manufacturing process, including end-to-end simulation for the entire robotics and automation pipeline, saving time and cost.

Exxon is also incorporating <unk> into its manufacturing process, including end to end simulation for the entire robotics and automation pipeline saving time and cost.

Speaker 3: We have 2 new on reverse cloud services for automotive digitalization available on Microsoft Azure, a virtual factory simulation engine and autonomous vehicle simulation engine.

Now two new on reverse cloud services for automotive digital monetization available on Microsoft Azure, a virtual factory simulation engine and autonomous vehicle simulation engine.

Speaker 3: Moving to automotive. Revenue was $261 million, up 3% sequentially, and up 4% year-on-year, primarily driven by continued growth in self-driving platforms based on NVIDIA DRIVE ORIN SOC and the ramp of AI cockpit solutions with global OEM customers.

Moving to automotive.

Revenue was $261 million up 3% sequentially and up 4% year on year, primarily driven by continued growth in self driving platform based on Nvidia drive Orange associates, and the ramp of AI cockpit solutions with global OEM customers.

Speaker 3: We extended our automotive partnership with Foxconn to include NVIDIA DRIVE SOLARC, our next-generation automotive SOC.

We extended our automotive partnership with Fox Con to include Nvidia drive for our next generation automotive associated.

Speaker 3: Foxconn has become the ODM for EVs. Our partnership provides Foxconn with a standard AV sensor and computing platform for their customers to easily build a state-of-the-art, safe, and secure software-defined car.

Foxconn has become the ODM for EV.

Our partnership <unk>.

<unk> Fox Con with a standard Avi sensor and computing platform for their customers to easily build a state of an art safe and secure software defined car.

Now, we're going to move to the rest of the P&L.

Speaker 3: Gap gross margin expanded to 74%, and non-gap gross margin to 75%, driven by higher data center sales and lower net inventory reserves, including a 1 percentage point benefit from the release of previously reserved inventory related to the Ampere GPU architecture.

GAAP gross margin expanded to 74% and non-GAAP gross margin to 75% driven by higher data center sales and lower net inventory reserves, including a one percentage point benefit from the release of previously reserved inventory.

Related to the <unk> GPU architecture products.

Sequentially GAAP operating expenses were up 12% and non-GAAP operating expenses were up 10%, primarily reflecting increased compensation and benefits.

Speaker 3: Sequentially, GAAP operating expenses were up 12% and non-GAAP operating expenses were up 10%, primarily reflecting increased compensation and benefits.

Let me turn to the fourth quarter of fiscal 2024.

Speaker 3: Let me turn to the fourth quarter of fiscal 2024.

Total revenue is expected to be 20 billion plus or minus 2%.

Speaker 3: Total revenue is expected to be $20 billion, plus or minus 2%.

Speaker 3: We expect strong sequential growth to be driven by data center with continued strong demand for both compute and networking.

We expect strong sequential growth to be driven by data center with continued strong demand for both compute and networking.

Speaker 3: Gaming will likely decline sequentially as it is now more aligned with notebook seasonality.

Amy will likely decline sequentially that is now as it is now more aligned with notebook seasonality.

Speaker 3: Gap and non-gap gross margins are expected to be 74.5 and 75.5 respectively, plus or minus 50 basis points.

GAAP and non-GAAP gross margins are expected to be $74, five and 75, 5%, respectively, plus or minus 50 basis points.

Speaker 3: GAAP and non-GAAP operating expenses are expected to be approximately $3.17 billion and $2.2 billion, respectively. GAAP and non-GAAP other income and expenses are expected to be an income of approximately $200 million, excluding gains and losses from non-affiliated investments.

GAAP and non-GAAP operating expenses are expected to be approximately $3, one 7 billion and $2 2 billion, respectively, GAAP and non-GAAP. Other income and expenses are expected to be an income of approximately $200 million, excluding gains and losses from non affiliated.

Yes.

GAAP and non-GAAP tax rates are expected to be 15% plus or minus 1%, excluding any discrete items.

Speaker 4: Gap and non-gap tax rates are expected to be 15% plus or minus 1% excluding any discrete items.

Further financial information.

Speaker 3: are included in the CFO commentary and other information available on our IR website.

Included in the CFO commentary and other information available on our IR website in closing, let me highlight some upcoming events for the financial community. We will attend the UBS Global Technology Conference in Scottsdale, Arizona on November 28th Wells Fargo, TMT Summit and Renshaw.

Speaker 3: In closing, let me highlight some upcoming events for the financial community. We will attend the UBS Global Technology Conference in Scottsdale, Arizona on November 28th, the Wells Fargo QMC Summit in Rancho Palo Verde, California on November 29th, the ARETA Virtual Text Conference

Palo Verde to California on November 29.

Iretta virtual <unk> conference on December seven.

Speaker 4: and the J.P. Morgan Healthcare Conference in San Francisco on January 8th.

And the Jpmorgan healthcare conference in San Francisco on January eight.

Speaker 4: Our earnings call to discuss the results of our fourth quarter and fiscal 2024 is scheduled for Wednesday, February 21st.

Our earnings call to discuss the results of our fourth quarter and fiscal 2024 is scheduled for Wednesday February 21st.

Speaker 4: We will now open the call for questions. Operator, will you please call for questions?

We will now open the call for questions. Operator will you. Please poll for questions.

Yeah.

Speaker 1: At this time, I would like to remind everyone in order to ask a question, please press start and the number one on your telephone keypad will pause for just a moment to compile the Q and a roster.

At this time I would like to remind everyone in order to ask a question. Please press star and the number one on your telephone keypad.

For just a moment to compile the Q&A roster.

Yeah.

As a reminder, please limit yourself to one question.

Your first question comes from the line of Vivek Arya of Bank of America. Your line is open.

Speaker 1: first question comes from the line of Vivek area of Bank of America. Your line is open.

Speaker 5: Thanks for taking my question, just collect wanted to clarify what China contributions are you expecting in q4 and then gentlemen, the main question is for you, where do you think we are in the adoption curve?

Thanks for taking my question just wanted to clarify what China contribution that you're expecting in Q4, and then Jensen. The the main question is for you where do you think we are in the adoption curve.

Speaker 5: in terms of your shipments into the generative AI market, because when I just looked at the trajectory of your data center.

In terms of your shipments into degenerative AI market, because when I just look at the trajectory of your data center.

Speaker 5: is growth. It will be close to nearly 30% of all the spending in data center next year. So what metrics are you keeping an eye on to inform you that you can continue to grow? Just where are we in the adoption curve of your products into the generative AI?

It will be close to nearly 30% of all the spending and data center next year. So what metrics are you keeping an eye on to inform you that you can continue to grow just where are we in the adoption curve of your products into the agenda to be odd market. Thank you.

So first let me start with your question Vivek on export controls and the impact that we're seeing.

Speaker 3: First, let me start with your question, but back on export controls and the impact that we are seeing in our Q4 outlook and guidance that we've provided.

In our Q4 outlook and guidance that we provided.

Speaker 3: We had seen historically over the last several quarters that China and some of the other impacted destinations to be about 20 to 25% of our data center revenue.

We had seen historically over the last several quarters that China and some of the other impacted destinations to be about 20% to 25% of our data center revenue, we are expecting in our guidance.

Speaker 4: We are expecting in our guidance for that to decrease substantially as we move into Q4.

For that to decrease substantially as we move into Q4.

Speaker 4: The export controls will have a negative effect on our China business. And we do not have good visibility into the magnitude of that impact, even over the long term.

The export control will have a negative effect on our China business and we do not have good.

Visibility into the magnitude of that impact even over the long term.

Speaker 3: We are though working to expand our data center product portfolio to possibly offer new regulation compliant solutions that do not require a license.

We are though working to expand our data center product portfolio.

Possibly offered new regulation compliance solutions that do not require a license.

Speaker 4: These products, they may become available in the next coming months, however, we don't expect their contribution to be material or meaningful as a percentage of the revenue.

These products they may become available in the next coming months.

