Yes, even Nvidia's head of automotive is fighting for compute

Summary

---

Decoder with Nilay Patel — Episode Summary

Title: Yes, even Nvidia's head of automotive is fighting for compute
Date: July 13, 2026
Host: Nilay Patel, Editor-in-Chief, The Verge
Guest: Xinju Wu, Head of Automotive, Nvidia

---

Episode Overview

In this episode, Nilay Patel speaks with Xinju Wu, the head of Nvidia's automotive division, about how the auto industry is grappling with the push toward electrification and autonomy amidst economic and geopolitical turbulence. They examine how Nvidia's technology powers the next generation of vehicles, the shift to software-defined and AI-defined cars, challenges facing legacy automakers, synthetic data and model training, and what it’s like to compete for resources inside one of the world’s most in-demand tech companies. They also discuss the current and future states of full self-driving, including the controversial role of LiDAR, regional trade barriers, and what truly defines a mainstream autonomous driving experience.

---

Key Topics & Discussion Points

1. The Fundamentally Changing Nature of the Car

[01:19–07:36]

• Software-Defined and AI-Defined Vehicles:
  Xinju Wu describes the industry’s ongoing transition: from largely mechanical machines, to software-defined vehicles (SDVs), and now to AI-defined vehicles, where generative AI rewrites vehicle software and enhances capabilities through over-the-air (OTA) updates.

  “What we call the AI-defined vehicle... we’re using AI to rewrite most of the software in car. And that is... accelerating the development pace of the vehicle capability.” — Xinju Wu [05:18]

• Centralized Compute Architecture:
  The conversation highlights the industry's move away from hundreds of electronic control units (ECUs) to single or dual centralized computers, following trailblazers like Tesla and Rivian.

  "I have no doubt basically the world is evolving that direction." — Xinju Wu [08:58]

2. China’s Head Start in Electrification and Platform Innovation

[10:20–12:37]

• China’s Advantage:
  Wu, formerly at Xpeng, explains how China’s lack of legacy burdens and heavy state support enabled faster adoption of fresh EV platforms.

  "Not only the new OEMs, but even the global player there... have to adapt to the China pace. Otherwise... you won’t be able to compete." — Xinju Wu [11:16]

• Industry Consensus:
  Moving to centralized compute and eliminating ECUs is now considered a necessary baseline for next-gen car architectures.

3. The Bumpy Road for Legacy Automakers

[12:37–16:17]

• Transition Difficulty:
  Nilay probes why the shift to SDVs has been slower and rougher than predicted. Wu cites the massive automotive supply chain, need for decade-long support commitments, the challenge of shifting company cultures, and rapidly evolving technology.

  “To make a change on the architecture... you have to support it for 10, 15 years... It’s almost insane.” — Xinju Wu [13:54]

4. Inside Nvidia Automotive: Team Structure and Resource Fight

[16:17–22:08]

• Organizational Structure:
  The Automotive division numbers in the thousands, globally distributed but mostly US-based, with multidisciplinary teams: hardware, software, model development, mapping, etc.

• Resource Competition Within Nvidia:
  Despite Nvidia’s AI-fueled dominance, Wu’s team must “fight for” GPUs and fab capacity, often requiring CEO Jensen Huang's direct involvement.

  “Believe it or not, even Nvidia... we do have a limited supply of GPU for compute... Sometimes we need Jensen to help.” — Xinju Wu [19:41]

• Leadership & Strategy:
  Wu admires Jensen’s technical depth and strategic vision:

  "It’s a once in a lifetime experience and opportunity for me to be able to learn from Jensen." — Xinju Wu [21:17]

5. Nvidia's Platform & Revenue Vision: From Chips to Miles

[22:08–25:20]

• Trillion-Dollar Ambition:
  Wu predicts a world where “everything that moves will be autonomous” and Nvidia will take a cut from the “revenue per autonomous mile.”

  “We hope that we can get a percentage of the revenue that the ecosystem can get from every mileage... This is where the trillion dollar opportunity can come from.” — Xinju Wu [22:33]

• Platform Approach:
  Nvidia offers layers — from chipsets, operating systems (“Halos”), and open-source models to simulation and training — allowing OEMs to engage at any tech level.

6. Industry Dynamics: Openness, Data, and Training

[29:23–38:56]

• Open Model for Automakers:
  OEMs can pick their level of engagement, from in-house chip design (like Tesla) to turnkey solutions.

  “The beauty of Nvidia’s business model... our platform is completely open.” — Xinju Wu [32:42]

• Data Sharing and Synthetic Data:
  To catch up with leaders (e.g. Waymo, Tesla), Nvidia's platform encourages OEMs to share real and synthetic driving data, leveraging massive simulation and neural reconstruction to generate rich training scenarios.

  “The cost saving is... enormous. Data collection, running a fleet of huge size... is capital spending for anybody who wants to do that.” — Xinju Wu [38:21]

7. AI Models, Foundations, and Safety

[38:56–49:54]

• From Classical Stacks to Reasoning Models:
  Nvidia runs both classical (module-based, safety-certified) and end-to-end machine learning models in parallel for redundancy.

  “The classical stack is acting like... a safety guardrail.” — Xinju Wu [42:50]

• Language-Based Reasoning in Cars:
  Next-gen models will reason in natural language, allowing for introspection and querying by users or engineers.

  “Short answer is yes. In our next generation model... we will have the model trained with language embedded. So being able to reason through language is very important.” — Xinju Wu [46:47] “You can ask the model about what it's doing and then you can also ask the model to speed up or slow down and make a lane change.” — Xinju Wu [47:19]

• Latency & Real-Time Performance:
  Despite rich reasoning, inference times are kept under 100ms to ensure safety.

• Do Models Have Emotions?
  Nilay jokes about the possibility, Wu responds:

  “We will use the guardrail to make sure it doesn’t get too moody.” — Xinju Wu [50:06]

8. Local vs. Connected Processing

[53:44–55:40]

• Validation and model training happens offline; real-time safety checks and trajectory comparison run locally in-car.
• Cars must be able to operate safely even without connectivity (“blind spot” scenarios).

9. Hardware Requirements, Cost, and Sensor Debate

[56:41–62:09]

• Falling Hardware Costs:
  Wu notes dramatic drops in radar and camera prices thanks to economies of scale, and expects compute to become more affordable.

• Balancing Cost and Capability:
  Nvidia’s Hyperion platform provides different sensor sets: cost-effective L2 (no LiDAR) vs. high-end L4 (multiple LiDARS and radars).

• On LiDAR for Autonomy:

  “We believe that LiDAR is the important sensor... for level four autonomy. But... it’s difficult to say it’s 100% necessary.” — Xinju Wu [62:28]

10. Tesla, FSD, and the State of Level 4 Autonomy

[63:23–66:36]

• Where Tesla Stands:
  Wu acknowledges Tesla’s L2 ADAS dominance through its vision-based approach, but sees L4 (true autonomy) as more open—pointing to Waymo’s proven, if limited, deployments.

  “Nobody would argue that Elon is ahead of everybody in the L2 or basically ADAS kind of market... For level four... it’s more open.” — Xinju Wu [64:07]

• LiDAR Debates with Tesla:
  Wu’s team has interacted with Tesla about LiDAR, though he hasn’t personally; he’s open to further dialogue.

11. Geopolitics: U.S.–China Tensions and Regulatory Barriers

[66:36–71:28]

• Regional Fragmentation:
  Due to trade restrictions and rules on data, Nvidia must keep region-specific model variants.

  “We are conformed to all the local kind of regulation... the model will behave differently in different regions based on the input.” — Xinju Wu [70:17]

• Who Will Win in L4?:
  Despite China's dynamism, Wu credits Waymo’s San Francisco service as the current global leader in mainstream L4, and disputes predictions that China will necessarily get there first.

