LIVE | Nvidia CEO Jensen Huang Unveils AI Vision At Computex 2026 | APT

Added: 2026-06-03

[00:00:00]coming up with a plan that you can act that it acts on. That orchestration path is orchestrated by some software. And so this is fundamentally a agent. It deals with short-term memory called working memory, long-term memory just like we do. We have long-term memory. And so the memory management system is incredibly important. This entire system is called an agent. The large language model is used to do the thinking and the harness connects everything together just like an operating system. Okay. And so this is the new computing model and this is what an agent it could do incredible things. This is the big breakthrough.

[00:00:47]The simultaneous conver the convergence of large language models that are now able to do a really good job thinking, reasoning, planning, using tools and the fact that we have now these harnesses that manages memory. The orchestration uses tools we can now do amazing things.

[00:01:06]Let me give you some example. This is this is a prompt. This is the prompt.

[00:01:12]This is the code that is generated and this comes out.

[00:01:18]This is the input.

[00:01:21]This is the input and that's the output.

[00:01:26]Do you guys what do you guys think? It's pretty amazing, right?

[00:01:31][applause] We use cloud code here, but codeex does an incredible job as well. Here's another example. This is the input.

[00:01:39]Create a GIF. Nvidia gen green dots on black scatter form Taiwan 101 building morph to GTC Taipei 20 2026 morph to Nvidia I logo then scatter and repeat right so you saw that that was the prompt here's the next one I lost my remote control battery clip it looks like this create a CAD file it uses a tool create a CAD file ready for 3D printing to create a new new one.

[00:02:11]Make sense? This is now the new computing pattern. Whereas we used to launch an application, click and type, we now replace that with explain to the AI what we want, our intent, and the AI generates the code or uses tools and produce the necessary output.

[00:02:36]This is how computers are going to work in the future. This is agentic AI. For two years, we've been building towards this and now it has arrived. Now, one of the big breakthroughs, of course, is tool use. A lot of people have said, you know, Jensen, AI is coming. Agentic AI is coming. Therefore, all of the software companies are going to go out of business. I said, it's exactly the opposite because there are going to be so many agents.

[00:03:04]The world is no longer limited by the number of people. Therefore, those agents are going to use more tools than ever.

[00:03:14]This is actually an incredible time to be a software company. But the software has to be presented to the agent in a way that the agent can use it. This is a break big breakthrough. And in fact, what we have done as you know what Nvidia's treasure is is all of our CUDA libraries. I call them CUDA X libraries.

[00:03:36]This is Nvidia's treasure. Today, we're able to now pres present these CUDA X libraries to agents who can use it much more effectively than even h humans. And so this is a wonderful time for CUDA X libraries. Let's take a look.

[00:03:56]20 years ago, we built CUDA, [music] a single architecture for accelerated computing. We reinvented computing. A thousand CUDA X libraries help developers make [music] breakthroughs in every field of science and engineering.

[00:04:12]CUDA X libraries [music] are tools for agents.

[00:04:16]CU Litho for computational lithography.

[00:04:20]CU Opt for decision optimization.

[00:04:25]CDSS for direct [music] sparse solvers.

[00:04:29]AIQ for deep research across structured and unstructured documents.

[00:04:35]Aerial [music] for AI ran, warp for differentiable physics, parabrics [music] for genomics.

[00:04:46]At their foundation are algorithms and they are beautiful.

[00:05:05]>> [music] [music] [music] [music] >> Heat. Heat.

[00:05:56]Heat. Heat.

[00:06:21]>> [music] [music] >> Heat up here.

[00:06:41][music] >> [music] >> Heat up here.

[00:07:02][music] [music] >> [music] >> Heat.

[00:07:18][music] Heat.

[00:07:26][music] >> [music] [applause] >> A round of applause for math. Math is beautiful.

[00:07:54]The computing pattern, the computing pattern of software is going to change.

[00:07:58]In fact, let's come back to this. This is the agent. It is the ultimate disagregated and distributed computing computing model.

[00:08:11]So many different computers are going to be activated in order to process this agent. The agent consists of model harness, tools and skills, and a runtime.

[00:08:29]All of that is running at different places in a data center.

[00:08:33]You can think of the model as the brain, the harness as the body, the tools that it uses working in a runtime. Think of it as a workshop.

[00:08:48]So this is a person, a worker working with tools in a workshop. Of course, this is being done at extraordinarily large scales and each one of those steps are running in a different part of the computer. And you could see the large language model is thinking, context processing, observing, understanding the environment, reasoning, coming up with a plan, and acting on the plan. Every single time that happens, an entire rack of Grace Blackwell MVL link 72 is activated. It's thinking with a large language model.

[00:09:27]Whenever it uses a tool, a CPU use is used. That tool could be a C compiler, it could be Python, it could be JavaScript or it could be accelerated computing. Today's agents are rel relatively simple users of tools.

[00:09:45]Tomorrow they're going to be very sophisticated users of tools, which is the reason why the CUDA X libraries that I showed you are going to be incredibly popular with agents. They solve some of the most important problems the world knows. And all of our CUDA X libraries are now now going to come with skills that the AI could learn how to use. So the CUDA X library some skills basically a manual the AI reads it and go aha that's how you use it.

[00:10:18]The ability to use these libraries by agents are going to be incredible. And so the tools run on CPUs and GPUs and large language models. The security harness runs on CPUs and a security processor called a DPU. Nvidia's blue field. The orchestration of all this runs on a CPU. This is the entire harness and the CPU is orchestrating all of the work. One of the hardest parts is memory. You could just imagine the working memory is called KV caching.

[00:10:52]What to remember? Compaction, not just compression, but how to retrieve. Do you retrieve structured data? Do you retrieve unstructured data? What is the ontology? The relationship of all of these different data to itself. That entire processing is incredibly complicated. The memory system, the memory system of AIS is going to cause the storage system to be completely revolutionized.

[00:11:20]As you could see, every aspect of this computing model, this computing pattern, this new application called an agent is fundamentally different than the way that applications used to run. A whole bunch of software sitting inside a binary, sitting inside an operating system. This is the reason this disagregated, this distributed, this heterogeneous computing problem is precisely the reason we built our next generation.

[00:11:53]Vera Rubin, Vera Rubin is not one chip.

[00:11:58]Vera Rubin is not a GPU only. It starts with the GPU. But Vera Rubin is incredible.

[00:12:07]This entire thing is Vera Rubin from end to end. It has GPUs Vera Rubin MVLink 72. It is orchestrated by Vera CPUs that I'm going to tell you more about the storage systems revolutionary Vera along with CX9 our software stack called DOA the security processor that's inside so that everything is encrypted at rest.

[00:12:38]in motion as well as in use.

[00:12:42]Everything across this is secure because the AI model is so precious. This is the reason why this entire system obeys confidential computing.

[00:12:53]Each one of these systems would be a complete revolution in itself. Veryl Rubin is the most ambitious endeavor in the history of our company.

[00:13:04]The whole company worked on Vera Rubin across all 40,000 engineers. Not to mention all of you. All of you participated in the creation of this entire system. Vera Rubin is really a miracle. And it's not just one chip. It is so many.

[00:13:22]Well, it's even beyond that. A long time ago, Nvidia used to be a GPU company, but over the years, we've evolved to become a systems company. You're looking here now for the most complex system. It's most complex and groundup system ever designed.

[00:13:41]But ultimately, our customers, our partners don't want to buy computers.

[00:13:48]They want to build AI factories. which is the reason why Nvidia has really started to transform oursel yet again.

[00:13:56]You could see so much of our technology is now at the entire infrastructure scale. Our partners are at infrastructure scale. power generators, cooling systems, the grid providers, so many industrial companies are now part of our ecosystem because ultimately we're trying to build an entire stack just like GPUs, just like when we were building Grace Blackwell MVLink 72, just like now we are building a full stack system so that our customers could build amazing AI infrastructure. Let's take a look.

[00:14:37]>> [music] >> The world is racing to build AI factories. The largest infrastructure buildout in human history. AI factories are incredibly complex. Every layer, chip, rack, network, power, cooling, grid must be designed together [music] from end to end because compute is revenues.

[00:14:59]>> [music] >> NVIDIA DSX is the blueprint, a reference design for building and operating AI factories at maximum efficiency and profitability.