However, we don't expect their contribution to be material or meaningful as a percentage of the revenue in Q4.

Generative AI is the <unk>.

Speaker 6: largest TAM expansion of software and hardware that we've seen in several decades.

A larger Tam expansion.

Software and hardware that we've seen in several decades.

The.

Speaker 6: The at the core of it, what's really exciting is that that what was largely a retrieval based computing approach, almost everything that you do is retrieved off of storage somewhere has been augmented now.

At the core of it what's really exciting is that that was largely a retrieval based computing approach almost everything that you do is retrieved office stores somewhere.

Has been augmented now.

Had it with a general test method.

Speaker 6: And it's changed almost everything. You could see that text-to-text, text-to-image, text-to-video, text-to-3D, text-to-protein, text-to-chemicals. These were things that were.

And it's changed almost everything you could see that.

Tax the tax techs damage techs. The video takes this <unk> protein petrochemicals.

These were things that were <unk>.

<unk>.

You know.

Speaker 6: Typed in by humans in the past and these are now generative approaches the way that we access data is

Typed in by humans in the past and these are now generative approaches the way that we access data has changed.

Speaker 6: It used to be based on explicit queries, it is now based on natural language queries, intention queries, semantic queries.

It used to be based on explicit queries. It is now based on natural language query and hedge inquiries for magic words.

Speaker 6: And so, we're excited about the work that we're doing with SAP and Dropbox and many others that you're gonna hear about.

And so on.

<unk>.

Excited about the work that we're doing with S&P and Dropbox and many others that you're going to hear about.

Speaker 6: And one of the areas that is really impactful is the software industry, which is about a trillion dollars or so, has been building tools.

And and one of the areas that is really impactful.

The software industry, which is about a build a trillion dollars or so.

Has been building tools.

Speaker 6: that are manually used over the last couple of decades, and now there's a whole new segment of software called Copilots and Assistants. Instead of manually used, these tools will have Copilots to help you use them.

That are manually used over the last couple of decades and now there is a whole new set.

Segment of software called co pilots assistance instead of instead of manually used these tools will have co pilots to help you use it.

Speaker 6: And so instead of licensing software, we'll continue to do that, of course, but we will also hire co-pilots and assistants to help us use the software. We'll connect all of these co-pilots and assistants into teams of AIs, which is gonna be the modern version of software, modern version of enterprise business software.

And so instead of licensing software.

We'll continue to do that of course, but we will also higher co pilots and assistance to help US use these used to software will connect all of these copilot and assistance into teams.

Of of ice, which is going to be the modern version of software modern version of.

Enterprise business software.

Speaker 6: And so, so the transformation of software and the way that software is done is driving the hardware underneath. And you can see that that.

And so some of the transformation of <unk>.

Software and the way that software has done is driving the hardware underneath.

And you could see that that it's transforming.

Speaker 6: hardware in two ways. One is something that's largely independent of generative AI. There's two trends.

Hardware in two ways. One is there is something that's.

Largely independent of degenerative AI there is two trends.

Speaker 6: One is related to accelerated computing. General purpose computing is too wasteful of energy and cost.

One is related to accelerated computing general purpose to general purpose computing is too wasteful of energy and cost.

Speaker 6: And now that we have much, much better approaches, called accelerated computing, you could save an order of magnitude of energy, you could save an order of magnitude of time, or you could save an order of magnitude of cost by using acceleration. And so accelerated computing is transitioning, if you will, general purpose computing into this new.

And now that we have.

Much much better approaches called accelerated computing you could save in order of magnitude of energy you can save in order of magnitude of time or you can say 19 should cost.

By using acceleration and so accelerated computing is transitioning if you will general purpose computing into this new approach.

Speaker 6: And that's been augmented by a new class of data centers. This is the traditional data centers that you were just talking about, where we represent about a third of that. But there's a new class of data centers. This new class of data centers is unlike the data centers of the past.

That has been augmented.

By a new class of data centers.

The traditional.

Data centers that you were just talking about where we represent about a third of that but there is a new class of data centers. This new class of data centers. Unlike the datacenters in the past.

Speaker 6: where you have a lot of applications running used by a great many people that are different tenants that are using the same infrastructure, and that data center stores a lot of files.

Where you have.

A lot of applications running used by a great. Many.

People that are different tenants that are using the same infrastructure.

Data center stores a lot of files.

Speaker 6: These new data centers are very few applications, if not one application, used by basically one tenant.

<unk> new data centers are.

Very few applications, if not one application used by basically one tenant.

Speaker 6: And it processes data, it trains models, and it generates tokens, it generates AI. And we call these new data centers AI factories.

And bid processes data and trains models and it generates tokens to generate.

And we call. These new data centers factories, we're seeing AI factories being built built out.

Speaker 6: We're seeing AI factories being built out everywhere, in just about every country.

<unk>.

Just about every.

Every country.

Speaker 6: And so if you look at the way, where we are in the expansion, the transition into this new computing approach.

<unk>.

So if you look at the way, where we are in the expansion.

The transition to this new computing approach. The first wave you saw with large language model startups generative AI startups and.

Speaker 6: The first wave you saw with large language model startups, generative AI startups, and consumer internet companies. And we're in the process of

Consumer Internet companies.

In the process of ramping that.

Speaker 6: Meanwhile, while that's being ramped, you see that we're starting to partner with enterprise software companies who would like to build chatbots and copilots.

Meanwhile, that's b ramp.

You see that we're starting to partner with enterprise software companies, who we'd like to Bill chat bots and co pilots.

Speaker 6: and assistance to augment the tools that they have on their platform.

And assistance to augment the tools that they have on their platforms.

Speaker 6: You're seeing GT specialized CSPs cropping up all over the world.

Seeing GPU specialized csp's cropping up all over the world.

Speaker 6: And they're dedicated to doing really one thing, which is processing AI.

And they are dedicated to doing really one thing which is processing AI.

Speaker 6: You're seeing sovereign AI infrastructures, people, countries that now recognize that they have to utilize their own data, keep their own data, keep their own culture, process that data, and develop their own AI. You see that in.

Are you seeing sovereign AI infrastructures people.

Countries that now recognize that they have to utilize their own data keep their own data keeps our own culture, a process that data and develop their own AI. So.

You see that in India.

Speaker 6: about a year ago in Sweden, a year seen in Japan, last week a big announcement in France. But the number of sovereign AI clouds that are being built is really quite significant. And my guess is that almost every major region will have, and surely every major country will have their own AI cloud.

Several several.

About a year ago in Sweden are youre seeing in Japan last week.

Big announcement in France.

The number of sovereign AI clouds that are being built.

It's really quite quite significant and my guess is that almost every major region, we'll have.

And surely in every major country will have their own clouds.

Speaker 6: And so I think you're seeing just new developments as the generative AI wave propagates through every industry, every company, every region. And so we're at the beginning of this inflection, this computing transition.

And so I think youre seeing just new new developments as they are generally to the AI wave.

<unk> through every industry.

Every company every region and so where we're at the beginning of this of this this inflection does this computing transition.

Your next question comes from the line of Aaron Rakers of Wells Fargo. Your line is open.

Speaker 1: Your next question comes from the line of Aaron Rickers of Wells Fargo. Your line is open.

Speaker 7: Yeah, thanks for taking the question. I wanted to ask about kind of the networking side of the business. You know, given the growth rates that you've now cited, I think it's 155 percent year over year and strong growth sequentially. It looks like that business is like almost approaching a two and a half to three billion dollar quarterly level.

Yes, thanks for taking the question.

I wanted to ask about kind of the networking side of the business.

Given the growth rates that you have now.

Cited I think it is 155% year over year and strong growth sequentially. It looks like that business is like almost approaching $2 $5 billion to $3 billion quarterly level.

Speaker 7: I'm curious of how you see Ethernet evolve, you know, evolving, and maybe how you would characterize your differentiation of SpectrumX relative to the traditional Ethernet stack as we start to think about that becoming part of the networking narrative above and maybe beyond just InfiniBand as we look into next year. Thank you.