  “I actually don’t think that’s true... I think the only one who really have proven that L4 can be safely deployable... is still in US, not in China.” — Xinju Wu [71:28]

12. Looking Ahead: What's Next from Nvidia?

[73:47–74:03]

• Roadmap:
  By end of 2026, Nvidia’s ADAS tech will be standard in all Mercedes vehicles, with expanded deployments worldwide. The L4 platform is rolling out in partnership with companies like Uber.

  “80% of the mass production OEMs are in Nvidia’s Hyperion basically ecosystem for L4.” — Xinju Wu [74:03]

• Bold Prediction:
  Wu forecasts mainstream, L4-level autonomy in under five years.

  “I would say five years, less than five years.” — Xinju Wu [73:47]

---

Notable Quotes & Memorable Moments

• On the AI Revolution in Cars:

  "Now technology advanced towards generative AI... we're using AI to rewrite most of the software in car." — Xinju Wu [05:18]

• On Competing Internally at Nvidia:

  "Believe it or not, even Nvidia... we do have a limited supply of GPU for compute... Sometimes we need Jensen to help." — Xinju Wu [19:41]

• On Openness to Automaker Engagement:

  "We're not picking winners per se. We try to help OEMs based on their capability at different levels." — Xinju Wu [32:42]

• On Synthetic Data Sharing:

  "The cost saving is... enormous... data collection, running a fleet of a huge size... is capital spending... it's kind of repetitive as well." — Xinju Wu [38:21]

• On Making AV Safe:

  "Safety is so important to us... we have developed very sophisticated development protocol and validation protocol... ISO 26262..." — Xinju Wu [42:50]

• On Language Reasoning Models in Cars:

  “Short answer is yes... we will have the model trained with language embedded. So being able to reason through language is very important.” — Xinju Wu [46:47]

• On the Humor and Weirdness of “AI Chatting With Itself” While Driving:

  "The model is talking, constantly explaining what it's trying to do. It can be quite annoying... if you hear everything the model is trying to reason about." — Xinju Wu [48:02]

• On the LiDAR Debate:

  “We believe that LiDAR is the important sensor... to get to really high level of both urban and highway level 4 capability... but theoretically people can prove out... without it.” — Xinju Wu [62:28]

---

Timestamps for Key Segments

• State of the Auto Industry & Architecture Shifts: [01:19–12:37]
• Why Legacy Automakers Struggle: [12:37–16:17]
• Inside Nvidia Automotive, Resource Competition: [16:17–22:08]
• Nvidia’s Platform & Revenue Per Mile: [22:08–25:20], [29:23–31:04]
• OEM Engagement & Data Sharing: [31:04–38:56]
• Synthetic Data & Model Training: [35:35–38:56]
• AI Reasoning/Language Models & Safety: [38:56–49:54]
• Local/Cloud Processing, Connectivity: [53:44–55:47]
• Hardware, Cost, Sensor Debate, LiDAR: [56:41–62:28]
• Tesla & Vision of Full Autonomy: [62:09–66:36]
• Geopolitics, US-China Divide: [66:36–71:28]
• China vs. US & Level 4 Mainstream: [71:01–73:47]
• Nvidia’s Roadmap & Industry Prediction: [73:47–75:08]

---

Conclusion

This episode is a comprehensive look at the auto industry’s technological crossroads, viewed from inside Nvidia. Wu presents a candid perspective on industry shifts, the realities of transforming automotive electronics, the crucial (and competitive) role of compute, and the complexity of building trustworthy AI for autonomous vehicles. The conversation is technical, transparent, open about challenges, and peppered with both optimism and realism—especially as it relates to data sharing, the race for autonomy, and the tension between serving legacy automakers and startups alike in a geopolitically divided world.

Transcript

[Sponsor Announcer]

Support for the show comes from ServiceNow. AI is moving fast across the enterprise, but without visibility, it's just chaos. Different tools, different models, different teams using AI in completely different ways. ServiceNow turns that chaos into control. With the AI control tower, you see all your AI across the business in one place, what it's doing, what it's done, and what it's about to do. So you stay in control. To put AI to work for people, visit ServiceNow.com,

[Xinju Wu]

this Monday.com ad was created by a team of people and AI agents. The agents wrote the copy and managed the timelines, while our human creative director made sure it all made sense.

[Sponsor Announcer]

Easy.

[Xinju Wu]

Create your own AI agent today on Monday.com when you need to build up

[Neil C. Patel]

your team to handle the growing chaos at work, Use Indeed Sponsored Jobs.

[Xinju Wu]

It gives your job post the boost

[Sponsor Announcer]

it needs to be seen and helps

[Xinju Wu]

reach people with the right skills, certifications and more. Spend less time searching and more time

[Sponsor Announcer]

actually interviewing candidates who check all your boxes.

[Xinju Wu]

Listeners of this show will get a $75 sponsored job credit@ Indeed.com podcast. That's Indeed.com podcast. Terms and conditions apply. Need a hiring hero? This is a job for Indeed Sponsored

[Neil C. Patel]

Jobs hello and welcome to Decoder. I'm Neil I. Patel, Editor in Chief of the Verge, and Decoder is my show about big ideas and other problems. Today I'm talking about Jinju Wu, who is head of Automotive at Nvidia. Nvidia is obviously in the news constantly right now because of the AI boom, and it's one of the most valuable companies in the world because the AI industry can't get enough of the company's GPUs. But Nvidia is also a key supplier to the auto industry. It's had chips in cars for years now, and Jinju has been instrumental in building a complete autonomous driving system that automakers can just use. It's already in place in newer Mercedes EVs, as you'll hear him mentioned several times. So I really wanted to get his perspective on how the auto industry is handling the big transition to self driving EVs. The goal that every carmaker and supplier will tell you is coming coming, but which seems maybe farther away in 2026 than ever. The EV adoption cycle in the United States is fully off track. Self driving seems to forever be stuck trying to solve the final 20% of situations, and cars themselves just keep getting more expensive even as consumers are feeling the squeeze of inflation and rising energy prices across the board. You'll hear Xin Zhu say that there's actually startling progress in reinventing the fundamental nature of the car itself, something the industry has long called the software defined vehicle controlled by just a handful of powerful computer instead of dozens or even hundreds of independent electronic control units or ECUs. If you're a decoder listener, you have heard so many carmakers talk about the need to get away from ecus. Jinzhou says that moment is basically here. We also talked a lot about the Chinese car industry and how it's been able to essentially get a head start on all of this because it began building on EV architectures and platforms instead of having to manage a transition away from gas cars and all of those ecus. Xin Zhu used to work at a Chinese oem, so he has quite a bit of insight here. He also talks about working at Nvidia itself. It's a unique company with a unique leader in Shenzhen. Wang and Xinju said his three years there so far have been a rapid learning experience. He didn't shy away from the reality of needing to compete for resources and manufacturing capacity against the company's booming AI business. His description of what wins those arguments, especially when his customers are as slow and cost averse as automakers, that's fascinating. Of course, we also talked about AI and how Nvidia's approach to autonomy brings together what Xinju calls the classical stack and the ability for reasoning models to actually operate the car. There's a lot here, including the idea that you'll have an AI model literally talking to itself to figure out how to drive your car, which I find both incredibly interesting and incredibly funny. Of course you can't talk about electric cars or autonomous vehicles without talking about Tesla and Elon Musk. So I assumed you pretty directly where Tesla is on the full self driving curve and whether that technology can actually do what Elon claims can do without having to put lidar sensors on the car. You tell me if you think his answer holds up. Okay. Xinju Wu Head of Automotive at Nvidia Here we go. Xinjiai Wu, you are the head of automotive at Nvidia. Welcome to Decoder.

[Xinju Wu]

Thanks for having me.

[Neil C. Patel]

I am really excited to talk to you. It feels like the very nature of what a car is is up for grabs. It feels like the automotive industry is in a period of massive realignment. Almost as though there was a sense of where the the car was going to end up as a product for several years and that is because of EV transition difficulties because of U S China trade war difficulties. All of that seems More messy than ever before. A lot of carmakers are retrenching and it feels like your position in Nvidia gives you a pretty wide view of what's going on in the car industry because you supply so many of the major automakers in virtually every country. So let's just start there. What's your view of where the car industry is on this kind of long winding road to both autonomy and electrification?