[00:15:09]It starts with [music] DSX SIM. With the DSX SIM Omniverse blueprint, partners design and validate an NVIDIA Vera Rubin AI factory before a single rack [music] lands.

[00:15:20]They plan the layout, simulate the power and cooling, design [music] the network, validate every integration, test every change in the digital twin.

[00:15:36]The factory powers on. DSXOSS [music] takes over and provisions, operates, monitors, and remediates the infrastructure, turning the installed systems into trusted, multi-tenant, resilient, AI ready capacity.

[00:15:54]Today's [music] AI factories overprovision power by up to 40%. DSXM Max LPS [music] lets operators safely deploy more GPUs inside the same power budget, adding billions in annual revenue.

[00:16:10]Breakthrough [music] hot liquid cooling at 45° C uses less water and energy.

[00:16:16]More power going [music] to revenue generating compute. Incredible.

[00:16:22]Dynamic power allocation steers power from rack to rack, recovering stranded [music] watts, sending them where work is happening.

[00:16:30]In rack [music] power smoothing flattens peak current spikes and power surges.

[00:16:36]Throughout the factory, [music] teams of AI agents work with DSX Max LPS, continuously [music] coordinating to balance cooling and power to meet workload demand.

[00:16:49]DSX AI factories are flexible energy assets that operate cooperatively [music] with the grid. DSX Flex reads realtime grid signals and dynamically adjusts factory power [music] when the grid needs relief.

[00:17:05]100 gawatt of AI [music] factories will come online before the end of the decade. NVIDIA DSX AI factories run at [music] highest efficiency, produce the lowest cost tokens, and make the grid stronger.

[00:17:23]>> [applause] >> I've shown you ecosystem slides of the past where Nvidia's computing layers and software and software and computing stacks are integrated into other people's platforms, third party platforms and libraries that serves end markets. That was a computing ecosystem.

[00:17:46]This is an AI factory ecosystem. This is way downstream of all of you. Upstream of me is all of you and downstream of us is this ecosystem. Because Nvidia ultimately is not just building a GPU, not just building a system. We're helping customers build these AI factories, these AI infrastructure that is so immensely complex. Each one of these at one gigawatt level started at 30 2030 billion dollars. It is at 5060 billion dollars and soon it will be 80 hundred billion dollars per gigawatt $100 billion into an AI factory.

[00:18:32]It must work the first time and it must work right away. The cost of capital is incredible. The complexity is incredible. So as you see we used to design a chip inside a computer and then we simulated a system inside a computer. Today you saw just now everything was built in Omniverse.

[00:18:57]I've been working with Omniverse with all of you for a long time. This was the dream come true so that we can build these gigantic systems as large as the world wants to build inside a digital framework inside a digital simulator in a digital world long before we build the first break ground and put our money to work. So this is our ecosystem our we call it DSX.

[00:19:24]RTX is for our GPU, DGX for our systems, and now DSX basically infrastructure.

[00:19:31]Because of the work that we do here across this entire stack, including our systems and software, it's the reason why we could work with small companies and enable them to be worldclass AI clouds. Every one of these I'm about to show you are small companies just recently. And now Coreweave is worth 50 60 70 billion and growing incredibly fast. Recently we worked with Nbius and again they're growing incredibly fast.

[00:20:00]Each one of these clouds have incredible customers. Cursor the software coding company, Black Mountain Labs, Image Generation, World Labs, World Foundation model, Revolute, the leading uh financial services AI company, and Shopify. Here's another one. This is Nscale and their customers are British Telecom, Google. Google is using one of our AI clouds, Thinking Machines, a Frontier Labs company. Super exciting.

[00:20:32]Here's Neighbor Cloud in Korea, Bank of Korea, Hyundai.

[00:20:38]So many incredible companies. Here's one in in India, Yoda.

[00:20:43]Incredible companies. Here's one uh based in Singapore building in a Australia together AI AI Singapore. This is one in Indonesia. Each one of these companies each one of these companies are serving regional as well as global customers. AI is going to run everywhere. Every company will be powered by it. Every region will build it. Endosat here in in Indonesia. Here in Taiwan, GMI here in Taiwan. GMI. It's okay to clap.

[00:21:23][applause] So, incredible, incredible uh incredible companies, incredible opportunity, but all of them need several things. Of course, they need the computing stack.

[00:21:37]This entire stack underneath this is what made Nvidia famous. All of our hardware and software and libraries, our connection into the world's ecosystem of third party developers makes it possible for anyone to stand up an AI cloud.

[00:21:54]However, the AI cloud is so complex.

[00:21:56]Now, this is the software version. This is the computer science version.

[00:22:02]The money version, the asset version is what I showed you earlier. It's a giant factory. Having this ability alone is not enough which is the reason why Nvidia has become an AI infrastructure company. Now doing this well and becoming incredibly good at dep at helping customers build AI factories and deploying AI factories is incredibly important. And the reason for that is this. Compute is revenue. Now compute is profit. the absence of revenues and profit is loss. And so it's really important to realize that this is when this is an example of an AI infrastructure coming online. It could take it could be coming online quickly. It could take a while. Its throughput could be high. It could be low. Its resilience and reliability could be good or bad. And its lifetime of usefulness could be long or short because this represents 50 60 going to a hundred billion dollars.

[00:23:17]This curve matters greatly which is the reason why Nvidia is such a great partner working with us because of our fully integrated capability. We didn't just come up with a PowerPoint slide. We created the entire infrastructure. We connected everything together. We built out billions and billions of it ourselves to make sure that everything works well. As a result of that, our time our time to first token, our time to first token, our time to first inference, our time to training turned on is much faster. Second, because our throughput per watt, our tokens per watt is utterly world class. And the reason for that is because we integrate everything. We design everything from the ground up. We simulate the entire system and we use extreme code design just like I showed you just now with the Vera Rubin rack. Everything was designed in order to deliver on this incredible throughput.

[00:24:25]If your data center, if your factory has one gigawatt, it will not have more. One gawatt means one gigawatt. That's all the power generation you could do. If you have one gawatt of power, then throughput per watt is revenues because every token is profitable. Every token is revenues.

[00:24:53]This is the future.

[00:24:55]Compute is revenues. Performance per watt is your revenues. Choosing the wrong architecture.

[00:25:02]Just because the chips are cheaper doesn't translate, doesn't make sense.

[00:25:09]You need to make sure that your revenues per watt, the more you buy, the more you make.

[00:25:16]And so tokens per wide.

[00:25:19]And then lastly, very last oh second, third is reliability.

[00:25:25]If you ever get a chance to see these data centers, there are so many moving parts, millions of cables, the ability for all of those computers to work harmoniously, reliably is extremely low. It is just extremely difficult. We have now been operating very large scale for a very long time. That experience matters. That difference meanantime between interrupts extremely important. And then lastly, this is very hard.

[00:25:58]The lifetime of these systems, the lifetime of these systems, the software is changing all the time. Four years ago, which is in the time of Hopper, AI has completely changed.

[00:26:12]Six years ago, this is the time frame of Ampear, AI has completely changed.

[00:26:18]We started out talking about CNN's, here we are. Then we talked about transformers and then we talked about mixture of experts. Now we're talking about agentic systems.

[00:26:30]Every single generation, every single few months, the software industry is coming up with new technology. If your architecture is not flexible, if your ecosystem is not rich, then this curve cannot be long. You cannot predict how long your system can last. I can. Nvidia systems is all over the world. Software developers start with Nvidia CUDA and by definition therefore the life, the ecosystem, the useful asset is going to be much longer. The difference is essentially cost. You could think of it as revenues, but the other side of revenues is cost.

[00:27:16]If the life of the asset is long, the TCO is low.

[00:27:21]This is the difference. This is what it looks like when compute.

[00:27:30]The more you buy, the more you make.

[00:27:37][applause] Now, all of you are experiencing this with me. Isn't that right?

[00:27:46]>> All of your demand, your factories are working so hard, your people are working so hard all across Taiwan because everybody wants to make money. They realize that AI, useful AI is here.

[00:28:03]Profitable AI is here. Compute demand is incredibly high and compute demand is the constraint. And so let's go work super super hard and help the world stand up AI factories everywhere. This is why it's so important. I'm so happy here I am standing in front of you.

[00:28:25]Vera Rubin is in full production.

[00:28:30][applause] Vera Rubin is in full production.