I'm curious of how you see Ethernet involved evolving and maybe how you would characterize your differentiation of spectrum Max relative to the traditional Ethernet stack as we start to think about that becoming part of the networking narrative by Bob and maybe beyond just infiniband as we look into next year. Thank you.

Yes. Thanks for the question our networking businesses already under 10 billion dollar plus run rate and term.

Speaker 6: Yeah, thanks for the question. Our networking business is already at a 10 billion dollar plus run rate and it's going to get much larger.

It's going to get much larger.

Speaker 6: And as you mentioned, we added a new networking platform to our networking business recently.

And as you mentioned, we added a new.

Networking platform to our networking business recently.

Yes.

Speaker 6: The vast majority of the dedicated large-scale AI factories.

The vast majority of the dedicated large scale factories.

Speaker 6: standardized on InfiniBand. And the reason for that is not only because of its data rate and not only just the latency, but the way that it moves traffic around the network is really important. The way that you process AI and the a multi-tenant hyperscale Ethernet environment, the traffic pattern is just radically different.

Standardized on Infiniband and the reason for that is not only because of its <unk>.

Data rate and not only just.

The latency, but the way that the move traffic around the network is really important the weight of your process AI envy.

A multi tenant hyperscale Ethernet environment, the traffic patterns just radically different.

Speaker 6: And with InfiniBand and with software-defined networks, we could do congestion control, adaptive routing, performance isolation and noise isolation, not to mention, of course, the data rate and the low latency and the very low overhead of InfiniBand that's natural part of InfiniBand. And so InfiniBand is not so much

And.

With with Infiniband and with software defined networks, we could do congestion control.

Our GAAP adaptive routing.

Performance isolation and noise isolation not to mention of course, the data rate in the low latency that in the very low overhead of Infiniband is natural part of Infiniband and so.

Infiniband is is not so much.

Speaker 6: just a network, it's also a computing fabric. We put a lot of software-defined capabilities into the fabric, including computation. We will do floating-point calculations and computation right on the switch and right in the fabric itself.

Just a network. It's also a computing fabric, we put a lot of software defined capabilities ended a fabric, including computation and we will do to 40 point calculations and computation right on the switch and writing the fabric itself and.

And so that's the reason why.

Speaker 6: that difference in Ethernet versus InfiniBand or InfiniBand versus Ethernet for AI factories is so dramatic. And the difference is profound. And the reason for that is because you've just invested in a $2 billion infrastructure for AI factories, a 20, 25, 30% difference in overall effectiveness, especially as you scale up, is measured in hundreds of millions of dollars of value.

The difference in Ethernet versus Infiniband room for an event versus Ethernet for AIG factories is so dramatic and and the difference the differences is.

A profound and reason for that is because we've just invested in a $2 billion infrastructure for our factories.

A 2025, 30% difference in.

And overall effectiveness, especially as you scale up is measured in hundreds of millions of dollars of value.

Speaker 6: And if you were renting that infrastructure over the course of four or five years, it really, really adds up. And so InfiniBand's value proposition is undeniable for AI factor.

And if you were rent renting that infrastructure over the course of four or five years.

It really really adds up and so <unk> value proposition is undeniable for AI factories, however, as we move AI into enterprise.

Speaker 6: However, as we move AI into enterprise.

Speaker 6: This is enterprise computing where we'd like to enable every company to be able to build their own custom AI.

This is enterprise computing, what we'd like to two.

Enable every company to be able to build their own custom AI.

Speaker 6: We're building custom AIs in our company based on our proprietary data, our proprietary type of skills. For example, recently we spoke about one of the models that we're creating, it's called Chipnemo. We're building many others. There'll be tens, hundreds of custom AI models that we create inside our company.

We're building customer <unk> and our company is based on our proprietary data are proprietary.

Type of skills. For example, the recently we spoke about on one of the models that we're creating as called ship Nemo. We're building many others there'll be there'll be tens hundreds of custom AI models that we create inside our company.

Speaker 6: And our company is, you know, for all of our employees.

And our company is.

For all of our employees it doesn't have to be as high performance as the factories, we used to train the models and so we would like the.

Speaker 6: It doesn't have to be as high performance as the AI factories we use to train the models. And so we would like the AI to be able to run in an Ethernet environment.

AI to be able to run an Ethernet environment.

Speaker 6: And so what we've done is we invented this new platform that extends Ethernet, doesn't replace Ethernet, it's 100% compliant with Ethernet.

And so what we've done is we invented this new platform that extends Ethernet does it replace Ethernet is a 100% compliant with Ethernet and is optimized for east West traffic, which is which is where the computing fabric is.

Speaker 6: and it's optimized for east-west traffic, which is where the computing fabric is. It adds to Ethernet with an end-to-end solution with Bluefield, as well as our spectrum switch that allows us to perform some of the capabilities that we have in InfiniBand, not all but some. And we achieve...

It adds to Ethernet.

With the end to end solution with Bluefield.

As well as our spectrum switch.

That allows us to.

Perform some of the capabilities that we have infant infiniband not all but some.

We achieved excellent results.

Speaker 6: And the way we go to market is we go to market with our large enterprise partners.

And the way we go to market as we go to market with our large enterprise partners, who already offer our computing solution and so HP Dell and Lenovo has the Nvidia AI stack Nvidia AI enterprise software stack.

Speaker 6: who already offer our computing solution. And so HP, Dell, and Lenovo has the NVIDIA AI stack, the NVIDIA AI Enterprise software stack. And now they integrate with Bluefield, as well as Bundle take to market their Spectrum switch.

And now they integrate with bluefield as well as bundle.

I'll take the market there our spectrum switch and there'll be able to offer an enterprise customers all over the world Winter.

Speaker 6: And they'll be able to offer enterprise customers all over the world with their vast sales force and vast network of resources.

<unk> sales force.

Vast network of resellers.

Speaker 6: in a fully integrated, if you will, fully optimized at least, and AI.

And then in a fully integrated.

<unk>, if you will fully optimized at least.

And AI solution.

Speaker 6: So that's basically bringing AI to Ethernet for the world's enterprise.

And so so that's basically bringing bringing AI to ethernet or the world's enterprise.

Thank you. Your next question comes from the line of Joe Moore of Morgan Stanley. Your line is open.

Speaker 1: Your next question comes from the line of Joe Moore of Morgan Stanley . Your line is open.

Speaker 8: Great, thank you. I wonder if you could talk a little bit more about Grace Hopper and how you.

Great. Thank you I Wonder if you could talk a little bit more about great Hopper and how you see the ability to leverage the microprocessor or how do you see that as a as a Tam expander and what applications do you see using grace hopper versus more traditional H 100 applications.

Speaker 8: ability to leverage some of the microprocessor, how you see that as a as a TAN expander, and what applications do you see using a gray topper versus more traditional?

Yes, Thanks for question.

Speaker 6: Grace Hopper is in production.

Grace Hopper.

It is in production.

In high volume production.

Speaker 9: In high volume production now we're expecting next year just with all of the design ones that we have in high performance computing and AI AI infrastructures.

We're expecting next year, just with all of the design wins that we have.

And high performance computing and AI AI infrastructures.

Speaker 9: We are on a very, very fast ramp with our first data center CPU to a multi-billion dollar product line. This is going to be a very large product line for us.

We are we are.

We are on a very very fast ramp.

With our first data center CPU.

A multibillion dollar product line.

This is going to be a very large product line for us.

Speaker 6: The capability of Grace Hopper is really quite spectacular. It has the ability to create computations.

The the capability of the Grace Hopper is really quite quite spectacular.

It has the ability to create.

Computing nodes that.

Speaker 9: simultaneously has very fast memory as well as very large memory.

Simultaneously has very fast memory as well as very large memory.

Speaker 9: In the areas of vector databases or semantic search, what is called RAG, Retrieval Augmented Generation,

In the areas of Vectra databases or semantic search what is called Rag retrieval augment to generation.