[Xinju Wu]

That's an excellent question. Actually, I've been working in the auto industry. Well, not exactly in the auto industry, but you know, let's say working in the automotive sector for probably 15 years, starting from my career in Qualcomm. I was heading the Qualcomm automotive team for a while. Obviously we have heard the word basic software defined radio, sorry, software defined vehicle. And then basically right now with the AI technology, it's really getting to the next phase. What do we call AI defined vehicle? Essentially, with this massive technology innovations, as you said, the auto, you know, industry is, is changing pretty rapidly, I would say over the last decade. As you know, I also worked as part of a Chinese OEM for a while, for five years, heading their automotive, you know, autonomous driving team. Now basically I'm, you know, Nvidia. So what I've seen over my 15 years of career is really basically have the, I would say the opportunity to witness this massive change the car from a, let's say mostly mechanical, obviously a plus electrical, basically machine, to some things that basically we can kind of upgrade the capability through OTA software. OTA pretty rapidly. That's what we call this software defined vehicle era. And now basically technology kind of advanced towards a generative AI. We are seeing, basically we're using AI to rewrite most of the software in car. That's what we call the AI defined vehicle essentially. And that is also, I think in one hand accelerated the development pace of the vehicle capability. And in the other hand it's also basically other than software. It also basically changed the way how we define vehicle as well. AI is impacting the whole industry at every level. So this is really exciting to see how the world will evolve from here with this new technology innovations.

[Neil C. Patel]

Let me pull apart some terms there. I hear them a lot from carmakers love to come on the show and tell me what's going to happen to cars. But I think some of these terms are a little bit fuzzy on the edges. So you said software divine. Vehicle.

[Xinju Wu]

That's right.

[Neil C. Patel]

That's a pretty fuzzy term. Right. I think the idea there is we're going to get rid of all of the use in a car that currently control lots and lots of different systems and we will centralize all of those components into maybe one or two big compute centers in a car. Tesla is very famous for having done this. Rivian is. They've made a huge bet on that. Wasim Ben said from Rivian was just on the show talking about that Other legacy carmakers have tried to do this. We had GM on the show, they said look, that we don't need to do that, we're fine, we'll do it our way. Ford tried to do this in big ways. They had to set up a skunk works and build an entirely new kind of way of making a car that they're very proud of. There'll be a truck coming out from that effort sometime soon, we're told. I don't think the industry got there. That's basically what I'm saying. Like the startup car makers got to the point where they could claim to have a software defined vehicle where there were one or two big computers in the car controlling every system. The legacy automakers for the most part have not succeeded yet. And I'll just put an asterisk, maybe, maybe Ford will succeed with this new truck, but we don't know yet. Do you think the industry broadly is going to get to software defined vehicles or you think the legacy automakers are

[Xinju Wu]

going to stay where they are 100%? Again I had the, let's say, opportunity to witness what happened in China from 2018 to 2023 and the whole industry went through this massive change just in five years over there. Not only the new OEMs, auto OEMs, but also the legacy ones, they have to adapt and everybody is adapting to a basically single central computer kind of electrical architecture because that's how you compete in rest of the world as well. Obviously we have our partners as well through basically drive and drive AV basic collaboration. For example, our partner Mercedes, their current generation is basically essential compute based architecture. It's going to be in all their vehicles. And for the other basic OEMs, we're obviously working with all of them and trying to help them basically upgrade the architecture to one or two computers route because there will be infotainment, there will be basically driving or adas, ecu. But I think definitely the world is actually moving pretty rapidly in that direction. Well, some of them obviously will be slower, some of them will be faster. That's the nature of this business. But I have no doubt basically the world is evolving that direction.

[Neil C. Patel]

I'm actually Curious about your history. You worked at Xpeng, which is a Chinese car maker. It feels to me, sitting where I sit in the United States and being a car fan for a long time, that that Chinese automakers had a fairly unique advantage in that they were not big global automakers, they were not operating at massive scale. Electrification came Tesla obviously built a bunch of capability in China to make cars. We all know how the Chinese manufacturing ecosystem works and they got to reset, they got to design a bunch of cars as evs clean sheet. Basically the way the startup car makers United States got to do and build globally competitive cars from a totally new foundation without having to worry about a bunch of the stuff that I don't know, legacy American car makers would have to worry about. And then the Chinese government obviously subsidized all that at huge rates. You worked there. Was that your experience? Is that, is that basically how it went? They got to start fresh?

[Xinju Wu]

I think that's just one side of it. Definitely have less legacy. Basically a burden to worry about is an advantage. But what I also see is not only as I said, the new OEMs, but even the global player there, they have to adapt to the China pace. And basically, at least from what I learned over there, everybody is going through that pace. Otherwise again, you won't be able to compete. But again, as you said, the wave software defined vehicle has been there for a long time and Tesla is the one that's really basically I think taking it to full production. I'm not sure of the first one, but basically definitely to the, I would say largest extent and the only way to get there is to get to first of all the architecture described, this kind of architect can enable kind of software upgrade without have many, many let's say discrete ecus. Actually I haven't heard people arguing against that recently. Maybe you heard something different, but I think that's really a necessary step for everybody at this stage. It's really almost like a table stake for the next generation architecture. Obviously we are talking to a lot of OEMs, but this is, I think to say the least there's a consensus that the industry is moving towards.

[Neil C. Patel]

Yeah, I'm just curious about the pathway there because I agree with you that many, many people have said that is the end state and that enables everything that's going to. It just feels like the, the path there has been much bumpier than the industry expected. And part of that is, I don't know, the Trump administration doesn't like EVs. So EV sales and the tax credits Here went away. And maybe EV sales spiked as all that demand got pulled forward. And maybe everybody wants a gas car. And maybe all of this is harder when you don't have a giant battery that can power all of these systems in perpetuity and you actually need to start the engine to get power to all these systems instead of having a 12 volt battery. Or maybe it's the Chinese automakers are so competitive and so subsidized that the cost to do it for the legacy automakers is hard to overcome. Right. Because they do have the legacy infrastructure and dealer networks, the United States to care for. And we're just going to hold off on it. Right. There's something about the path to this agreed upon future state of the car that seems harder than I thought it would be or that anyone on the show over the past five years has said it would be. And I'm curious, from your perspective, like you're the supplier, you're trying to sell the vision, you're trying to put the chips in all the cars. From your perspective, what has made that path harder?

[Xinju Wu]

The auto industry is very heavy. You know, it involves basically a massive supply chain and lots of companies, lots of employees essentially. And to make a change on the architecture. And whenever you push out a car, you have to support it for 10, 15 years, basically. Nvidia, obviously as a supplier, we also make a similar commitment to our customers for travel technology we supply, including chipset, including other platforms and our AV technology, we will have the commitment to support for the same generation for 10, 15 years. Even for the current generation chip. If you think about it from Silicon Valley, from Silicon provider kind of perspective, it's almost insane. But that's the nature of auto business. It has a basically the nature of the business. It will kind of slow things down a bit. And that's one. And the other thing is basically because of the technology is changing so fast from let's say the automotive as we know before, and to software defined vehicle to AI defined vehicle. You have to go through almost like a different talent pool to be able to set up the company in proper way to adapt to this new wave of technology innovations. And that's why Nvidia can come in and help essentially. Right? Because we believe the technology is getting to, you know, we are talking about autonomous vehicle, obviously, you know, mainly here the technology is getting to a level of maturity and we are going to take in this technology to mass production and the supplier can come in. And that's why we are not only, you know, provide the AV technology, but we are providing the whole basically platform, you know, starting from obviously chip, but also to operating system, also to open source model, and also to what we call the Halos, the safety kind of operating system to help the OEM to be able to adapt to this new world faster. And the nature of the business is basically not everybody can run at the same speed. So for sure, and it will take some time obviously for everybody to get to the finish line. But again, my job in Nvidia is to try to help everybody to get to this, everything that moves, that will be autonomous, this kind of vision, as long as possible.