[00:28:37]it. The um the supply chain we created for Vera Rubin is twice as large as Grace Blackwell.

[00:28:46]Not Yeah, it's incredible. And And what used to take two hours to assemble one Grace Blackwell rack now only takes five minutes. So, not only is the capacity higher, the throughput is a lot faster and we need it all to support the demand.

[00:29:07]This ecosystem is extraordinary.

[00:29:10]Millions of square feet has been put online to support Grace Blackwell and preparing now, ramping up now, Vera Rubin. I want to thank all of you. Vera Rubin is now in full production. Thank you.

[00:29:24][applause] Let's take a look.

[00:29:31]Large language models generate answers.

[00:29:33][music] Now AI agents can do work. But processing agentic AI is a whole different kind of problem. Agents [music] observe, reason, plan, use tools. They manage massive context, juggling working memory and long-term memory. [music] They spin up sub aents, specialists on demand. NVIDIA Vera Rubin is a [music] multi-rackck podscale system built to process Agentic AI and is now in full production. The manufacturing [music] automation and orchestration across the supply chain a miracle to witness. Our journey [music] started when we launched the first AI supercomputer Nvidia DGX1.

[00:30:14]Over the next decade, we pushed every chip and system to the limit. from Pascal [music] and the first MVLink to Grace Blackwell the first rack scale AI supercomputer and now Vera Rubin the first multirack [music] pods scale supercomputer built for the agentic age it starts at TSMC the seven new chips [music] that make up Vera Rubin take shape through hundreds of processing steps 3 nanometer [music] process co-wr packaging HBM4 [music] memory from Micron SKH and Samsung the Vera Rubin [music] compute board 6 trillion transistors with over 18,000 components on one board. Vera Rubin NVL72 [music] does the thinking prompt and context understanding reasoning and planning.

[00:31:03]Next, a new modular compute tray streamlined with a new PCB midplane design, super chips, connect X9 [music] super nicks, and Bluefield 4 DPUs, all made in place [music] with no cables for resiliency at AI factory scale, 18 compute trays, nine hot swappable NVLink switch trays, new high efficiency manifolds, liquid cooled bus bars carrying over 5,000 amps, the equivalent of 20 electric cars at full acceleration. Together, 1.3 million components [music] formed this third generation MGX rack design.

[00:31:41]Congratulations to Microsoft for their operational Vera Rubin MVL72 engineering rack. Congratulations to Dell and Coreweave as well for [music] standing up their Vera Rubin MVL72 engineering rack. Then the Vera CPU rack. 256 CPUs in a single [music] liquid cooled rack.

[00:32:01]Orchestrating the models, shuffling memory, launching tools. At Foxcon [music] and Quant, Gro 3 LPX takes shape. 256 Gro 3 LUS [music] across 16 trays, 40 pabytes per second of SRAMM bandwidth for ultra low latency.

[00:32:20]While MVL72 [music] generates tokens at the highest throughput, Gro LPX generates them at the lowest latency.

[00:32:28]Vera Bluefield 4 STX, [music] where AI keeps its memory. Storage processing accelerated by Bluefield 4, connecting memory, storage, and insilicon security.

[00:32:41]and NVIDIA Spectrum X Ethernet [music] photonix. The world's first Ethernet switch with 200 gigabit [music] co-packaged optics. TSMC's coupe process chip scale packaging and ultra highowered laser [music] dice on indium phosphide.

[00:32:59]Vera Rubin, five connected [music] rack scale systems, a supercomput for AI agents, 150 supply chain partners [music] across Taiwan, millions of square feet of factory floor, hundreds of sites, chips, packages, systems, and [music] data centers pushed to the limits of size, power, and scale. This is what we call extreme code design. We did this with Taiwan. Together, we [music] reinvented computing for the age of AI. Taiwan was with us at the beginning and here today as [music] we bring Vera Rubin to the world. Thank you Taiwan. [music] [applause] Ladies and gentlemen, Vera Rubin.

[00:33:47]Vera Rubin was not just built for AI. AI. Vera Rubin was not built just to run AI. Vera Rubin was built to run agents. This is an agentic system. Imagine the complexity which is the reason why agents is the last computer science breakthrough. It has taken this many years for agents to realize its potential and become useful. It stands to reason that the computer that runs it is the most advanced in the world. This is Vera Rubin. Let's take a look. Can we bring out Vera Rubin, please?

[00:34:38]>> [applause] >> And Janine, do we have the do we have the racks, the systems?

[00:34:53]It looks heavy.

[00:34:56]This is This is Vera Rubin. Vera Rubin MVLink 72.

[00:35:02]This is the Gro LPX. At the next GTC, I'm going to talk to you about a lot more of this today. We have so much to talk to you about. This is Vera CPU rack. 256 CPUs, all liquid cooled. Let me tell you about Vera in just a moment.

[00:35:20]This is the Vera Bluefield storage processing system and also security system. And of course this is our Melanox networking the world's first CPO.

[00:35:34]This is Vera Rubin. Incredible technology all coming together. Now when we built when we built Hopper, we built Hopper as you know for pre-training.

[00:35:46]pre-training was the most important application, the most important workload we were working on at the time. Then when we worked on Grace Blackwell, everybody said, "Jensen, you know, Nvidia is really good at pre-training.

[00:35:59]Inference is so easy." Do you remember that? People used to say, "Inference is so easy. We could do that, too." But as you know, inference equals money. And the mo models are so complicated and to do it at incredibly high response time, fast interactivity, and high throughput at the same time is incredibly hard, which is the reason why we created MVLink 72. Today, Nvidia's token cost is the lowest in the world. Not by 10%, by X factors, orders of magnitude.

[00:36:37]All because we did extreme code design.

[00:36:40]All because we understood the computing model, the computing pattern of inference and we were able to create MVLink72. Now with Vera Rubin, it is beyond inference. It is now inference in an agent agentic system. This is Vera Rubin. No cables, no hoses, no fans.

[00:37:07]What used to take the last time when I showed this to you, we had cables everywhere.

[00:37:14]The cables were amazing to look at, but now there's a PCB in the middle which connects both sides. What used to take two hours now takes 5 minutes. The reliability and the resilience of Vera Rubin is going to be off the charts.

[00:37:31]This is our Vera CPU tray. the most advanced CPUs that has ever been built.

[00:37:38]I'm going to show you that in just a second. And this is our storage tray.

[00:37:45]Two Vera CPUs, four CX9. Incredible amounts of software.

[00:37:53]This is our new LPX LPU30, the Gro system designed for very low latency inference. The throughput is delivered by Vera Rubin and extended with MVLink 72. If you want to extend that even further, you can have Gro LPUs. Here we have the Vera Rubin MVLink the switch tray. This is the switches in the middle. And this is revolutionary because of Vera Rubin's because of MVLink72 and the MVLink switches that we created and invented. And this is our Ethernet switches for scale out. What's amazing is we introduced these two systems for Grace Blackwell. These two systems were created for Grace Blackwell and today Nvidia is the largest networking company in the world. I'm so proud of the networking team.

[00:38:56]This is such an incredible enabler for everything that we do. I'm going to now talk to you about the next major industry we're going to be part of.

[00:39:07]Thank you, Janine.

[00:39:10]Thank you.

[00:39:12][applause] I think there are 2,000 people back there pulling that.

[00:39:30]Okay, let's talk about CPUs.

[00:39:35]Vera CPUs. CPUs built for the age of AI.

[00:39:40]All of the CPUs until now were created for people. We were the users.

[00:39:48]We were the users. We were the renters.

[00:39:52]The way we use CPUs, we live in a world counted by seconds.

[00:39:57]The way we rent CPUs in the cloud, each one of them more you can more CPU cores you have, the more you can rent. the economics of the old the use case of the old CPU and the economics of the old CPU fundamentally different than agents.

[00:40:16]Agents are impatient. They don't live in a world that is in seconds. They live in a world that's in nanoseconds.

[00:40:25]When it uses a tool, it wants the response time to be as fast as possible.

[00:40:31]When it access database, it has to come back as soon as possible. Every moment that the agent is waiting keeps it from going to the next step, the next step, the next step. It is vital that we make the CPUs as low latency as possible, as interactive as possible. So we created Vera CPU for the age of AI. Now inside our system, it's used for three different ways. The first way of course is Vera Rubin for thinking and inside the Vera Rubin rack there already two CPUs as you know we are building and selling millions of Vera Rubins. We have sold millions of Grace black walls. Nvidia already is one of the largest CPU makers in the world.