So that you could you could have the agenda to the AI model.

Speaker 9: so that you could have a generative AI model be able to refer to proprietary data or factual data before it generates a response. That data is quite large. And you could also have applications or generative models where the context length is very high. You basically store.

Well to refer to proprietary nano or factual data.

Before it generates a response that data is quite large and.

You could also have applications or generative models, where the context length is very high.

Basically stored an entire book.

Speaker 9: into its system memory before you ask the questions. And so the context length can be quite large. This way, the generative models have the ability to still be able to naturally interact with you on one hand. On the other hand, be able to refer to factual data, proprietary data, or domain-specific data, your data, and be contextually relevant.

Into end to end system memory before you ask your questions and so the context length and can be quite large that's why these generative models has the ability to still being able to naturally interact with you.

On one hand on the other hand.

Being able to refer to backhaul data proprietary data or domain specific data your data.

And be contextually.

Relevant.

And so reduce hallucination and so so.

Speaker 9: and reduce hallucination. And so that particular use case, for example, is really quite fantastic for Grace Hopper. It also serves the customers that really care to have a different CPU than x86.

That particular use case for example is really quite fantastic for Grace Hopper.

It also serves as the customers that that really care to have a different Cpus and X 86, maybe.

Speaker 9: Maybe it's European supercomputing centers or European companies who would like to build up their own ARM ecosystem and like to build up a whole stack, or CSPs that have decided that they would like to pivot to ARM because their own custom CPUs are based on ARM. There are a variety of different reasons that drives the success of Grace Hopper, but we're off to just an extraordinary start. This is a home run product.

Maybe maybe.

Hi.

European.

Supercomputing centers or European companies, who would like to build up their own arm ecosystem and I would like to build up a full stack or csp's that have decided that they would like to pivot to arm.

Because their own custom Cpus are based on arm.

There are a variety of different reasons that that drives the success of Grace Hopper, but we're off to a just an extraordinary start this is a home loan product.

Speaker 1: Your next question comes from the line of Tim Arcuri of UBS. Your line is open.

Your next question comes from the line of Tim Arcuri of UBS. Your line is open.

Speaker 10: Hi, thanks. I wanted to ask a little bit about the visibility that you have on revenue. I know there's a few moving parts. I guess on one hand, the purchase commitments went up a lot again.

Hi, Thanks, I wanted to ask a little bit about the visibility that you have on revenue I know, there's a few moving parts I guess on one hand, the purchase commitments went up a lot again, but on the other hand, China bands.

Speaker 10: But on the other hand, China bans would arguably pull in when you can fill the demand beyond China.

Arguably pull in when you can fill that demand beyond China. So I know, we're not even into 2024, yet and it doesn't sound like Jensen you think that next year would be a peak in your data center revenue, but I just wanted to sort of explicitly ask you that do you think that.

Speaker 10: So I know we're not even into 2024 yet, and it doesn't sound like, Jensen, you think that next year would be a peak in your data center revenue, but I just wanted to sort of explicitly ask you that. Do you think that data center can grow even into 2025? Thanks.

Data center can grow.

Even into 2025.

Absolutely believes that data center can grow.

Speaker 9: through 2025, and there are, of course, several reasons for that. We are expanding our supply quite significantly. We have already one of the broadest and largest and most capable supply chain in the world. Remember.

Through 2025 and.

There are of course.

Several reasons for that.

We are expanding our supply.

Quite significantly we have already won.

The broadest and largest and most capable supply chain in the world.

Remember.

People think that the GPU is a chip.

Speaker 6: People think that the GPU is a chip. But the HGX H100, the Hopper HGX, has 35,000 parts. It weighs 70 pounds. Eight of the chips are Hopper.

But the <unk> Gx H 100, the Hopper <unk> at 35000 parts it weighs 70 pounds.

Eight of the chips are hopper.

The other 35000 or not.

Speaker 9: It is, it has, even its passive components are incredible, high voltage parts, high.

It is it has <unk>.

Given its passive components are incredible.

Voltage parts.

Frequency parts high current parts.

Speaker 9: high current parts. It is a supercomputer, and therefore the only way to test a supercomputer is with another supercomputer. Even the manufacturing of it is complicated, the testing of it is complicated, the shipping of it is complicated, and installation is complicated.

It is a supercomputer and therefore, the only way to test the supercomputers with another supercomputer given the the manufacturing of it is is complicated the testing of it is complicated to shipping event is complicated and installation is complicated.

Speaker 9: And so every aspect of our HGX supply chain is complicated, and the remarkable.

And so every aspect of <unk>.

Of.

H Gx supply chain is complicated and the remarkable.

Speaker 9: team that we have here has really scaled out the supply chain incredibly.

The team that we have here has really scaled out the supply chain incredibly.

Speaker 9: Not to mention, all of our HDXs are connected with NVIDIA networking, and the networking, the transceivers, the NICs, the cables, the switches, the amount of complexity there is just incredible. And so I'm just, first of all, I'm just super proud of the team for scaling up this incredible supply chain. We are absolutely world-class.

Not to mention.

All of our all of our <unk> are connected with Nvidia networking and the networking.

C versus the mix the cables the switches the amount of complexity. There is just incredible and so im just first of all just super proud of the team for scaling up this incredible supply chain, we are absolutely world class.

Speaker 9: But meanwhile, we're adding new customers and new products. So, we have new supply. We have new customers, as I was mentioning earlier. Different regions are standing up GPU specialist clouds. Sovereign AI clouds coming up from all over the world as people realize that they can't afford to export their country's knowledge, their country's culture for somebody else to then resell AI back.

But meanwhile, we're adding new customers and new products. So we have new supply we have new customers as I was mentioning earlier.

Different regions are are standing up.

<unk> specialist clouds.

Sovereign AI clouds coming up from all over the world as people realize that they can't afford to export their country's knowledge their country's culture.

For somebody else to then resell AI back to them.

Speaker 9: They have to, they should, they have the skills, and surely with us, in combination, we can help them do that, build up their national AI. And so the first thing that they have to do is create their AI cloud, national AI cloud. You're also seeing us now growing into enterprise.

They have to they should they have the skills and surely with us.

In combination we can help them do that buildup, there national AI and so the first thing that they have to do is create their AI cloud national AI cloud and you're also seeing us.

Now growing into enterprise.

Speaker 9: The enterprise market has two paths.

The enterprise.

Market has has to pass.

Speaker 9: One path, or if I could say three paths, the first path, of course, is off-the-shelf AI. And there are, of course, CatchyPG, the fabulous off-the-shelf AI, there'll be others. There's also a proprietary AI, because the software companies like Service...

One path.

Or if I could say three paths the first half of courses to off the shelf AI.

And are there.

<unk> <unk>.

Fabulous off the shelf AI there'll be others.

Theres also a proprietary AI because the software companies like service now in S&P.

Speaker 9: There are many, many others that can't afford to have their company's intelligence be outsourced to somebody else. And they are about building tools, and on top of their tools, they should build custom and proprietary and domain-specific co-pilots and assistants that they can then rent to their customer.

There are many many others I can't afford to have their company's intelligence be outsourced to somebody else and then they are about building tools and on top of their tools, they should build custom and proprietary and domain specific co pilots and assistance that they can.

Rent to their customer base.

Speaker 9: This is, they're sitting on a goldmine. Almost every major tools company in the world is sitting on a goldmine. And they recognize that. They have to go build their own custom AIs. We have a new service called an AI Foundry, where we leverage MVF capabilities to be able to serve them in that. And then the next one is enterprises building their own custom AIs, their own custom chatbots, their own custom regs.

They are sitting on a goldmine almost every major tool.

Towards company in the World is sitting on a goldmine and they recognize that they have to go build their own custom AI.

We have a we have a new service called an AI foundries, where we leverage <unk> capabilities to be able to serve them in that and then the next one is on <unk>.

Enterprises building their own custom.

Is their own custom chat bots do their own custom rags and this capability is.

Speaker 9: and this capability is spreading all over the world.