[Neil C. Patel]

Let me ask about your part of Nvidia now, because I think this brings us to the decoder questions. I think everyone listening to the show is probably very familiar with the run Nvidia has been on with AI. It's one of the most valuable companies in the world. Every GPU that Nvidia can make is accounted for. How many people work at Nvidia Automotive?

[Xinju Wu]

We have actually quite a sizable team, somewhere between basically in the order of thousands essentially in the automotive team. But it's a pretty, again, because we are working on the whole platform. So there's a hardware, software and model and the infrastructure. So it's a pretty sizable team. And also we have a lot of things we can leverage from the other teams as well. For example, we have pretty sure you heard about the Cosmos and Nemotron. These are our basic open source foundation models. We're leveraging heavily from work from that side as well.

[Neil C. Patel]

And how is your team organized? You mentioned you've got hardware, software, you've got models. Is that the basic structure of the team or is it organized differently?

[Xinju Wu]

Yes, I would. Well, on the engineering side, obviously we have product, we have strategy, we have something kind of behind the scene. Sometimes we call them unsung heroes. Right. The MAP team, for example, which is still very critical for L3L, for the high level autonomy paths and the data

[Neil C. Patel]

infrastructure, the literal navigation maps. That's what you're talking about?

[Xinju Wu]

Well, there's HD map as well. Okay, so it's roughly. That's, you know, I divide my team this way. Yes.

[Neil C. Patel]

And then is that all global? Is that mostly in the United States? Where's that located?

[Xinju Wu]

Mostly in the United States, but we do have a presence in China and Europe as well. Obviously we are building a global product, a global platform. So we need the support team everywhere.

[Neil C. Patel]

You mentioned that you rely on some of the foundation models Nvidia has developed. More broadly, how is your team structured? Inside of Nvidia, does it fit into the AI strategy? Is it set apart? Are you more siloed? How does that work?

[Xinju Wu]

So in Nvidia we have let's say centralized hardware team which are responsible for the hardware roadmap on our GPU basically and the CPU and all the chipset basically strategy and the productization and we have centralized software team and Automotive is a separate, I would say organization which is very much more automotive, basically focused with the mission of really building the automotive platform to leverage the work from our hardware team and the software team and adapt to Automotive. And then basically we have the model team as well. Open source model team, actually part of the opa. Nvidia also have a culture of virtual teams. For example our open source model for Nemotron and Cosmos, they all have across our research team and the software team and the hardware team, but they are virtual teams. That's basically to work on these open source foundation models and we can leverage basically those work and then basically in the automotive organization to build up model, for example, as hopefully you have heard about it, to help the AV industry basically have a powerful open source model to work on.

[Neil C. Patel]

As I said, basically every GPU Nvidia can manufacture is accounted for in some way. It's the nature of the AI industry right now. And they're going to go into some NEO cloud somewhere. Do you have to fight for resources and attention against that business which is growing at the speed and the scale it's growing at?

[Xinju Wu]

Yes, believe it or not, of course. So basically, for example, believe it or not, even Nvidia basically we do have a limited supply of GPU for compute, so we have an internal priority and I'm working with my colleagues basically almost on a weekly basis to decide how to to set aside these different compute. Sometimes for training, sometimes for test resources for different thread of work in the company. And sometimes we need Jensen to help.

[Neil C. Patel]

But yeah, how does that work? What does that debate look like? Is it a ROI debate? If we put this much money in, we'll get this much money out from our customers. Is it a market size debate? What are the parameters of the conversation?

[Xinju Wu]

Well, I think it's all of the above. As you can imagine, revenue is important obviously, but also basically Nvidia as you know, is a very strategic company. We value what sometimes Jensen calls the $0 trillion business. We are looking for new opportunities which can create a trillion dollar business all the time. So there need to be strategic kind of priorities we set inside the company. This is the new direction we go and you probably also know that we are not a market share company. So it's a balance between basically what's the current, what brings the money right now and what can create the future, create the opportunity for the company in future.

[Neil C. Patel]

Nvidia is a very uniquely run company as you've mentioned. Jensen's deeply involved in everything. I've seen an interview with him where he said he doesn't have one on one meetings, he just meets with everyone all at once and everyone just hashes it out. What's that like?

[Xinju Wu]

Well, I have been Nvidia for three years. I think it's very unique, honestly. It's not obviously not everybody all at once. Right. It's different groups, we all have technical strategy, product, different kind of part of the business reviews. With Jensen it's a super exciting for me actually a learning experience basically to learn from his strategic thinking and how he think about a product, how he think about a strategy. He's also uniquely technically deep. So it's also quite inspiring basic experience as well to just also to see how much he's keep up to date on the technical side as well. Again it's really, I would say once in a lifetime experience and opportunity for me to be able to learn from Jensen. Yeah.

[Neil C. Patel]

When you describe the opportunity for autonomy, particularly in the future, because that seems like the big bet, right? We're going to bring to bear Nvidia's compute excellence and the power of AI to cars and have them drive themselves. What does that revenue model look like? Does it look like you're just selling chips and software to automakers? Does it look like consumers pay a subscription and some of that flows back to you? Where, where does a trillion dollars come from?

[Xinju Wu]

So basically if you look at it basically right now we firmly believe that everything that moves will be autonomous. Every mile driven by the car in the future will be autonomous. So right now if you look at it basically among all the cars we drive 13 trillion miles basically per year and right now the percentage of autonomous miles among automatically driven is probably, let's say negligible. I think it's 0.006% or something like that. So this is really the opportunity in front of us. So Nvidia's view is basically will help the ecosystem to get there as soon as possible by providing basically all the foundation technology piece again starting from chip to operating system and then basically to what we call Halos. Again the Halos operating system is really important because it doesn't you. It not only provides the SDK and the APIs for folks to develop models on our hardware but also provide basically the safety guardrail for developer to put a model on it. And then basically we also define what we call the Hyperion basically hardware platform that's a production ready platform which include both the computer resource, the ECUS and also the sensor suite. We think it's necessary to achieve different level of autonomy and on top of that we provide basically the Mayo basically open source model which we trained and open source not only the model architecture but also the parameter and the data that basically you can use to fine tune the model on our platform. And on top of that we also provide basically all the infrastructure needed. For example simulation. Right now it's really important for developing av. We usually call it the AV problem is becoming three computer problem right as the training computer. There's the simulation compute and then there's the inference computing in the car. All these technology piece we want to provide to the ecosystem in a platform which we call Nvidia Drive essentially so that folks can develop their technology on top of our platform. And we hope that we can get a percentage of the revenue that the ecosystem can get from every mileage that driven autonomously in the future. This is where the trillion dollar basically opportunity can come from.

[Neil C. Patel]

We have to take a quick break here. We'll be back in just a minute.

[Xinju Wu]

IT.