[00:41:28]Vera in the Vera Rubin rack are two CPUs. One for orchestrating and managing the GPUs, managing the KV cache dealing with all of the software that runs in the rack. We also have the grace blue field that is used for security and isolation.

[00:41:50]The Vera compute is used for the harness, the orchestration of the AI models, tool use, accessing the database. And the data servers are right here, Vera Bluefield, the fastest storage, fastest storage servers, the fastest storage system the world has ever made. And the reason why this is so vital is because agents are accessing memory, accessing memory so incredibly fast. These systems, the storage server and the CPUs are now the critical path of the most expensive part of the data center. This is the most expensive for a good reason.

[00:42:36]The economics, the economics of the AI factory is tokens and the tokens are created here. And so of course you want to manufacture and generate as many tokens as possible.

[00:42:51]This is where you put all of your economics and this has to not be in the way. And so Vera CPU has great pressure on the Vera on the CPU architecture which is the reason why we built a brand new architecture from the ground up. A CPU the world has never seen before. We call it Vera.

[00:43:12]This is CPU for agents. All the CPUs of the past we built for humans. This CPU is built for agents. Well, there are four things to keep in mind. The four takeaways.

[00:43:27]The first takeaway is that the instructions per clock of Vera has to be incredibly good because we need the latency to be short. We need the processing time. Singlethreaded performance, not throughput.

[00:43:43]Singlethreaded performance has to be world class. Absolutely the best singlethreaded performance. Which is the reason why the IPC the instructions per clock of Vera is so high. is the highest in the world. 10 instructions fetched, decoded, and executed per clock. Number one. Number two, the bandwidth necessary to move data in and out for the CPU has to be utterly world class. The second thing is bandwidth per core. The third is just bandwidth period. We're moving. Remember I said earlier agentic systems is fundamentally disagregated and distributed. Disagregated and distributed. When computing is disagregated and distributed, networking becomes the problem. Therefore, we have to move the data around as fast as possible between the CPU cores and between the CPU and the storage, the CPU and the GPU.

[00:44:48]The bandwidth around the system and inside the CPU core has to be utterly worldclass.

[00:44:55]This is the first CPU that's been built a long time that is literally at retical limits with a fabric that connects all of the CPU cores that is speed of light 3.6 terabytes per second.

[00:45:10]No chiplet tax, no chip boundary crossings because we need to have everything because the CPU cores are talking to each other with extremely high bandwidth. They're not rented core per core per core. They're all working together.

[00:45:27]The cross-sectional bandwidth of Vera is off the charts. It's the first one to be PCI Express Gen 6. It is also the first one to have LPDDR DDR5 with 1.2 terabytes per second. Three times two to three times the bandwidth of the highest performance CPUs on the outside, three times the bandwidth on the inside. The bandwidth per core and the bandwidth period is world class. Now remember I showed you earlier.

[00:46:00]The number of CPU cores, the number of CPUs is going to be quite high and the reason for that is very simple.

[00:46:09]We created CPUs for humans in the past and humans there only 1 billion of us.

[00:46:18]There will be billions of agents and these agents are going to be using the CPUs with very little patience. because the cost of the GPU they sit next to is too high and therefore too valuable, too precious. Therefore, these CPUs are going to be both performant, but they also have to be extremely energy efficient so that we can cram as much CPU as we can into the factory without taking away power from the token generation, which we know is how we make money.

[00:46:56]These four properties, instructions per clock or single threaded performance, bandwidth per core, the total bandwidth around the chip and inside the chip and energy efficiency defines Vera. It is absolutely world class. When you compare it to the highest performance x86, it is just off the charts. When you compare it in real singlethreaded performance, real performance, it's off the charts.

[00:47:28]It is incredible to be able to deliver 5% improvement on CPUs. It is incredible to be able to deliver 10%. But this kind of performance speed up is just unheard of. This is Nvidia Vera.

[00:47:44]What do you think?

[00:47:46][applause] Let's take a look.

[00:47:54]>> Aentic AI [music] changes the role of the CPU. The CPU is now the conductor and the GPU is the orchestra.

[00:48:01]Traditional CPUs [music] were built for a different era.

[00:48:05]Maximizing cores per socket. Slice them up, virtualize, rent by the hour.

[00:48:11]In the age of agents, the CPU [music] is now a bottleneck to GPU utilization, directly affecting token throughput, latency, and user experience.

[00:48:22]NVIDIA Vera is the CPU built for the Agentic loop, combining NVIDIA's [music] custom data center CPU core with a scalable coherency fabric for the right balance [music] of performance cores and bandwidth to maximize AI factory output.

[00:48:36]At the heart of [music] Vera is the NVIDIA Olympus core built for modern data center workloads, [music] branchheavy Python runtimes, tool calls, and sandbox code execution. Each core is tuned for throughput. A neural branch predictor [music] evaluating two taken branches per cycle. A 10-we decode engine brings in more [music] work each cycle. A large out-of- orderer engine keeps instructions moving. Advanced prefetchers [music] with a novel graph engine anticipating the next data path.

[00:49:07]But fast cores only matter when data [music] arrives correctly and on time.

[00:49:13]Vera is the first CPU to use LPDDR5X memory while correcting multiple errors simultaneously [music] without compromising bandwidth. Vera achieves 40% lower peak memory latency versus x86, keeping cores fed on time through retrieval, [music] analytics, and sandbox execution. NVIDIA's second generation scalable coherency [music] fabric unifies all 88 Olympus cores on a monolithic mesh with separate dies [music] for memory and IO. cores are not split across chiplets, enabling 50% [music] faster core to core communication than traditional CPUs. And memory coherent NVLink chip to chip connects GPUs directly [music] to the fabric. Beyond GPUs, NVLink chip to chip can scale Vera up to multiple sockets, enabling massive bandwidth between CPUs.

[00:50:07]Vera delivers 1.8 >> [music] >> 8 times the agentic sandbox performance of x86 CPUs. Standalone Vera racks run agent sandboxes, tools, code, [music] and data pipelines. Tightly coupled to Reuben GPUs, Vera keeps accelerated workflows moving. Nvidia Vera Bluefield [music] 4 STX powers context memory and AI storage.

[00:50:32]Compute, networking, storage. Vera is the CPU for the age of agents.

[00:50:43][applause] This is going to be our new major growth driver. The reviews are already coming out and it's pretty good.

[00:50:55]That's pretty good stuff.

[00:51:01]>> [applause] >> Now remember, Grace and Vera are also the most highly qualified CPUs in the world of AI because every single data center, every single cloud, every single enterprise, every company that works with Nvidia on AI has already qualified.

[00:51:26]Grace the entire software stack has already been optimized for grace. Every company will be qualifying Vera.

[00:51:35]Vera will be the most optimized agentic CPU in the world simply because it's going to go with Vera Rubin simply because we made the the big hard switch.

[00:51:46]In fact, during Grace Blackwell transition, the biggest risk was going from external CPU x86 into Grace Blackwell.

[00:51:56]That transition was extremely dangerous, but we did it with incredible execution.

[00:52:02]Now, Grace is literally synonymous with Grace Blackwell. When people say Blackwell, they say Grace Blackwell because it is utterly now everywhere.

[00:52:12]Every company's software stack has been optimized for it. Everybody's security stack has been optimized for it and now here comes Vera. I'm super excited about that. Now look at some of the performance numbers.

[00:52:25]Speedups says one thing. It is extremely hard to speed up SQL.

[00:52:31]SQL the most famous domain specific language DSL that has ever been created before SQL. You know before CUDA there was SQL before OpenGL there was SQL invented by IBM today it is the structured database engine of the planet everybody uses SQL this is SQL running three times faster not 10% faster not 25% faster 10 times f three times faster incredible this is real time the next one is real time stream process processing. Remember, your AI is going to be not just reading documents. Your AI is going to be watching for telemetry, especially inside a factory, inside a stock exchange.

[00:53:26]You're going to be looking for telemetry continuously. The burst of data that's coming in goes into a CPU. This is Vera CPU running real time stream processing for New York Stock Exchange. Lynn Martin the president of New York Stock Exchange has been so gracious to partner with us.