Spreading all over the world.

Speaker 9: And the way that we're going to serve that marketplace is with the entire stacks of systems, which includes our compute, our networking, and our switches.

And the way that we're going to serve that marketplace is.

With the entire.

Stocks of systems, which includes our compute our networking and our switches running our software stack called Nvidia AI enterprise, taking it through our market.

Speaker 9: running our software stack called NVIDIA AI Enterprise.

Speaker 9: taking it through our market, partners, HP, Dell, and Lenovo, so on and so forth. And so we're seeing the waves of generative AI starting from the startups and CSPs.

Partners, HP Dell and Lenovo.

So and so forth and so so we're just we're seeing the waves of generative AI star.

Starting from from the startups and Csp's.

Speaker 9: moving to consumer Internet companies, moving to enterprise software platforms, moving to enterprise companies, and then

Moving to consumer Internet companies moving to enterprise software platforms, moving to enterprise companies, and then and ultimately one of the areas that.

Speaker 9: and ultimately one of the areas that you guys have seen us spend a lot of energy on has to do with industrial generative AI. This is where NVIDIA AI and NVIDIA Omniverse comes together, and that is a really, really exciting.

You guys have seen us spend a lot of a lot of.

Energy on has to do with industrial generative AI. This is ware, Nvidia AI and Nvidia Omnivores comes together and that is that is really really exciting work and so I think the.

Speaker 9: And so I think we're at the beginning of a basically across the board.

We're at the beginning of a team.

It basically across the board industrial transition to generative AI to accelerated computing. This is going to affect every company every industry every country.

Speaker 9: to generative AI, to accelerated computing. This is gonna affect every company, every industry, every country.

Your next question comes from the line of <unk> Hari of Goldman Sachs. Your line is open.

Speaker 1: Your next question comes to the line of Tashia Hari of Goldman Sachs. Your line is open.

Speaker 11: Hi, thank you. I wanted to clarify something with Collette real quick, and then I had a question for Jensen as well. Collette, you mentioned that you'll be introducing regulation compliant products over the next couple of months, yet the contribution to Q4 revenue should be relatively limited. Is that a timing issue? And could it be a source of re-acceleration and growth for data center in April and beyond? Or are the price points such that?

Hi, Thank you I wanted to clarify something with collect real quick and then I had a question for Johnson as well.

You mentioned that you'll be introducing regulation compliant products over the next couple of months yet.

Contribution to Q4 revenue should be relatively limited.

Is that a timing issue and.

Could it be a source of reacceleration in growth for data center in April and beyond or are the price points such that the contribution to revenue going forward. It should be relatively limited and then the question for Jensen.

Speaker 11: the contribution to revenue going forward should be relatively limited. And then the question for Jensen, the AI Foundry service announcement from last week, I just wanted to ask about that and hopefully have you expand on it. How is the monetization model going to work? Is it primarily, you know, services and software revenue?

The AI foundry service announcement from last week I just wanted to.

To ask about that and hopefully have you expand on it.

How is the monetization model going to work is it primarily services and software revenue.

Speaker 11: you know, how should we think about the long-term opportunities, and is this going to be exclusive to Microsoft, or do you have plans to expand to other partners as well? Thank you.

Should we think about the long term opportunity set and is this going to be exclusive to Microsoft or do you have plans to expand to other partners as well. Thank you.

Okay. Thanks to share on the question regarding potentially new products that we could provide to our tenant comes to levels.

Speaker 3: Thanks, Tashia, on the question regarding potentially new products that we could provide to our China.

Speaker 3: It's a significant process to both design and develop these new products.

It's a significant process to both design and develop.

These new products.

Speaker 3: As we discussed, we're going to make sure that we are in full discussions with the US government of our intent in these products as well. Given our state about where we are in the quarter, we're already several weeks into the quarter. That's just going to take some time for us to go through and discussing with our customers. Other needs and desires of these 2 products that we.

We discussed we're going to make sure that we are in full discussions with the U S government of our intent in these products as well.

Given our stayed about where we are in the quarter were already several weeks into the quarter, let's just going to take some time.

For us to go through.

In discussing with our customers on their needs and desires of these new products that we have.

Speaker 3: Moving forward, whether that's medium-term or long-term, it's just hard to say both the ideas of what we can produce with the U.S. government and what the interest of our China customers is. So we stay still focused on finding that right balance for our China customers, but it's hard to say at this point.

Moving forward.

Whether that's medium term or long term, it's just hard to say.

Okay.

Most of what we can produce with the U S government and with the interest of our China customers. So.

So we stay still focused on finding the right balance for our China customers, but it is hard to say at this time.

Sure. Thanks for the question there is a glaring opportunity in the world for AI foundry.

Speaker 9: Thanks for the question. There is a glaring opportunity in the world for AI foundry.

And it makes so much sense.

First.

Okay.

Speaker 9: Every company has its core intelligence. It makes up our company, our data, our domain expertise.

Every company has its core intelligence it makes up our company.

Our data.

Our domain expertise.

In the case of many companies we create tools.

Speaker 9: and most of the software companies in the world are tool platforms and those

And most of the software companies in the world are tool platforms.

And those tools are used by people today and.

Speaker 6: And in the future, it's going to be used by people augmented with a whole bunch of AIs that we hire.

And in the future, it's going to be used by people augmented with a whole bunch of <unk> that we that we hire.

And.

Hi.

And these Dci platforms, just gotta go across go across the world and you'll see and we've already announced a few.

Service now.

Dropbox.

Getty.

Many others are coming and the reason for that is because they have their own proprietary AI.

Speaker 9: And the reason for that is because they have their own proprietary AI. They want their own proprietary AI. They can't afford to outsource their intelligence and hand out their data.

They want their own proprietary AI, they can't afford to outsource their intelligence and hand out their data and us handout.

Speaker 9: and hand out their flywheel for other companies to build the AI for them. And so they come to us. We have several things that are really essential in a foundry, just as TSMC is a foundry.

Handout their flywheel for other other companies too to build the AI for them and so they come to US we have some things that are really essential in the foundry just as TSMC as a boundary.

Speaker 9: You have to have AI technology, and as you know, we have just an incredible depth of AI capability, AI technology capability.

You have to have AI technology and as as you know we have just an incredible depth of AI capability AI technology capability.

Speaker 9: And then second, you have to have the best practice, known practice, the skills of processing data through the invention of AI models to create AIs that are guardrails, fine-tuned, so on and so forth, that are safe, so on and so forth.

And then second you have to have the best.

Best practice known practice the skills of processing data.

Through the invention of AI models to create.

AI Center.

That our guardrails fine tuned.

So on and so forth that are safe so on so forth.

Speaker 9: And the third thing is you need factories. And that's what DGX Cloud is. Our AI models are called AI foundations.

And and and the third thing is you need factories and Thats, what <unk> cloud is.

Our AI models are AI foundations.

Speaker 9: Our process, if you will, our CAD system for creating AIs are called NEMO, and they run on NVIDIA's factories we call DGX Cloud. Our monetization model is that with each one of our partners, they rent a sandbox on DGX Cloud where we work together. They bring their data. They bring their domain expertise. We bring our researchers and engineers. We help them build their custom AI. We help them make that custom AI.

Our process, if you will our CAD system for creating a ice are called Nemo and they run on Nvidia factories, we called <unk> cloud our monetization model.

That with each one of our partners they rent a <unk>.

Sandbox on <unk> cloud, where we worked together they bring their data they bring their domain expertise, we bring our researchers and engineers, we help them build their custom AI.

We help them make that customer incredible than.

Speaker 9: then that custom AI becomes theirs.

And then that customer becomes pairs.

Speaker 6: and they deploy it on a runtime that is.

And they deploy it on a run time that is.

Enterprise grade enterprise optimized our outperformance optimized runs across everything Nvidia, we have a giant installed base in the cloud on Prem anywhere.

Speaker 6: Enterprise-grade, enterprise-optimized, or performance-optimized, runs across everything NVIDIA. We have a giant installed base in the cloud, on-prem, anywhere.