[Sponsor Announcer]

Support for the show comes from Quo. Running a business means you're actually reachable. Until you're not. And the moment you miss that call, that text, that follow up a competitor is the one that picks up. That's why today's episode is brought to you by Quo, spelled Q U O. The business phone system built so you never miss an opportunity. All your calls, texts and voicemails live in one place, so anyone on your team can pick up a conversation, see the full history and respond fast. Quo is the number one rated business phone system on G2, trusted by over 90,000 businesses who rely on it to stay reachable and look professional. Every day, set up in minutes on any device, keep your existing number and add teammates as you grow. No it, no hassle. And Kuo says it's easy. Calls, text, voicemails, transcripts and contact details are all in one clean view, so your team always has the full picture and can show up for every customer conversation, ensuring a seamless and more personalized experience. Money is on the line. Always say hello with quo. Try quo for free plus get 20% off your first six months when you go to quo.com decoder that's q u-o.com decoder support for the show comes from Shopify. Landing on a good idea for a business is basically catching lightning in a bottle. What are the odds it'll happen more than once? You have to jump on every opportunity to make it work. And the last thing you need is to get slowed down by clunky logistics. You want tools that are ready to go right out of the box. And if your business involves E commerce, you need Shopify. Shopify powers millions of businesses worldwide. From household names like Mattel and Gymshark to small businesses just getting started. You could choose from hundreds of beautiful templates that you can customize to match. Your brand setup is fast with Shopify's built in AI tools that write product descriptions and headlines and help you edit product photos. If you've been sitting on a business idea, it's easy to make it happen. With Shopify Shopify, everything you need to start selling is included and ready from day one with Shopify. Nothing stands between your idea and a real business. So go make it one. Start your free trial at shopify.com decoder that's shopify.com decoder shopify.com decoder Support for this show comes from Vanta. Every new tool your team signs up for, every vendor that turns on AI features, every new integration has a chance for something to go wrong. And most security programs weren't built for AI's pace of growth. Enter Vanta. Vanta is the number one agentic trust platform used by over 16,000 fast moving companies like Ramp Cursor and Harvey to ensure they're always audit ready. And now Vanta is helping companies like yours watch for risks that show up between audits across your vendors, your AI tools and your whole environment. How the Vanta agent works like a 24.7grc engineer in the background. Finding issues, drafting fixes for you and cutting vendor assessment time up by 50%. Whether you're a fast growing startup or a global enterprise, Vanta is here to help you automate your security and compliance and earn and prove trust. Get started [e-mail] decoder that's V A N T A dot com decoder.

[Neil C. Patel]

Welcome back. I'm talking with Nvidia's head of automotive, Shinzhou Wu about how Nvidia fits into the larger auto industry industry. So revenue per mile, that that sounds like the core metrics that you're chasing. Where does revenue per mile come from for a user? When I drive a car, do I pay a subscription or are you thinking it's robo taxis everywhere and they're being monetized per ride. Where does revenue per mile come from and how does that number go up?

[Xinju Wu]

That's right. Well, I think the world will, I, I think we'll, we'll embrace both models. One is basically robot taxi. You know, as you see there's quite a few successful ones basically doing, you know, in China and in US and in the world. And we will see more hopefully going down this path. And basically, you know, we'll have like a taxi like fleet where you can enjoy and taking you from, you know, place A to place B without a driver in the car. And then I think the passenger fleet will also still, you know, continue to exist for a long time because you know, there's many people who still prefer a private base, basically, let's say space during travel. It's like many people still prefer their house as compared to rent an apartment. Obviously there's economy behind this as well. So we think both models will thrive. That's why we are working with both the Robotaxi companies and also the auto OEMs by providing and obviously the AV software developer companies to help them basically by supplying different technology piece from Nvidia to them.

[Neil C. Patel]

One of the interesting dynamics through I would say at least the electrification portion of over the past five years has been legacy automakers realizing that they had become insurance companies and financing companies and their suppliers were making the cars right and they had lost control of car design in like a big way. The tier one suppliers to the big automakers were in many ways in charge of big subsystems of the cars. And when they wanted to do an over the air update, they had to go talk to 15 different suppliers to get that done. I've heard this complaint dozens and dozens of times on the show and they all kind of realized, oh, we need to take back the engineering of the car. We need to be more, much more firmly in control of the platform of the car. It sounds like in autonomy, for a variety of reasons, Nvidia sees an opportunity to become the main supplier to a wide variety of carmakers. That's obviously in tension with them thinking, oh, we need to take control of the car, right? I think Tesla might run Nvidia chips, but they are very proud of the fact that they wrote every line of that code and that is their platform and they've made their technology bets Rivian. I think Weston is very proud of the fact that he is in charge of that platform company and he's going to build that platform. RJ is certainly very proud of the Fact that Rivian is that kind of company. What's the dynamic there? Because it, it doesn't seem like every car maker can stand up the technology bet and forward invest on the hope that the revenue will pay off, that they will need a supplier like Nvidia to show up with a ready made platform and business model. Is that tilting more in your favor now? Have we, have we gotten out of those woods or is it still up in the air?

[Xinju Wu]

I think the beauty of the Nvidia business model in the automotive side is really our platform is, is completely open. We provide multiple layers of services and depends on basically what OEM need or Robotaxi company needs. They can select what they want to work with us basically up to which layer. For as you mentioned, Basically Tesla, some OEMs, they're so capable they even want to build their own inference chip in the car. Yeah, even for that we're okay, we'll still continue working with them. Actually we are working with Tesla and many OEMs who are building using their own inference chip by collaborating with them in the cloud, by providing them, we even try to help optimize their models. Basically again with different OEMs we have different basic collaborations because we still have the simulation computer and the training computing in the infrastructure. We are working with them and for some of the OEMs basically they would like to have a more towards a turnkey solution. We are very happy to work with them as well. In that case we are going to go all the way. We are working like a Tier 1 or Tier 1.5 essentially just go hand by hand. This is our driver AV kind of partners, for example Mercedes. We work very closely with them to define the products they want and then also adapt our driver AV stack to work seamlessly in their vehicle. And actually the engineers from both sides work pretty closely to make it really adapt well into the Mercedes, let's say design DNA and the customer experience they would like to offer. We are not picking winners per se. We try to help OEMs based on their capability at different levels. So as I said, the openness is really important for our basically kind of engagement model with OEMs.

[Neil C. Patel]

One of the reasons I'm so curious about this is you mentioned training models, you mentioned, I think in other interviews that you're doing synthetic data to train autonomy in different ways. I'm very curious about that. It just strikes me looking at the industry, Waymo has this gigantic lead in autonomous miles driven and they're very proud of it. And that's helped make Their cars as successful as they are in the markets they're in. Tesla obviously has a huge number as well because they're training on the actual cars that are being driven. Not every automaker can figure out how to get to a billion autonomous miles driven.

[Xinju Wu]

Right.

[Neil C. Patel]

They're going to have to rely on some third party to get them to at least the status quo, if not beyond that. Feels like Nvidia is sitting there ready to be that third party. Is that a lot of the sell to the automakers that you can just buy our technology at the cell for in whatever open capacity that you want and we will just quickly get you to a competitive state?

[Xinju Wu]

I would say this is the one compelling point based for OEMs to engage with Nvidia in the Hyperion ecosystem, in the DRIVE ecosystem. Because one of the key thing for Hyperion basically again which define the COMPUTE architecture and also the sensor architecture is the data sharing for anybody who engage become a drive partner. Nvidia drive partner. We not only we share data through our kind of existing program which we collect basically millions of hours a day and also basically through the different car programs, basically we are also accumulating that data from different OEMs and then basically we can build a model first of all, basically which can work which are trained with all these data. And also we make sure at least the data basically collected in our different car program is shared with. That's number one. Number two is in the new era, we strongly believe COMPUTE is data as well. As you mentioned, there's a lot of synthetic data and also there's neural reconstruct data which we call neural. This is a very important piece of technology in simulation where we have collected data from the field. But we can use neural reconstruction sometimes to fuzz the data to change the background or change the car trajectory. We can bas basically generate a lot of variants of the same data. And all these data, again they need a compute obviously to generate this kind of millions, tens of millions of data. And we can share with them, with everybody who's engaged in our ecosystem and in this way collectively from all the players that engaged with the DRIVE ecosystem, we can catch up on the data gap, which is very important content.

[Neil C. Patel]

So the synthetic data I think I understand right, you're going to collect a bunch of real world driving examples. You'll put it into a simulator. The simulator will then blur the data. Right. I think the example that I've heard you give is there was a pedestrian that came out and we can just delay the pedestrian and make it make that person come out later and the car will have to react to it as though it's real. You're going to run lots of training against lots of different variations of the same data. That's fascinating to me. Me, I understand why all the car makers would buy into that. Why would they buy into the data sharing? Is it just a recognition that collectively they stand a better chance of catching up? Is it they just don't want to pay the money? Is it, is it cheaper? Like, why would they participate in their competitors in that kind of data sharing arrangement?