[00:53:46]This system is run all over the world in real time real time stream processing vera CPU six times all because of the bandwidth the single single threaded instruction execution the bandwidth inside between the cores the bandwidth outside Vera is completely revolutionary that's vera [applause] you know X factors is something you you talked about when you're talking about GPUs. It is quite rare that somebody talks about X factors on real workload real workload that is associated with CPU. So I'm so proud of the team. You guys did such a great job. We have an extraordinary road map coming.

[00:54:34]But what's really exciting is almost everybody is supporting Vera. They're as excited as we are. This is Vera opening up. It's opened up a brand new market.

[00:54:48]Agents, agents is a new workload.

[00:54:52]We built CPUs for humans in the past. We need CPUs for agents. Agentic systems, their properties are different. Why would the old CPUs be the same? We are building millions and millions of errors. Millions of errors. and to go to market with us. Taiwan's ODMs and computer makers, all the OEMs, and you could see the early adopters.

[00:55:20]The early adopters are the agent companies. This is the beginning of a new market, a market that never existed before. It's not going to take away from the old markets, but this is a new market.

[00:55:35]CPU for agents. And this will this c this market will surely be larger than the last and the reason for that is because there'll be a lot more agents than there are people and then there the agents are very impatient. So Nvidia Vera CPU thank you.

[00:55:54][applause] This is the most important slide really.

[00:56:02]This is the takeaway. The takeaway here is that this is the application pattern.

[00:56:07]This is the computing pattern of the next decade.

[00:56:12]Agents, harnesses orchestrating large language models.

[00:56:19]Every company will run it. Every company will be an agent company. Every company will have agents running inside.

[00:56:28]Every company will see that agents will need its own operating system. Every company's asking us how do we run agents safely? How do we build agents for our own workloads? And so we have the NVIDIA agent toolkit for enterprise AI. You've seen me build this in plain sight.

[00:56:54]Almost everything that Nvidia does, as you know, at every GTC, if you go back and look at my GTC 5 years ago or 10 years ago, you will see today.

[00:57:02]This you've seen me talking about for several years now because we've been building for this moment. There are four things that companies need in order to build agents as a service or build agents to operate.

[00:57:18]The first thing you need is you need models. Of course, large language models, the smarter the better, the cheaper the better, the faster the better. The second is you need a harness to orchestrate the whole thing. The third, these a these models want to use tools and these tools come with its skills and I showed you CUDA X libraries. Those are going to be amazing tools for the agents in the future. And then lastly, you need a runtime. You need the operating system that holds it all together. This is the Nvidia toolkit for agents. It includes it includes models that you can modify. Nvidia's worldclass open models. And I'll show you more. You can run agents from anybody. You could run uh cloud code incredible agent codeex incredible agent. You could run it inside this harness called open shell which will be highly secure for your inside the enterprise. The shell protects the agent keeps it grounded in security policies.

[00:58:32]Privacy is protected. Its rights and privileges are given. Its identity is protected. And so this open shell is being adopted all over the world. Nvidia open shell is open source. You're going to see so many companies adopt it. Red Hat, Canonical, Microsoft, it's going to be adopted everywhere. This is an important this is the runtime and this runtime is fully optimized for the NVIDIA AI platform which is everywhere.

[00:59:03]So you can run open shell in any cloud on prem and even on device. So you have you have now tools and libraries that they can use. You have models that you can modify or use asis or you have agents. This be open claw Hermes another incredible another incredible uh harness. These agent agentic harnesses can now run on prem or for you anywhere.

[00:59:34]Okay. So four things and this represents the operating system of the modern enterprise. Now how do we use this? One of my favorite use cases of agents is chip designers. It is the single most important thing that Nvidia does. And so of course we have to partner with cadence to build super agent a chip design super agent. It is orchestrated by codecs or cloud code. It has RTL and architecture di diagrams or schematics or uh specifications as input and whatever you need to fix. And together we created some super agents that are optimized for the NVIDIA runtime with Neotron. And let's take a look. It's really incredible.

[01:00:32]Cadence and [music] Nvidia are partnering to build chip design agents.

[01:00:37]Hundreds of thousands of NVIDIA [music] chips come together to make the AI factories that power the world's frontier AI models. Designing these chips and the systems they run in is one of the hardest engineering challenges.

[01:00:50]Trillions of [music] transistors, three-dimensional circuits, microscopic scale. Every gate, every wire synchronized to picos seconds must work in perfect harmony with no margin for error. Physical prototypes [music] are too slow and too costly. So engineers work in the digital realm. Each chip begins as a [music] set of architectural specifications, then translated into RTL, the language of chip design. RTL must be verified in simulation. A single bug can delay a chip by months. At NVIDIA, thousands of engineers, billions of compute hours per year, millions of tests written, run, and debugged. A cycle that takes teams weeks. To compress this cycle, [music] Cadence and Nvidia built a design verification agent. Codex orchestrates the process.

[01:01:41]Cadence [music] chipstack launches the RTL verification loop powered by Neatron and secured by Nvidia OpenShell. Calling on expert sub agents in [music] RTL generation, testbench creation, regression testing and debug, the system drives itself. [music] The chipstack agents run hundreds of simulations with Cadence Excelium.

[01:02:02]Formal verification with Jasper. Design flaws revealed. Bugs in the code fixed.

[01:02:09]What once took weeks now takes [music] hours. Verification cycles over 40 times faster. Together, Nvidia and Cadence [music] are reinventing chip design with AI agents.

[01:02:24]From weeks, from weeks to hours, from weeks to hours, from weeks to hours, Nvidia has thousands of chip designers.

[01:02:33]We are going to hire hundreds of thousands of cadence super agents that work with us so that we can accelerate our company so that we can be even more ambitious create even more amazing things run even faster. You saw earlier that the toolkit with models harness tools the tools in this case are cadence simulators and verifiers formal verification systems. It is the reason why we're working with Cadence so hard to accelerate all of their tools on CUDA because the agents are impatient. The agents want the answer immediately. And so models, harnesses, accelerated CUDA accelerated libraries and tools and then the runtime. What you saw just now is all of that coming together. Now, one of the things that it starts with is a great model that Cadence could modify and tune to be expert at the cadence workflow at the cadence expertise so that they could create super agents that are proprietary to cadence with their proprietary knowledge.

[01:03:48]They have to start with an excellent model. We call it Neotron. Nvidia is dedicated to build open models for the world so that all of you, all of us could create our own agents. Today, we're announcing the Neotron 3 Ultra.

[01:04:08]Yep. Our next open model, and it is smart.

[01:04:13][applause] The Neotron models not only give you the model, we give you all the data that we use to train the model. And because we have a coalition of incredible partners, you can see all of our partners down here.

[01:04:31]We work together, contribute data to each other. Neotron is trained on one of the largest suites of longunning reasoning models, long running tool task solving tool using data sets in the world because of all of our great partnerships. All of this from the model, the training script and the data made completely available to you. This is open models at its best. the best open model system policies in the world.

[01:05:06]Simple goal is so that you can take all of it, add to it, make it even better, make it yours. Neotron 3 Ultra is five times faster.

[01:05:19]This is the world's first model based on a hybrid architecture of SSM state space models with mixture of experts. The architecture is incredibly fast. We made a fast so that you could think fast.

[01:05:34]When you think fast, you could think longer at the same cost. So five times faster. It is also 30% cheaper. 30% lower cost to run in total flops and total inference time than even the most cost effective in the world. We're comparing against the world's best open models. Frontier Smart, five times faster, 30 30% cheaper, completely open. We're completely dedicated to this. This is now Neotron 3. We're currently working on Neotron 4.

[01:06:11]So this entire toolkit from models, harnesses, tools and skills and runtimes is the reason why every enterprise company in the world has the ability now to create their own agents just like Cadence did with their super agents. And we're working with so many companies cadence and crowdstrike and dissol and palunteer sap and service now people were always said Jensen the agents are going to disrupt these markets. I said completely opposite and you can now see it. agents is going to create the largest opportunity ever for my partners and friends and we have the Nemo the the NVIDIA agentic toolkit for enterprise AI to help them.

[01:07:03]So there you go.

[01:07:06][applause] First, Vera Rubin in full production.

[01:07:13]Two, Vera CPU CPU built for a new generation for agents. And three, Nvidia's enterprise AI toolkits so that every enterprise and every enterprise software company can build agents.