Speaker 6: And it's secure.

And it's secure.

Speaker 6: securely patched, constantly patched, and optimized and supported. And we call that NVIDIA AI Enterprise. NVIDIA AI Enterprise is $4,500 per GP per year. That's our business model.

Securely patched constantly.

Astutely patched and.

Our optimized and supported as.

We call that Nvidia AI enterprise and medium enterprise is $4500 per GP per year, that's our business model.

Speaker 6: Our business model is basically a license. Our customers then, with that basic license, can build their monetization model on top of.

Our business model is is basically a license our customers then with that basic license can build their monetization model on top of it.

Speaker 6: In a lot of ways where wholesale, they become retail, they could have a per.

A lot of ways, where wholesale they become retail.

They could have they could have occur.

Speaker 6: They could have subscription license base, they could per instance or they could do per usage. There's a lot of different ways that they could take to create their own business model, but ours is basically like a software license, like an operating system. And so our business model is help you create your custom models. You run those custom models on NVIDIA AI Enterprise.

They could've subscription license base they could.

For instance are they could they could do per usage theres a lot of different ways that they could take them create their own business model, but ours is basically like a software license like an operating system and.

And so our business model is help you create your custom models you run those custom models.

On Nvidia AI enterprise.

Speaker 6: And it's off to a great, great start. Yeah. NVIDIA AI Enterprise is going to be a very large.

And it's off to a great start.

Enterprise is going to be a very large business for us.

Your next question comes from line of Stacy <unk> of Bernstein Research. Your line is open.

Speaker 1: Your next question comes from Stacy Rasgun of Bernstein Research, your line is open.

Hi, guys. Thanks for taking my questions.

Speaker 12: Hi, guys, thanks for taking my questions. Colette, I wanted to know if it weren't for the China restrictions, would the Q4 guide have been higher? Or are you supply constrained in just reshipping stuff that would have gone to China elsewhere? And I guess along those lines, if you give us a feeling for where your lead times are right now in data centers, is the China redirection such as this, is it lowering those lead times? Because you've got parts that are sort of immediately available to ship.

Scott I wanted to know if it weren't for the China restrictions would be Q4 guide had been higher.

Are you supply constrained in just reshaping stuff that would have gone to China elsewhere, and I guess, along those lines if we can.

Give us a feeling for where your lead times are right now in data center and just the China redirection, such as at lowering those lead times because <unk> got parts that are that are sort of immediately available to ship.

Speaker 3: Yeah, let me let me see if I can help you understand. Yes, the are still situations where we are working on both improving our supply each and every quarter. We've done a really solid job of ramping every quarter, which is defined our revenue.

Yeah.

Stacy let me, let me see if I can help you.

I understand yes. They are still are situations, where we are working on both improving our supply each and every quarter.

Done a really solid job of ramping every quarter, which is defined our revenue.

Speaker 3: But with the absence of China, for our outlook for Q4, sure, there could have been some things that we are not supply constrained that we could have sold to China, but we no longer can. So could our guidance have been a little higher in our Q4? Yes.

But with the absence of China.

For our outlook for Q4 sure there could have been some things that we are not supply constraint that we could have sold.

But we no longer can so could our guidance have been a little higher.

Our Q4, yes.

Speaker 3: We are still working on improving our supply plan on continuing growing all throughout next year as well to work for that.

We are still working on improving our supply.

Supply and plan on continuing growing all throughout next year as well towards that.

Speaker 1: Your next question comes from the line of Matt Ramsey of TD Cowan. Your line is open.

Your next question comes from the line of Matt Ramsay of TD Cowen Your line is open.

Speaker 13: Thank you very much. Congrats, everybody, on the results. Jensen, I had a two-part question for you, and it comes off of sort of one premise. And the premise is I still get a lot of questions from investors thinking about

Thank you very much Ah congrats everybody on the results.

Jensen I had a two part question for you and it comes off of sort of one premise and the premises.

I still get a lot of questions from investors I'm.

Thinking about AI.

Speaker 13: AI training as being NVIDIA's dominant domain and that.

AI training at being video dominant domain and that.

Speaker 13: somehow as inference, even large model inference, takes more and more of the time that the market will become more competitive, you'll be less differentiated, etc, etc. So I guess the two parts of the question are number one, maybe you could

In France, even large modeling friends takes more and more of the Tam that.

Market will become more competitive youll be less differentiated et cetera, et cetera. So I guess the two parts of the question our number one maybe you could.

Speaker 13: Spend a little bit of time talking about the evolution of the inference workload as we move to LLM and how your, your company is positioned for that rather than smaller model inference and second.

And a little bit of time talking about the evolution of the inference workload as we've moved to L. O ends and how your company is positioned for that rather than smaller modeling friends and second up until a month or two ago I never really got any questions at all about the data processing piece of the AI workloads. So the pieces.

Speaker 13: Up until a month or two ago, I never really got any questions at all about the data processing piece of the AI workload. So the pieces of...

Speaker 13: Manipulating the data before training, between training and inference, after inference, and I think that's a large part.

Manipulating the data before training between training and inference after entrants and I think that's a large part of the workload now maybe you could talk about how cuda is enabling acceleration of those pieces of the workload.

Speaker 13: of the workload now, maybe you could talk about how CUDA is enabling acceleration of those pieces of the workload. Thanks.

Sure.

Hi inference.

Inference is complicated it's actually incredibly complicated if you if you.

Speaker 6: Inference is complicated. It's actually incredibly complicated. If you, if you,

We reached weakest quarter announced.

Speaker 6: We this quarter announced one of the most exciting new engines.

One one of one of them.

Most exciting new engines.

Speaker 6: optimizing compilers called TensorRT LLM. The reception has been incredible. You go to GitHub, it's been downloaded a ton, a whole lot of stars integrated into stacks and frameworks all over the world.

Optimizing compilers called tensor RT O&M the reception has been incredible.

You Gotta get hub, it's been downloaded a ton a whole lot of stars integrated into into our stacks and frameworks.

All over the world.

Speaker 6: almost instantaneously. And there are several reasons for that, obviously.

Almost instantaneously.

And.

There are several reasons for that obviously.

Speaker 6: We could create TensorRT LLM because CUDA is programmable.

We could create tensor RT online because crude is programmable.

Speaker 6: If CUDA and our GPUs were not so programmable, it would really be hard for us to improve software stacks at the pace that we do. TensorRT LLM on the same GPU without anybody touching anything improves the performance by a factor of two.

If cuda and our Gpus were not so programmable.

Would really be hard for us to improve software stacks at the pace that we do.

Tensor RT <unk> on the same GPU.

Without anybody touching anything improves improves the performance by a factor of two.

Speaker 6: And then, on top of that, of course, the pace of our innovation is so high, H200 increases it by another factor of two. And so, our inference performance, another way of saying inference cost, just reduced by a factor of four within about a year's time. And so, that's really, really

And then on top of that of course, the pace of our innovation is so high H 200 increases by another factor of two and so our our inference performance and.

Another way of saying difference cost just reduced by a factor of four within about a year's time.

And so that's really really hard to keep up with.

Speaker 6: Now, the reason why everybody likes our inference engine is because our install dates.

Now the reason why everybody.

Likes our inference engine is because our installed base.

Speaker 6: And we've been dedicated to our installed base for 20 years, 20 plus years.

We've been dedicated to our installed base for 20 years 20 plus years.

Speaker 6: We have an install base that is not only largest in every single cloud, it's in every, you know, available from every enterprise system maker. It's, you know, used by companies of just about every industry. And every, anytime you see a NVIDIA GPU, it runs our stack.

We have an installed base that is not only largest and every single cloud it's in every.

Available from every enterprise system maker.

Used by companies.

Companies are just about every industry.

Anytime you see a nvidia GPU it runs our stack, it's architecturally compatible something we've been dedicated to for a very long time, we're very disciplined about it.

Speaker 6: architecturally compatible. Something we've been dedicated to for a very long time. We're very disciplined about it. We make it, if you will, architecture compatibility is job one.