[Xinju Wu]

Both are absolutely true. And actually the cost saving is, is, is, is enormous. Basically, you know, data collection, running a fleet of, of huge size. Essentially it's, it's to say, capital spending for anybody who wants to do that. And also it's kind of repetitive as well. If you can find, for example, what we provide in the drive platform or the drive ecosystem, it can save a lot of effort and basically money from our customers.

[Neil C. Patel]

I'm curious about that because the idea that you're going to train stuff and then you're going to have a model in the car and we'll have an AI defined car. The sort of classical approach to self driving was we're going to throw more and more data at the problem and eventually the car will kind of know how to do everything and it will have mapped all the roads on top of everything. Right. So you're going to, you know, I have a, I have a Cadillac EV and the way Super Cruise works is it works on roads that are mapped and eventually, you know, the bet is they'll map more and more roads and more and more things and the car will become more capable. It feels like Nvidia's approach is for the car to be smart enough to do anything with or without the map apps. And that requires a different approach to data collection, a different approach to commute and then obviously a bigger bet on AI. Is that split real? Have you just made that jump? Is that the future of the platform or are you in the middle?

[Xinju Wu]

Well, the approach we take right now for what we call the L2 essentially is Mapless, as you said correctly. So basically the model would definitely need more data and to cover more corner case. And the model is obviously getting bigger, bigger as we speak as well. Basically for this generation, next generation, we are going to use a much bigger model with more parameters and also foundation models will play a big role here and being able to make this model very capable, essentially more data is very critical. But on the other hand, though the trend of using foundation model which is already trained with Internet data that can help coming help as well. That's why I emphasized quite a few times on the connection with the foundation model effort inside Nvidia with the reasoning model and the foundation model. These are the things that we can leverage from let's say the frontier model perspective and the leverage Internet basically kind of scale data to be able to help the vehicle to generate better even without vehicle specific data. So this is one of the, I would say the main direction we are betting on towards let's say higher level of autonomy, especially level four. This is one of the main work thread we are focusing on right now and back to oem. I think being able to leverage basically what we have built upon through our collaborations with the existing base, basically engagement and our massive capability of basically data generation using synthetic data set and the neural reconstruction and also being able to leverage the foundation model capability which are trained from more general data but which will help the model to reason better, to generalize better. These are the things we can offer to our customers. Customers.

[Neil C. Patel]

I feel like I have to ask about safety now. I'm sure it's more complicated than this, but you're talking about a foundation model reasoning through self driving. And all I have in my head is chatgpt apologizing to me because it got it wrong while the car crashes or one of those horrible long latency loops where the model goes off in the wrong direction and realizes it. And then like you can look at the chain of thought and it's like oh, it got it totally wrong and it's, it feels bad, you know, in the way that like Anthropic believes that Claude feels bad. None of that seems compatible with the very real time nature of driving a car. How do you bridge that gap? How? Latency. The, the need to have one of those big models in the background. The sort of reasoning tangents that the models can go on. How is that compatible with driving a car?

[Xinju Wu]

Safety is so important to us and obviously so important for the AV industry. So let me answer your question from our kind of approach in different layers of our offering. So to address safety, this is obviously not new for the auto industry and we have developed very sophisticated, basically even development protocol and also validation protocol to be able to prove this software is safe. All right. That's called ISO 26262 and we actually develop our hardware and operating system OS level software and application level software to the high standard, which is very important, which is very critical to be able to deploy Anything to drive the car. That's number one. And number two is basically we take a slight different approach than some of the, you know, players in this space. We are actually have a redundant stack even for our L2 plus plus or ADAS basically function other than the end to end model, which is basically you have pixel in, you have trajectory out. We also have a classical stack. Classical stack means it's more developed based on this safety standard as we know it with a component. Basically it's a stack with many components. And each component can be verified using this known standard. That's what I refer to as a classical stack. And when you have two stack basically kind of run in parallel, the classical stack is acting like sometimes we call it the big brother, but essentially it's a safety guardrail. Try to verify all the trajectories from the end to end model and use it, use the, let's say known safety standard to verify safe at every frame. So that's a very important concept we have. And not only concept, but the implementation we have in our stack. And we will take this, obviously this will be so critical for higher level autonomy L4. So this is also the foundation of our kind of L4 stack where we have full redundancy not only at the sensor set but at the software architecture. So, so this is I would say the second point I want to make to answer your safety question. And the number three also basically when we develop the model, basically we are also trying to make the model reduce the hallucination as much as we can, right? So the way to do that is really basically through massive validation. We are looking at, we are building basically massive simulation test data as I said, for every model we release. Right now we are looking at in our program right now we are running 5 million basically tests every day. And obviously roughly every day we have 10 iteration of the model, the end to end model of Mayo. So we are doing really massive validation to make sure in all these scenarios you can think of that every tested test scenario the model is generating the right trajectory. So that's also super critical for us. So this is what we do to make sure our product is safe.

[Neil C. Patel]

Let me ask you a really dumb question I'm really curious about. You've talked a lot about the model and how it will operate the car. And yes, there's the classical stack. Is the safety guardrail. Is the model reasoning in language like every other model? Is it sitting there in the background saying I see a stop sign, what do I do? I'd better stop. I'M going to go hit the brakes. The way that any sort of general model reasons in language in the background

[Xinju Wu]

short answer is yes. And in our next generation model which we are going to deploy in the next generation of vehicles because the current generation is on Orin which has more or less limited compute. The next generation is soar based. We will have the model trained with language embedded. So being able to reason through language is very important. And also using can chat with the model, you can ask the model about what he's doing and then you can also ask model to speed up or slow down and make a lane change.

[Neil C. Patel]

For example, as it's literally driving, it's saying to itself I see a car over there, I need to change lanes to get ready for the exit that's coming in a couple miles. And it's doing that in language to operate the car.

[Xinju Wu]

I think it's combination of things. Language is already embedded in the model. But the vision signal is also super important as you know. So it's. I want to say it's multimodal but language is part of it Obviously as you know the model is black box. We don't exactly know basically what it is exactly doing. But you can ask about it and then the model will answer what it's trying to do and you can read.

[Neil C. Patel]

I just have this vision of a chatbot model just like freaking out as it careens down the highway at 55 miles an hour.

[Xinju Wu]

The recent basically GTC I think Jensen did GTC Taiwan released a video that the model is talking about talking constantly explaining what's trying to do. It can be quite annoying as well if you really try to hear everything the model is trying to reason about.

[Neil C. Patel]

What's the latency on that is that I mean obviously you're deploying the systems, it must be working. But is there an attempt to reduce the latency of that? I feel like language is inherently slow compared to what you need to do to drive. I'm not thinking in language when I drive my car.

[Xinju Wu]

That's why I said it's multi model. Right. But reduce the end to end model. End to end latency is super important. Actually that's one of the key advantage of deploying drive the car with a model because if you think about it the old basically stack or the classical stack which has multiple component it usually basically takes multiple hundred millisecond. But with a model, because it's just inference time, it's separated between input which is pixel and trajectory. You can reduce the basically depends on the compute obviously capability you have. But even in the current generation we control it to be within 100 milliseconds. Millisecond, which is pretty fast. And regarding the language reasoning, obviously if you think about it, well, that's human brain, right? But if you think about the language, basically I would say the information rate, it's already abstracted, the information rate is not super high. And we are obviously using the Internet data to train this kind of language based reasoning capability. I think the latency is well under control, let me put it that way. And again, you're not driving the car with language only. That's a key. As I said, usually the reasoning part is, I believe it's slower. Again, we don't know exactly what the model is doing. But the pixel part, that's what drives the basic instantaneous kind of reaction of the vehicle.

[Neil C. Patel]

If you ask Anthropic, they will tell you that Claude has feelings and emotions and he can get scared. Do you think about that? Do you think your models have emotions when they're driving the car?

[Xinju Wu]

We will use the guardrail to make sure it doesn't get too moody.