[01:07:36]>> [applause] >> My relationship with you started here.

[01:07:44]And many of you, many of you, many of my friends and partners here in Taiwan, your companies started here.

[01:07:53]This is in a lot of ways the beginning of the modern computer industry. 40 years now. Nvidia is 33 years old. the PC industry was already starting to get the P Windows one and Windows 2 and Apple Mac Apple Apple one and Apple 2 and by the time that we came along Windows 3.1 was the PC and as you know Windows 95 made PC personal. It took PC from enterprises companies and made it into a consumer electronics device.

[01:08:29]Everybody should have one and everybody does. This is the beginning. This computing platform did several things incredibly smart.

[01:08:39]Windows was not just disagregated. As you know, Windows was properly abstracted. It was architected just right. systems biosis, open chipsets, the operating system with drivers, drivers that could be connected and installed at runtime and an abstraction layer with a multimedia API that was that opened up the PC to what we all know today. Each one of these elements were essential in making the PC so popular.

[01:09:18]40 years later, Microsoft and Nvidia are going to reinvent the PC.

[01:09:25]This is going to be the new PC. Now, tomorrow night, tomorrow night, I think it's tomorrow night our time, but I'm going to be with Satia where we're going to talk a lot more about the work that we're doing together. Microsoft Nvidia over the last three years. It took this long to completely reinvent how the PC is going to work so that we could be ready for this moment. As I mentioned earlier, that compute pattern called the agent. It's going to run in AI clouds.

[01:09:57]It's going to run inside enterprises. It is also going to run on your PC.

[01:10:04]What's going to happen to that PC when it has an autonomous agent? An agent that's helping you, that understands you. You could talk to it. It could look at you. You could ask it to read files, go help you do some research. It could do a lot more that I'll show you. But the new operating system is, of course, the old operating system plus large language models. large language models in a lot of ways is the modern version of DirectX.

[01:10:37]It has of course input and output, understands prompts, it understands computer vision, it can generate video, it can generate sounds. It is the modern extension, the intelligence extension of the PC, of a computer.

[01:10:53]On top of that, the application as I mentioned before is going to be replaced by now an agentic runtime and that is the modern application an agent. Let's now take a look at what it can do.

[01:11:09]It started with a spark, an idea to reimagine the PC for the first time in 40 years for the age of AI. What becomes of our [music] personal computer in a world of agents?

[01:11:25]Agents running natively, connected to models, local [music] or in the cloud.

[01:11:31]Our personal AI sandboxed for security, running continuously, getting work done.

[01:11:38]The chips and the OS must evolve.

[01:11:42]Introducing RTX Spark. Everything we've learned over 33 years distilled into one chip.

[01:11:51]Blackwell RTX [music] GPU with 6,144 CUDA cores. One pedlop of AI performance. A custom [music] 20 core grace CPU built in partnership with MediaTek. Fused by MVLink.

[01:12:09][music] 128 GB of unified memory. TSMC3 nanometer process, 70 billion [music] transistors, and in close collaboration with Microsoft, a Windows [music] platform for agents.

[01:12:26]We're reinventing the personal computer, for creating, for gaming, [music] for agents. This is the dawn of a new personal computing revolution and it starts with NVIDIA RTX [music] Smart.

[01:12:52][applause] Here it is.

[01:13:02]Of course, I got to show you the most beautiful part, which is video games.

[01:13:08]It is. It's also the closest to our heart. This is Forza. This is 007, by the way. The new 007 game. I'm looking forward to playing it. I look a little bit like him.

[01:13:21]Ladies and gentlemen, Nvidia's RTX Spark laptops. Now, [applause] thank you.

[01:13:37]I have too many things in my pocket.

[01:13:43]Okay. All right. This is the most amazing chip the world has ever built.

[01:13:48]This is the N1X that we built in partnership with MediaTek. I think I saw I saw Rick earlier. This is N1X. This is a beautiful chip. This is this is a a a chip that frankly would take 33 years to build. And the reason for that is because a 100% of NVIDIA software stack runs here. If you want to run uh uh digital biology, no problem. If you want to do seismic processing, no problem.

[01:14:17]You want astrophysics, no problem.

[01:14:19]Everything associated with CUDA, all the physics, all the biology, all the genomics, all the AI, no problem. All the computer graphics, no problem.

[01:14:28]Every single application Nvidia has ever created and every single application that Windows has ever run, Microsoft and Nvidia meticulously optimized everything so that this computer literally runs everything the world has ever created. Plus, it now runs agents. An incredible computer. I'm so proud of it.

[01:14:57]>> [applause] >> Okay.

[01:15:02]Now, I want you to keep that in mind in the next video. I just I'm going to show you. Just imagine everything here is going to run on your PC. Now that computer could have a local Neotron 3 Ultra model or Neotron 3 super model or it could have a cloud code or codeex or some other model in the cloud or something on the network and it's going to it's going to work and do something amazing. Let's play it.

[01:15:34]>> Every house starts as [music] an idea.

[01:15:36]Getting from idea to design takes a myriad of tools, expertise, and a lot of time.

[01:15:45]Now, [music] an agent running locally on RTX Spark can help me design a house using the tools on my laptop [music] with an open shell sandbox running the Hermes harness connected to Claude Sauna in the cloud. I select the site, share my concept sketches and mood [music] board of styles to inspire my design and the prompt, a text description of the requirements and the design [music] intent.

[01:16:13]My agent goes to work using the tools on my laptop. It opens Rhino and starts modeling the site, shaping terrain, setbacks, and [music] the building envelope. Then it proposes building forms optimized for cost, [music] comfort, and quality.

[01:16:31]With the form defined, my agent generates the interior layout. Walls, circulation, rooms begin to take shape.

[01:16:39]I jump in whenever I want [music] to adjust to change.

[01:16:46]Doors, [music] windows, and structural elements are placed automatically. My agent detects its own mistakes and fixes them.

[01:16:57]When I approve, the agent exports the model from Rhino into Blender. Materials and object properties transfer with [music] the design context intact. I fine-tune the materials, get the look just right. Then I pick the shots.

[01:17:13]Blender renders [music] the house. My agent using generative AI with the Flux 2 model makes them photoreal. Multiple viewpoints, lighting conditions. What was once a complex [music] workflow is now guided and simplified by my agent working with me on RTX Spark. Design at the [music] speed of imagination.

[01:17:40][applause] PC in the world of agents. The developers are so excited about it. This is an incredible computer. All of the acceleration, all the software capabilities associated with it, working with every developer to make it incredible for all of you. The next one, Adobe, incredible tool suite of course used by tens of millions of people around the world. They have re-engineered the architecture, the core of Adobe, Photoshop, and Premiere, and they're going to release it for RTX Spark. It is twice as fast. It's already fast. Now it's going to be twice as fast. And it it's also designed to be agent friendly with its MCP server. It can now interact with agents on your laptop.

[01:18:26]The number of customers, the number of partners that are so excited to bring RTX RTX Spark to the market is just incredible. You know, this is the first across the lineup of PC reinvention for 40 years and I'm just so happy that all of you and the ecosystem around the world has joined us. This is basically everybody everybody will support RTX Spark and will be building incredibly smart and powerful and beautiful laptops with all of us. Thank you very much.

[01:19:00][applause] But that's not all. That's not all.

[01:19:09]RTX Spark is a reinvention of laptop.

[01:19:14]But in fact, Microsoft Nvidia is reinventing all of PC. And today, we're announcing a whole new line.

[01:19:23]Three revolutionary Windows machines covering desktop, laptop, and workstations. All 100% Windows compatible, 100% CUDA, 100% Nvidia AI tensor core. Everything that runs that you see that runs on Nvidia in all these different platforms around the world runs here.

[01:19:46]This is the first completely re-engineered, reinvented line of PCs that has happened in 40 years. Now, what's really amazing is this. So, this is this is the RTX Spark laptop. This is the desktop. So, this one's from MSI Joseph. This one's yours. Okay. Look how beautiful it is. This agent could run 24/7 meter free and you could download your agent. You could raise your lobster dryer, your water cooler, your water heater, your everything, whatever you want, your security system, all connected to this. And this becomes your personal AI, your personal AI agent. And it gets smarter and smarter and smarter over time because today we have Neotron 3 Ultra. Tomorrow we have Neotron 4 and then Neotron 5, Neotron 6. And we just keep getting us smarter and smarter and smarter. And meanwhile, this is sitting at home helping you do things. If you want to book a travel, no problem. And if you if you want an incredible system, this is a DGX station for Windows, compatible with Windows, runs everything in Windows, and and it has 768 GB of memory. And so you could run a trillion parameter model. This is unbelievable. 20 pedaflops, eight terabytes per second of memory bandwidth, and this sits by your desk.