We make it our if you will architecture compatibility is job one and that.

Speaker 6: And that has conveyed to the world the certainty of our platform stability.

Is that is convey to the world.

The certainty of our platform stability.

Speaker 6: NVIDIA's Platform Stability Certainty.

N genius platform stability certainty.

Speaker 6: is the reason why everybody builds on us first and the reason why everybody optimizes on us first.

Is the reason why everybody builds on us first.

Reason why everybody Optimizes on us first.

Speaker 6: All of the engineering and all the work that you do, all the invention of technologies that you build on top of NVIDIA accrues to and benefits everybody that uses our GPUs, and we have such a large install base.

All of the engineering and all the work that you do all the invention of technologies that you build on top of Nvidia accrues to the benefits everybody that uses our gpus and we have such a large installed base large millions of millions of GPS and cloud 100 million GP Houston from peoples Pcs just.

Speaker 6: large millions and millions of GPUs in the cloud, 100 million GPUs from people's PCs, just about every workstation in the world, and they're all architecturally compatible. And so if you're an inference platform and you are deploying an inference application,

Every workstation in the world and they're all architecturally compatible and so if you're an inference platform and you are deploying and conference application.

Speaker 6: You are basically an application provider and as a software application provider, you're looking for a large install base.

You are basically an application provider.

As a software application provider and you are looking for large installed base.

Data processing.

Speaker 6: Before you could train a model, you have to curate the data. You have to dedupe the data. Maybe you have to augment the data with synthetic data. So you process the data, clean the data, align the data, normalize the data. All of that data is measured not in bytes and megabytes. It's measured in terabytes and petabytes.

Before you could train the model you have to carry the data you have to <unk>. The data maybe you have to.

<unk> the data with the synthetic data sorry.

So.

Process the data clean the data align the data normalize the data all of that data is measured not in <unk>.

That invites them megabytes, it's measured.

Terabytes, Petabytes and the amount of data processing that you do before data it's data engineering.

Speaker 6: And the amount of data processing that you do before, data engineering, before that you do training is quite significant. It could represent, you know, 30, 40, 50% of the amount of work that you ultimately do.

<unk>.

Before that you do training is quite significant.

It could represent you know, 30%, 40%, 50% of the amount of work that you ultimately do.

Speaker 6: in what you, in ultimately creating a data-driven machine learning service. And so data processing is just a massive part. We accelerate Spark. We accelerate Python. One of the coolest things that we just did is called cuDF Tandas. Without one line of

And.

And what you.

And ultimately, creating a data driven machine learning service and some data processing is just a massive port we accelerate spark we accelerate python.

One of the one of the coolest things that we just we just did.

Paul <unk>, Kansas without one line of code and US which is the single most successful data science framework in the world.

Speaker 6: Pandas, which is the single most successful data science framework in the world, Pandas now is accelerated by NVIDIA CUDA, and just out of the box, without a line of code, and so the acceleration is really quite terrific, and people are just incredibly excited about it. Pandas was designed for one purpose, and one purpose only, really, data processing for data science. And so NVIDIA CUDA gives you all of that.

Candace now is accelerated by Nvidia Cuda, and just just out of the box.

Down a line of code.

And so the acceleration is really quite terrific and people are just incredibly excited about independents was designed for one purpose one.

<unk> only really data processing for data science.

So Nvidia Cuda gives you all of them.

Speaker 1: Your final question comes from the line of Harlan Sir of JP Morgan. Your line is open.

Your final question comes from the line of Harlan sur of Jpmorgan. Your line is open.

Speaker 14: Good afternoon, thanks for taking my question. If you look at the history of the tech industry, right, those companies that have been successful has have always been focused on ecosystem, silicon hardware, software, strong partnerships. And this is importantly right in aggressive cadence of new products, more segmentation over time.

Good afternoon. Thanks for taking my question. If you look at the history of the tech industry rate. Those companies that have been successful has have always been focused on ecosystem silicon hardware software our strong partnerships and just as importantly, write an aggressive cadence of new products more segmentation over time.

Speaker 14: You know, the team recently announced a more aggressive new product cadence and data center from two years to now every year with higher levels of segmentation, training, optimization, inferencing, CPU, GPU, DPU networking. How do we think about your R&D, OpEx growth outlook to support a more aggressive and expanding forward roadmap? But more importantly, what is the team doing to manage and drive execution through all of this complexity?

The team recently announced a more aggressive new product cadence and data center from two years to now every year with higher levels of segmentation training optimization in Quincy and CPU GPU and CPU networking, how do we think about your R&D opex growth outlook to support a more aggressive in expanding core.

Roadmap, but more importantly, what is the team doing to managing drive execution through all of this complexity.

Speaker 6: Gosh, you know, boy, that's just really excellent. You just wrote NVIDIA's business plan, and you described our strategy. First of all, there is a fundamental reason why we accelerate our execution, and the reason for that is because it fundamentally drives down costs.

Gosh.

Boy, that's just really excellent you just wrote in video business plan.

And and so you described our strategy.

First of all there is a fundamental reason why we accelerate our execution and the reason for that is because it's fundamentally drives down cost.

Speaker 6: When the combination of TensorRT LLM and H200 reduced the cost for our customers for large model inference by a factor of four,

When the combination of tensor RT <unk> and <unk> hundred reduce the cost for our customers for large modern influenced by a factor of four.

And so that includes of course, our speeds and feeds but mostly its because of our software mostly the software benefits because of because of the architecture.

Speaker 6: And so that includes, of course, our speeds and fees, but mostly it's because of our software, mostly the software benefits because of the architecture.

Speaker 6: And so we want to accelerate our roadmap for that reason. The second reason is to expand the reach of generative AI. The world's number of data center

So we want to we want to accelerate our roadmap for that reason. The second reason is to expand the reach of generative AI the worlds number of <unk>.

Data center configurations.

Speaker 6: This is kind of the amazing thing, you know, NVIDIA is in every cloud, but not one cloud is the same.

This is kind of the amazing thing you know Nvidia is in every cloud, but not one cloud is the same.

Speaker 6: NVIDIA is working with every single cloud service provider and not one of their networking

Video is working with every single cloud service provider and not one of their networking.

Speaker 6: control plane, security posture is the same. Everybody's platform is different, and yet we're integrated into all of their stacks, all of their data centers, and we work incredibly well with all of them.

Control plane security posture is the same everybody's platform is different and yet we're integrated into all of their stacks all of their data centers.

And we work incredibly well with all of them and not to mentioned, we then take the whole thing and we create AI factories that are stand alone. We take our platform. We can put them into supercomputers now we can put them into enterprise, bringing AI to enterprises, something generative AIG enterprise something nobody has ever done before.

Speaker 6: And not to mention, we then take the whole thing and we create AI factories that are standalone. We take our platform, we can put them into supercomputers, we can put them into enterprises.

Speaker 6: Bringing AI to enterprise is something, generative AI to enterprise, something nobody's ever done before.

Speaker 6: And we're right now in the process of going to market with all of that.

And we're right now in the process of going to market with all of that and so the complexity.

Speaker 6: And so the complexity includes, of course, all of the technologies and segments and the pace. It includes the fact that we are architecturally compatible across every single one of those. It includes all of the domain-specific libraries that we create. The reason why every computer company, without thinking, can integrate NVIDIA into their roadmap and take it to market. And the reason for that is because there's

Includes of course, all of the technologies in segments and the pace and includes the fact that we are architecturally compatible.

Across every single one of those it includes all of the domain specific libraries that we create.

The reason why you every every computer company without thinking integrate nvidia into their roadmap and taken to market and the reason for that is because theres market demand for it.

Speaker 6: There's market demand in healthcare. There's market demand in manufacturing. There's market demand, of course, in AI, in financial services, in supercomputing, in quantum computing. The list of markets and segments that we have domain-specific libraries is incredibly broad. And then finally, we have an entity.