[Neil C. Patel]

I'm just curious. I mean it. Like you said, we don't know how the models are working. I just, I literally have a vision of the model being like, oh my God, I'm going so fast. But maybe the classical system will cut that down.

[Xinju Wu]

Yeah,

[Neil C. Patel]

we have to take on a short break here. We'll be right back.

[Sponsor Announcer]

Support for the show comes from upwork. If you're looking to grow, you need people who can take your business to the next level level. Upwork is a one stop platform to find, hire and pay expert freelancers across more than 125 categories, launch projects faster and scale support up or down. Without committing to full time headcount. You can browse profiles, review people's past work and get help scoping the role. Upwork also has Business plus which lets you access the top 1% of talent on Upwork. With AI powered shortlistings, you'll get matched to the right freelancer and utter six apps hours. And when it comes to operational stuff like contracts and payments, upwork has a handle on it all. That way you can actually spend more time at the helm. It's free to sign up and posting a job is easy. Visit Upwork.com right now and post your job for free. That is Upwork.com to connect with top talent ready to help your business grow. That's up w O-R-K.com upwork.com

[Neil C. Patel]

support for

[Sponsor Announcer]

the show comes from whatnot. Consistency is key. It's what can take a business side hustle into something that lasts. And with whatnot, you can turn that side hustle into a real business. Whatnot is the largest dedicated live shopping platform. Whether it's beauty, collectibles, electronics, luxury, fashion, even cookies, sellers are building real, thriving businesses. Businesses anyone can sell, whether your business is big, small, or yet to exist. And people selling on whatnot sell 10 times more than on other major marketplaces. That's because you're not just listing products, you're building real connections with buyers. And the company has seen that. Whatnot buyers spend more than an hour a day in the app, and that time is spent more than just browsing. They're engaging, buying, building community, and coming back as a seller. You go live, show off products in real time and turn what you love into real income. We have a colleague who's tried whatnot out and really enjoyed it. Search whatnot W h a t N o t in the app store download and you can start selling right away. Support for the show comes from ServiceNow. AI was supposed to handle the parts of the job you hate. Instead it just describes them, suggests what to do about it them, and then leaves you to do it. That's not help, that's homework. ServiceNow's AI specialists are different.

[Neil C. Patel]

They're not a tool.

[Sponsor Announcer]

Think of them as digital teammates who actually do the work from start to finish. Cases get resolved, requests get processed, loops get closed, and most importantly, no extra work for you. Because when you can truly delegate to AI, you can get back to the work. Only you can do the work. That requires a person with ideas and judgment and, you know, a pulse. To learn how to put AI to work for people, visit servicenow.com.

[Neil C. Patel]

Welcome back. I'm talking with Xinju Wu, Nvidia's head of automotive, about how all these AI systems very literally work. Is all this running locally and the car?

[Xinju Wu]

No, no, no, no. All these validation offline. But the second part, which, with the safety guardrail, when we run two stack in parallel, that's definitely in the car and in the car at every frame. The software in our ADAS ecu, we are comparing basically the trajectory from both the classical stack and the end to end model to make sure the model is outputting a safe basic trajectory.

[Neil C. Patel]

So do the cars require fast connectivity to be autonomous with your approach?

[Xinju Wu]

Not necessarily, but we do require some connectivity to get navigation information and some map information. Most of these are a navigation map. So not only the model side and also the classical stack which we do use some of the navigation map information to help us understand the world better. Essential.

[Neil C. Patel]

I'm only asking because I, I covered the launch of 5G networks in great detail and all of the telecom companies promised me that 5G would enable autonomous cars. And it seems like your approach is the one that will leave lean the most heavily on low latency networks in that way.

[Xinju Wu]

Well, which is that this is not wrong. But on the other hand, basically you know, the car have to, to drive autonomously in completely blind spot as well.

[Neil C. Patel]

Yeah.

[Xinju Wu]

Real time basically low latency, I would say content dependency. Have that dependency in the cloud, at least for the ADAS kind of application. L2 is what we call that which is meant to work everywhere. Building that dependency is not a good idea.

[Neil C. Patel]

When you get to level four, level five, that's when you have the connectivity dependency.

[Xinju Wu]

That's right, yes. Yeah.

[Neil C. Patel]

What happens when you lose the connectivity at level four Autonomy? When you're at level five and you don't have a steering wheel anymore and you lose connectivity, what happens?

[Xinju Wu]

You can think of connectivity is kind of a sensor and again the basic driving capability cannot have huge dependency on that. One of the core concept of developing level for technology is you have a sensor redundancy that's not only for basically gps but also for camera reduction, radar, everything you see. For every single point of failure the car have to be able to drive safely. It's like you suddenly lost your gps. But the car with local perception, it need to be able to get to a safe point and pull over. That's the minimum requirement an L4 system need to have. So this is just the L4 basic principle to be able to develop such a system.

[Neil C. Patel]

I'm very curious about where all of the sensor stacks live in the. How much compute is in the car, how much RAM we need to put in cars at a time of increasing RAM prices. This, all, this all seems like a lot of extra cost to layer into cars which are increasingly getting more expensive and which you know, consumers at least in the United States feel like they're rebelling against in lots of ways. Right. I can look at our own website traffic and I'm like everybody wants to buy a slate truck for $25,000 and it doesn't even have a radio. Right. Like that's just a, that's just a battery and wheels. That, that's that whole car. It doesn't even have paint job. Like we're, we're, we, we're, we're getting rid of Paint jobs on the cars now to keep the cost down. You're talking about a lot of compute in the car, a lot of connectivity, maybe a bunch of RAM to load the models on.

[Xinju Wu]

That's right.

[Neil C. Patel]

How does that play out? Is that push you more into that robo taxi model or do you think people are just going to buy expensive self driving cars?

[Xinju Wu]

Definitely building autonomous car need, you know, a lot of hardware. But the other trend, the hardware cost is I would say is dropping pretty rapidly as, as the technology become more mature. For example, radar. Right. Even in my career basically I have seen radar price probably drop by at least four or five times over 15 years because the volume just getting much bigger and bigger than basically the cost. IC have witnessed basically the drop of both actual camera sensor price drop as well. There's more competitors and the more competition and the competition bring lower price when the volume become bigger. The scale effect is definitely there. You know, right now in adas and all the components are much become much more and more basically mature and to some level commodity. And on the computer side, as you know the computer is growing at such a rapid pace. So we talked about Moore's Law, you know, in the semiconductor industry, you know, sometimes ago, but in the auto basically segment, in the autonomous driving segment, the compute has, the computer need has been growing basically at a really astonishing pace. Roughly we are talking about 10 times every two years. It's insane. With the success of AI and obviously Nvidia, we will be able to provide this kind of massive computer cars at affordable price.

[Neil C. Patel]

But in the cloud or in the car?

[Xinju Wu]

In the car.

[Neil C. Patel]

In the car. I asked you about fighting for training capacity earlier. Do you have to fight for fab capacity to too because those costs are going up for everybody.

[Xinju Wu]

Yes, of course, yes.

[Neil C. Patel]

But then kind of, I'm curious, it's Nvidia's demand that's driving up the cost for everybody. So how do you, how do you go get fab capacity when the other divisions at Nvidia are willing to pay whatever rates are anyone demands?

[Xinju Wu]

Well, the same answer I give you. Right again. You know, I don't know if there's anything I can say more. Right. Because you know we are, we are such a strategic company company and our automotive business is doing well as well. But not at the pace of our data center business is doing obviously. But basically we are strong believer, Jensen himself as well of the AV future. And we are keeping investing basically in this technology and in this future. Not only from allocating external computer, but from fab capacity as well but that's definitely one of the things we are looking into actually. Most likely even the chip price might need to go up essentially because of this intense kind of demand for every chip. Everybody can grab on essentially. But the positive side is basically the technology is really getting. I talked about the chip side and also I talk a little bit about sensor side and we are looking at basically for example, I talked about Hyperion which is basically kind of product ready compute plus sensor kind of platform. So we are looking, we are really trying to balance between the cost and what we can do. We are looking at what we call the, you know, sufficient necessary kind of sensor set to achieve high level of Autonomy. So in Hyperin 10 for example, we really offer two versions. One is the base which is mostly camera 10 camera, 3 radar, no lidar and you know, it's a very cost effective way to build a basically kind of L2 ADAS kind of vehicle. And on the other hand for the high end, what do we call the Hyperion high, we provide basically the sensor set required which have like I think 14 camera and three lidars and basically seven radars essentially to be able to drive, have enough sensor redundancy to be able to drive, you know, L4. We also provide, we you need the ECU redundancy as well. You need two basically kind of our next generation, well actually be more precise current generation soar based kind of computer platform. But just, just imagine basically you have a car really can drive by itself. We believe with this sensor set and this computer basically architecture we can get to that level of autonomy which can basically justify the cost.