[01:21:43]You basically, if you're a developer of large language models, you're a developer of agents, having this sit by your desk gives you all the compute you need. And then when you deploy it, you put it into the cloud. Now, there's something that if you look at this and think about this, something is happening here. Remember 15 20 years ago we used to have an idea called a phone.

[01:22:11]Today we have an idea called a PC.

[01:22:16]Today when you think about your phone the one thing you don't do with it is make phone calls.

[01:22:25]You do just about everything else. And so that phone means something very different to you than a phone of the past.

[01:22:34]I am certain what's going to happen here is that the PC 10 years from now and the PC that you think about today, a tool whether you launch applications, click and type and this PC is going to be completely different. Here's my theory.

[01:22:55]I can totally imagine just as every house today has a home theater where many houses have home theaters, big TVs, lawnmowers, dishwashers.

[01:23:08]I could totally imagine that someday there's actually an AI supercomput in your house and it's running all of your agents. that's running all of your assistants and they're doing all kinds of things for you all the time and you have to have it in your house just like you have a home theater in your house, you have stereoss in your house, you have game consoles in your house, you want to assist AI agent computers running in your house and these in time becomes a lot more like R2-D2 to you. It becomes more like C3PO to you than it feels like a PC to you.

[01:23:48]There is no question this reinvention of the computer is as big of a deal as the reinvention of the phone into what we now know as the smartphone. And so this is the beginning of that journey. This is the beginning of a new line. And so we have a roadmap for this. This is a brand new product family for us. every single generation of architecture, we will have a desktop, a laptop, a workstation, and then a desktop, a laptop and workstation. And the thing that I am just incredibly pleased, incredibly honored is that 100% of the world's PC industry has joined us to reinvent the PC. A new line, a new beginning. Thank you.

[01:24:35][applause] As you know, agentic AI is just a digital robot.

[01:24:54]It understands, it reasons, it plans, and it acts and use tools.

[01:25:01]Agented AI is going to run across all of these computers and you've seen me talk about each and every one of these over time. We're working on human or robotics computers, robotics computers of all kinds. We're working on self-driving car computers. We're working on satellites.

[01:25:19]You have GeForce which is has tensor cores. I just talked about a whole new line of PCs. agriculture equipment, manufacturing equipment, heavy industry equipment will all be agentic. You'll even have a little agentic helper for yourself.

[01:25:36]Even your base stations, the radio stations of the future are going to be agentic.

[01:25:42]understanding traffic and thinking about how to coordinate with the other base stations so that you could use as little energy as possible increase the utilization the efficiency of the spectral efficiency and so everything will run agents today Nvidia is largely in the center but I am pretty certain that there will be tens of billions hundreds of billions over time of agentic systems agentic computers that are going to be running around the world. The biggest problem is data.

[01:26:17]In the case of language models, all the English and all the language that we have on the internet that we trained on was from the perspective of us. We wrote it and we're reading it. However, in order to create a data for AI robotics, it has to be in the perception, the perspective of the robot. And most of the world's video data is from a third person, not first person. And so agentic systems, robotic systems, physical AI, the data is the hardest problem. You've seen us move up this ladder. We started with tea operations, which is basically human demonstration.

[01:26:58]This is no different than the big breakthrough of reinforcement learning, human feedback. This then we use simulation. This is where omniverse comes in. This is no different than reinforcement learning ver verifiable rewards. Okay. And so we use these systems to bootstrap the AI model, the physical AI model.

[01:27:21]Eventually we're able to learn from third per third person reproing it into first person. And now eventually through bootstrapping we have a world foundation model that can understand the physical world from any perspective you want.

[01:27:39]Third third third person first person outside in inside out doesn't matter.

[01:27:44]This is a big breakthrough indeed. And today we're announcing Cosmos 3. Cosmos 3 is the frontier of physical AI.

[01:27:59]We are at the frontier with language models. There are so many people working on it. However, in physical AI, we are absolutely the world's best. I am so proud of the team for doing this. This is the foundation model for all of your work. Whenever you want to create a robot, whenever you want to create a factory robot or a robot that works in a factory, any kind of robot that in that involves physical world, you now have a companion, a Cosmos 3 that can understand and reason, it can generate, it can simulate in the loop, it can even be the policy itself. It is on the top of leaderboards all over the all over the world. I am incredibly proud of Cosmos and today we're announcing Cosmos 3. Let's take a look.

[01:28:47]>> The real world is infinite [music] and unpredictable.

[01:28:51]Physical AI needs data, but real world data [music] is impossible to scale. For physical AI, compute is data. This is Cosmos, [music] an open frontier omnimodel for physical AI built on a new mixture of transformers architecture.

[01:29:09]Pixels, [music] action, sound, and language flow into the auto reggressive transformer, which reasons, plans, and instructs the diffusion transformer, which generates what comes next.

[01:29:21]Developers post-train Cosmos across embodiment and use cases. As a VLM, Cosmos watches the physical world, understands what's happening, describing scenes, and flagging what matters.

[01:29:34][music] As a world model, Cosmos generates physics accurate synthetic video from an image, text, or video.

[01:29:44]As a simulator, Cosmos closes the loop for policy training and evaluation. And as the foundation of NVIDIA Omnidreams, an action conditioned world model, Cosmos predicts the future frame by frame.

[01:29:59]Post-train Cosmos and it becomes a world [music] action model. Perceiving, reasoning, planning, generating actions for robots of every kind, for everything [music] that moves.

[01:30:16]A new kind of data, a new kind of teacher generated by compute.

[01:30:23]Cosmos, the foundation for developers of the age of physical [music] AI.

[01:30:37][applause] It's takes data plus compute, gives you AI.

[01:30:46]Now that we have AI, compute is data. And so use Cosmos 3, train a whole bunch of AI models. Cosmos is such an incredible open model system.

[01:30:56]It's exactly the same as Neotron. We open the model, we open the data, and we even opened how we trained it so that you could enhance it for yourself and turn Cosmos into your proprietary model.

[01:31:09]We have such incredible partners working with us in so many different industries.

[01:31:13]Now the model itself is the most of course the most understandable part of the AI stack but the AI stack is very complicated. It has generators the model simulators and the runtime just as just as it is for agentic systems. these cars or essentially a physical AI agentic robot that is a is a autonomous vehicle has also this complicated stack. Today we're announcing Alpammyo 2, an open model for self-driving cars. We're working with car companies across the world. If you look at these brands that have signed up for the Nvidia Hyperion that are building Nvidia Hyperion cars, this represents about 80% of the world's cars. The manufacturers represent 80% of the world's cars. We are going to have a whole lot of Nvidia Hyperion systems that are able to run Alpha Mayo or anybody else's AV stack.

[01:32:18]We are also connected into mobility services. Approximately 97% of the world's mobility services are connecting with us. So that when we deploy Alpaio on the Hyperion runtime with the Halos operating system, we will be able to connect to all of these services across the world. Let's take a look at this.

[01:32:41]>> Hey Mercedes, let's go to my favorite sandwich shop.

[01:32:45]>> Routing to your destination.

[01:32:49]Lane is clear. pulling out to start drive. Nudge left due to the stationary leave vehicle ahead [music] blocking our lane. Slow down to stop at the stop sign controlling the intersection. Stop to yield to the pedestrian since the person is in our lane. Yield to the cut in vehicle from the left. Nudge left to clear the stopped vehicle blocking on the right. Keep distance to the cut in vehicle since it is merging into our lane. Nudge left due to the stopped van blocking out of our lane. They lane [music] crossing ahead. Stop to keep distance to the lead vehicle. Keep distance to the vehicle directly ahead in our lane. Keep distance to the vehicle directly ahead in our lane. Stop the stop sign since the intersection is controlled to yield to the cross traffic since the vehicle is crossing ahead. Keep distance due to the truck blocking the right side of our lane. Due to the truck blocking the left side of our lane due to the truck side, >> your destination is on the right.