There is market demand and health care, there is market demand and manufacturing there is market demand and of course in AI in financial services in supercomputing and quantum computing the list.

The list of markets and segments that we have domain specific libraries.

Is incredibly broad and and then and then finally.

Now we have an end to end solution for data centers.

Speaker 6: InfiniBand networking, Ethernet networking, x86, ARM, just about every permutation, combination of solutions, technology solutions, and software stacks provided. And that...

Infiniband network Infiniband networking Ethernet networking X 86 arm just about every permutation.

Combination of solutions.

Technology solutions and software stacks provided.

And that that translates.

Speaker 6: to having the largest number of ecosystem software developers, the largest ecosystem of system makers.

Having the largest number of ecosystem software developers the largest ecosystem of system makers.

Speaker 6: the largest and broadest distribution partnership network, and ultimately the greatest reach. And that takes, surely that takes a lot of energy, but the thing that really holds it together, and this is a great decision that we made decades ago, which is everything is architecturally compatible.

The largest and broadest distribution partnership.

Network and ultimately the greatest reach and that takes that takes surely that takes a lot of energy.

But the.

The thing that really holds it together and this is this is.

A great decision that we made decades ago, which is everything is architecturally compatible when you when we develop a domain specific language that runs on one GPU. It runs on every GPU.

Speaker 6: When we develop a domain-specific language that runs on one GPU, it runs on every GPU. When we optimize...

When we optimized tensor RT.

Speaker 6: for the cloud, we optimized it for enterprise. When we do something that brings in a new feature, a new library, a new feature, or a new developer, they instantly get the benefit of all of our reach.

For the cloud we optimized it for enterprise when we do something that brings in a new feature a new library and new feature or new developer they instantly get the benefit of all of our reach and so that discipline that architecture compatible discipline that has lasted.

Speaker 6: And so that discipline, that architecture compatible discipline that has lasted, you know, more than a couple of decades now is one of the reasons why NVIDIA is still really, really efficient. I mean, we're 28,000 people large and serving just about every single company, every single industry, every single market around the world.

More than a couple of decades now is one of the reasons why <unk> is still really really efficient I mean, we're 28000 people large and serving just about every single company every single industry.

Every single market around the world.

Thank you I will now turn the call back over to Jensen Huang for closing remarks.

Speaker 1: Thank you. I will now turn the call back over to Jensen Huang for closing remarks.

Speaker 6: Our strong growth reflects the broad industry platform transition from general purpose to accelerated computing and generative AI. Large language model startups consume internet companies and global cloud service providers are the first mover.

Our strong growth reflects the broad industry platform transition from general purpose to accelerated computing and generative AI <unk>.

Large language models startups consume internet companies and global cloud service providers are the first movers.

Speaker 6: The next waves are starting to build. Nations.

The next waves are starting to build.

Nations.

Speaker 6: and regional CSPs are building AI clouds to serve local demand.

And regional Csp's are building AI cloud to serve local demand.

Speaker 6: enterprise software companies like Adobe and Dropbox, SAP and ServiceNow are adding AI co-pilots and assistants to their platforms.

Enterprise software companies like Adobe and Dropbox.

And service now are adding AI co pilots and assistance to their platforms.

Speaker 6: And enterprises in the world's largest industries are creating custom AIs to automate and boost productivity.

And enterprises in the world's largest industries are creating custom AI to automate and boost productivity.

So Joseph AI era is in full steam and has created the need for a new type of data center and.

Speaker 15: The generative AI era is in full steam and has created the need for a new type of data center, an AI factory.

And AI factory.

Optimized for refining data.

Speaker 6: and training, and inference, and generating AI.

And training and inference and generate generating AI.

Speaker 6: AI factory workloads are different and incremental to legacy data center workloads supporting IT tasks. AI factories run co-pilots and AI assistants, which are significant software TAM expansions.

AI factory workloads are different and incremental two legacy data center workloads supporting tasks.

<unk> run co pilots and AI assistance, which are significant software Tam expansion and.

Speaker 15: and are driving significant new investment, expanding the $1 trillion traditional data center infrastructure install base, and empowering the AI industrial revolution.

And are driving significant new investment.

Expanding the one trillion traditional data center infrastructure installed base empowering the AI industrial Revolution.

Speaker 6: NVIDIA H100 HGX with InfiniBand and the NVIDIA AI software stack define an AI factory today.

And with each 100, HTS with Infiniband and the Nvidia AI software stack defined and AI factory today.

Speaker 6: As we expand our supply chain to meet the world's demand, we are also building new growth drivers for the next wave of AI. We highlighted three elements to our new growth strategy that are hitting their stride, CPU, networking, and software and services.

As we expand our supply chain to meet the world's demand. We are also building new growth drivers for the next wave of AI.

We highlighted three elements to our new growth strategy that are hitting their stride.

<unk> networking and software and services growth.

Speaker 6: Grace is NVIDIA's first data center CPU.

<unk> is NVIDIA's first datacenter CPU.

Speaker 6: Grace and Grace Hopper are in full production and ramping into a new multi-billion dollar product line.

Grace and Grace Hopper are in full production and ramping into a new multibillion dollar product line next year.

Speaker 6: Irrespective of the CPU choice, we can help customers build an AI factory.

Irrespective of the CPU choice, we can help customers build an AI factory.

Speaker 15: NVIDIA networking now exceeds a $10 billion annualized revenue run rate. InfiniBand grew fivefold year over year and is positioned for excellent growth ahead as the networking of AI factor.

And video networking now exceeds a $10 billion annualized revenue run rate Infiniband grew.

<unk> grew five fold year over year and is positioned for excellent growth ahead as the networking of AI factories.

Speaker 6: Enterprises are also racing to adopt AI, and Ethernet is the standard network.

Enterprises are also racing to adopt AI and Ethernet as the standard networking.

Speaker 6: This week, we announced an Ethernet for AI platform for enterprise.

This week, we announced an Ethernet for AI platform for enterprises.

Speaker 6: NVIDIA Spectrum X is an end-to-end solution of Bluefield SuperNIC.

Video spectrum ex is an end to end solution of Bluefield Super neck.

Spectrum for Ethernet switch.

Speaker 6: and software that boosts Ethernet performance by up to 1.6x for AI workloads.

And software that boosts Ethernet performance by up to one six X for AI workloads.

Dell HP and Lenovo have joined us to bring a full generative AI solution of Nvidia AI computing networking and software to the world's enterprises.

Speaker 6: Dell, HPE, and Lenovo have joined us to bring a full generative AI solution of NVIDIA AI computing, networking, and software to the world's enterprises.

Speaker 15: NVIDIA software and services is on track to exit the year at an annualized run rate of $1 billion. Enterprise software platforms like ServiceNow and SAP need to build and operate proprietary AI. Enterprises need to build and deploy custom AI copilots.

Nvidia software and services is on track to exit the year at an annualized run rate of $1 billion enterprise.

Enterprise software platforms like service known as SAP need to build and operate proprietary AI.

Enterprises need to build and deploy custom AI co pilots.

Speaker 6: We have the AI technology, expertise, and scale to help customers build custom models.

The AI technology.

Expertise and scale to help customers build custom models.

With their proprietary data on Nvidia <unk> cloud and deploy the AI applications on enterprise grade Nvidia AI enterprise and.

Speaker 15: with their proprietary data on NVIDIA DGX Cloud and deploy the AI applications on enterprise-grade NVIDIA AI Enterprise. NVIDIA is essentially.

Nvidia is essentially an AI foundry.

Nvidia Gpus Cpus networking AI foundry services, and Nvidia AI enterprise software are all growth engines in full throttle.

Speaker 6: NVIDIA's GPUs, CPUs, networking, AI foundry services, and NVIDIA AI enterprise software are all growth engines in full throttle.

Speaker 15: Thanks for joining us today. We look forward to updating you on our progress next quarter.

Thanks for joining us today, we look forward to updating you on our progress next quarter.

This concludes today's conference call you may now disconnect.

Okay.

Okay.

Yeah.

Okay.