[01:33:36][music] Alpao, [applause] the world's first reasoning autonomous vehicle.

[01:33:51]If you let it talk all the time, it will drive you crazy.

[01:33:56]But we're very happy that it's talking to itself all the time. That's called thinking. And so, Alpamo is a reasoning car. The technology that we've created also applies to humanoids. Of course, there are many new breakthroughs that has to happen. The NVIDIA Isaac Groot is our humanoid robotic stack model data generation simulation the runs the runtime including the operating system this represents group platform the Isaac group platform every one of our systems as you can see the exact same pattern whether it's agentic system for the cloud agentic system for the PC, a robotic system for a self-driving car, a robotic system for a human or robot, all the same. And of course, in every single case, we build everything completely.

[01:34:56]We build everything vertically, completely integrated with code design, extreme code design, and then we open it up for everybody to use whichever part you like. And whatever you want to use, we even help you modify. But the one thing that is missing is we need a reference platform for robotic systems. These robotic systems are so complicated. So many motors, so many sensors, so fragile. And yet we need to have a way to deliver these reference platforms.

[01:35:29]Just like we do with PCs and DGXs and clouds and self-driving cars, we now are going to do it for robots. Today we're announcing the NVIDIA Isaac Groot, a reference humanoid robot, all fully integrated. 25 degrees of freedom on the on each hand made by Sharpupa. 31 degrees of freedom on the robot. 6 feet 150 lbs. Just like me.

[01:36:00]The first number is shorter. The num second number is bigger.

[01:36:04]Otherwise, pretty close. and and this platform runs the new Thor and our entire software stack, data generation stack, data simulation stack, the runtime, all integrated into a robot that is designed for everyone to use.

[01:36:21]Now, we built this for higher education and university researchers because for them to build this is in insanely hard to do. And so, let's take a look at that.

[01:36:32]The next leap in AI is generalpurpose robots, humanoids. But building one is hard. [music] Every team starts from scratch, stitching together simulators, teleyops systems, data pipelines, [music] and training infrastructure.

[01:36:46]Months of setup before research can start. NVIDIA Isaac Group, an open development platform for humanoid [music] robots, open models, simulation and training libraries, and data [music] generators.

[01:37:00]Plus the robot computer fully pipe clean, ready to go in hours. First, set up the simulation environment in Isaac lab.

[01:37:12]Capture demonstrations [music] with Isaac Teleyop on a real or simulated robot.

[01:37:20]Generate synthetic data with Omniverse and [music] Cosmos.

[01:37:24]Scaling one demonstration into thousands.

[01:37:28]Train policies, evaluate them in Isaac [music] Lab Arena, deploy through Isaac Ross running on Jets and Thor.

[01:37:49]Every element [music] modular, open, use ours or swap in your own.

[01:37:56]Groot is powering robotics research across every discipline for every domain from research labs to factory [music] floors.

[01:38:06]One open platform and now a new addition. Isaac Groot reference design robots built on NVIDIA's open platform ready for frontier research for any lab anywhere.

[01:38:26]The age of robotics starts here. Nvidia Isaac [music] Groot.

[01:38:34]So many robots. [applause] We're working with just about everybody who's working on robots in world or robotic systems in world. Let me tell you what I told you. The computer industry has been completely changed in the last six months. Everything changed.

[01:38:56]Everything changed because agents were realized and it converged with the latest frontier models and it made possible the AI to now do useful work.

[01:39:08]The computing pattern will repeat over and over and over again. This computing pattern of an agent that's a model, a harness that uses tools with skills and runs in a runtime. That runtime depends on whether it's in the cloud or on prem on a PC or in a robot. But the computing pattern is exactly the same for all of them. You will use different harnesses because of your preference. You'll use different models because of your preference. You will improve them for your proprietary use. You would create sub super agents that you can rent to other people to help them do their work.

[01:39:44]This agentic platform, this agentic pattern, Nvidia has an enterprise AI toolkit. This is a wonderful way for all of you to engage AIS and for us it's a wonderful growth opportunity.

[01:39:58]Vera Rubin is in full production whereas Grace Blackwell was created to process AI particularly inference. Vera Rubin was created to run agents. It is in full production. It is much much more than a GPU. It is an entire disagregated distributed agent processing system.

[01:40:21]Nvidia has really become an infrastructure company. Not just a GPU company, not just a systems company, but an infrastructure company to help you generate the maximum revenues, the maximum profit, and to get there as soon as possible.

[01:40:36]the agent world.

[01:40:39]This new way of doing computing where you build CPUs now for agents not for people CPUs for agents has its own special requirement and our Nvidia Vera is revolutionary. I'm so happy about its ramp the orders already. It's going to make it the fastest and the most successful product launch in our company's history. Nvidia and Microsoft has created a whole new line of PCs.

[01:41:05]This is a new beginning. And of course, that exact same agentic pattern that I agentic processing pattern, computing pattern that I just described is also going to run on all kinds of devices. I mentioned PCs, but in the future, it'll be robots and satellites and base stations and factories in the cloud on prim at the edge. This pattern, agentic AI system, this agentic computing pattern will be replicated in computers all over. How we think about the personal computer will very likely change. I want to thank all of you for your partnership, your friendship. We couldn't be here without everything that we do together. I am so proud of how you've been so successful this last year. The next year is going to be even more. I have one more thing for you.

[01:41:56]Let's take a look.

[01:42:08]>> [screaming] >> You ready, Tom?

[01:42:22]>> Let's do this. The keynotes done at Computex [music] Jensen. Show the world what's next.

[01:42:28]Useful AI has arrived. Agents working [music] by your side. But in case you miss things we said today, we're going to break it all down for you. [music] Taipei. Agents used to be misunderstood.

[01:42:41]Only movie stars had them inollywood.

[01:42:43]Now we all got [music] teams making dreams come true. Building companies from living rooms, but they need so much comput. [music] We hear you. That's why we created Vera >> Reuben stole the show. It's true. The cheapest tokens coming through. [music] >> 10 times faster inference heaven. More special agents than 007.

[01:43:03]>> Blue field keeps [music] agents memory.

[01:43:05]True.

[01:43:05]>> Now let's talk about it. CPU >> 50% faster. [music] That's outrageous.

[01:43:10]>> Not for Ver.

[01:43:11]>> IT'S BUILT FOR AGENTS. Envy link fusion blends A6 [music] smartly.

[01:43:15]>> Everyone's welcome to the Envy Link party.

[01:43:18]>> Well, if you like that introduction, Ruben in full production ultra lead [music] the run 5x work gets done keep the guard rails right open shell keeps [music] the sand tight >> your code migrated and reviewed all before this song is [music] a five layer CAKE NO mistake global [music] AI class with lots of gigawatt DSX keeps power connecting dots >> every watt optimized for [music] you >> so You can have your cake, >> can eat it, too.

[01:43:53]>> RTX is finally here.

[01:43:56]>> Biggest PC moment in [music] 40 years.

[01:43:58]>> Agents powering our workflow. Running anywhere Windows go.

[01:44:02]>> Harnesses run on CPU. [music] >> Models fly on GPU.

[01:44:07]>> Cosmos worlds that robots need.

[01:44:10]>> Turning comput. [music] is how [music] they learn to move.

[01:44:20]>> Learning skills and finding growth [music] is powered by >> the future humanoid.

[01:44:29]>> [music] [music] >> Oh yeah. Oh yeah. Oh [music] yeah. Oh. [singing] Oh.

[01:44:58]The future's bright. Come see what's [singing] next.

[01:45:05]>> Thank you, Taiwan. [sighs] >> Welcome to Computics.

[01:45:11][music] [music] [applause] Have a great tax. Thanks for an amazing year. Thank you for all your friendship and support. Thank you. Take care. Have a great complex.

[01:45:33][applause] [music] Woke [music] up feeling something shift.

[01:45:49][singing] Same room, but the air felt thick.

[01:45:52]>> [music] >> mirror said I'm still that kid I said you're bigger than this time I laces hand still shook first [music] step was the hardest one every lap every sideways look all fade when the races run I carried every doubt [music] lightweight now I'm lighter than [singing] the flame you can write it on the [music] page everything CHANGED A LEGEND WAS MADE TODAY FROM EVERY LITTLE SCAR every heartbreak took all of the nights I [music] almost walked And I turn in the name.

[01:46:25]This place