[Neil C. Patel]

The minimum sensor set for autonomy feels hotly debated. It's been hotly debated for a long time I think.

[Xinju Wu]

Yeah.

[Neil C. Patel]

Elon Musk saying that he thought lidar was a local maximum ages ago was the beginning of this debate. This debate has not quelled in any way shape or form. Do you think level four requires lidar?

[Xinju Wu]

The short answer is yes. We believe that lidar is the important sensor, you know, to provide the safety and redundancy required for level four autonomy. But on the other hand, hand it's difficult to say it's 100% necessary. We believe this is a very much feasible path based on, As I said, Hyperin 10 high sensor configuration to get to really high level of both urban and highway level 4 capability. On the other hand, theoretically people can prove out with massive mileage essentially to say that lidar may not be necessary, but it will come with odd limitation essentially.

[Neil C. Patel]

Sorry, what's an odd limitation?

[Xinju Wu]

Odd Is basically applicable. Basically domain you can deploy the technology. Obviously we have done quite a bit of analysis on this based on our current understanding and the framework we use to do this analysis. We believe that to deploy this L4 technology in all the odds that our customer benefit from it's much better to have LIDAR as compared to not having it.

[Neil C. Patel]

When you look at where Tesla is with full self driving and their vehicles and their absolute commitment to being a vision based system, do you think that they are currently ahead of you? Do you think they're at parity? Do you think they're behind you?

[Xinju Wu]

So there's two level of answer I guess to this question. And obviously basically for the basic L2 basically technology, Elon is probably ahead of everybody. Essentially he has the vision a long time ago and he has stick to the vision for a long time to be able to and develop and test the technology among massive fleet. Nobody would argue that Elon is ahead of Everybody in the L2 or basically ADAS kind of market. And everybody is playing catch up game essentially and we are very happy actually Elon is so successful. And also obviously Elon is a big customer for us as well for both SpaceX and Tesla in the GPU computer side and we are supporting him and his team to make sure they're successful. And for level four essentially I think it's more open I would say because obviously there's established players that who are proven, who are already basically like Waymo, who are doing basically already taking customers to really experience the L4 kind of experience using the methodologies they use. And Tesla is probably still trying to find the path there. And again we don't try to pick winners but we try to help everybody to be able to develop their technology. And our mission is really try to make make the AV ecosystem get to this vision of every my all everything moves that will be autonomous. This kind of vision becomes a reality.

[Neil C. Patel]

Have you had conversations with Tesla executives about using LiDAR? It seems oddly religious for no reason. Especially if the costs are coming down as you say, at some point if the better technology solutions is right there, it feels like everyone should just use it. Have you had those conversations?

[Xinju Wu]

Well, actually no, not myself. My team definitely has and well, I'm looking forward to have that conversation with them actually I would like to anyway. So as I said, much of this is just basically science and basically reasoning. So it's good to hear their view as well.

[Neil C. Patel]

I want to wrap up by talking about something that maybe is the least in your control. Models are going to Keep getting better. Nvidia is going to keep making chips. Maybe customers are going to keep demanding self driving. That all feels like something you have a handle on. But the auto market, the cutting edge of the auto market is happening in China. I, I think we can just agree on this. US consumers open TikTok and see car influencers talking about BYD vehicles and they complain in the comments that they can't get those cars. I watched a video of a Buick that is in China. It's a Buick EV that you can't get. United States and US customers are furious. The Buick is making better cars in China. They're making here. There's a lot of trade barriers between the United States and China. Nvidia sits in the middle of that fight in all kinds of ways. Whether it's tariffs on imports of car components, whether it's literal blocks on what chips can be sold and where the revenue from those chips go. As you try to push the car market forward, how does, how does the US China trade chaos play into it? Is that something you think about? Is it something that's slowing the industry down? Is it something that you can push through?

[Xinju Wu]

Well, basically. Well, I certainly believe the, you know, policymakers, they have their reasoning and basically rationale to make the policy. You know, as we see right now and as Nvidia again we are open ecosystem player, we still have a lot of customers in China. We try to basically help this. For example, we are still supplying actually in car inference chips because they are still basically below, let's say the threshold of basically what GPU is allowed to sell in the China market. And then basically we are also working with all the Chinese OEMs, actually not all of them obviously, but quite a few of them to help them on the infrastructure side by supplying them basically simulation tools. And we are working with them on open source models, Cosmos and Mayo and then basically on one hand we can help them to get their models better. On the other hand we can also learn from the competition in the China market. Obviously we are also working very closely with the rest of the world basically OEMs and try to supply all Nvidia basically platforms at different layers to different OEMs and help them to be successful as well. So again we don't pick winners and we try to basically work with everybody and the mission is super clear and we try to make AV this vision become reality as, as soon as possible.

[Neil C. Patel]

When you talk about sharing data between OEMs to train the models better and to make them more capable, are there any Regulatory roadblocks or competitive roadblocks between sharing data from Chinese OEMs and American and European OEMs.

[Xinju Wu]

Oh yes, of course. So we have to live with the regional basic. Actually not only China actually other regions have restrictions as well. For example, Europe has certain regulations regarding data. So we are conformed to all the kind of local kind of regulation to make sure we are compliant to all the basic regulations we need to be compliant to at different regions.

[Neil C. Patel]

Does that mean that regional variants of the models have different capabilities or they're better at different things? Because if the input data is different, it seems like maybe the output will be different as well.

[Xinju Wu]

Absolutely. Well, first of all, we try not to basically for the production model, we try not to basically fork it as much as we can, but there will be basically original kind of difference. So the model will behave differently in different regions based on the input. And some of the things are what we call the country coded. So you have to. Obviously the rules are quite different in different regions, like in Europe, as compared to us. Some adaptation is required and some parameters are different as well. Yeah. So it's quite an interesting journey trying to scale the technology into definitely different parts of the world.

[Neil C. Patel]

Do you think that based on the different regulatory approaches, the different data approaches, the different input data, the different configuration of the OEMs and what they're willing to invest in, the different subsidies from the governments, do you think China will get to level four as a mainstream self driving experience first? Because if I had to look at it, I would bet that level 4 self driving will happen in China way before it happens in the United States as a mainstream experience.

[Xinju Wu]

I actually don't think that's true. As you know, basically Waymo is already getting to everybody to L4 experience in at least in certain odds in San Francisco and they're scaling pretty fast. And China is, you know, obviously it's a much more dynamic, competing kind of market and there's quite a few players there. But my experience none of them has got to the maturity of Waymo, at least in Francisco. But again we're trying to help everybody in the ecosystem. So from OEM perspective it's a different competition landscape. But even basically on the OEM side, I think different regions have different kind of. Well, one side is probably the China streets is also much more challenging as compared to the US streets. So you know, to be able to. And the level four, I would sometimes call it a 01 game. You know, either you have it or you don't have it. Actually as of today I think the only one who really have proven that L4 can be safely deployable to every customer without driver in a kind of city, kind of size region without any limitation is still in us, not in China.

[Neil C. Patel]

Yeah, that's wayo. I think wh would would is going to be very flattered to hear them described as a mainstream experience. I will accept that for some subset of people in San Francisco, WAYO is a mainstream experience. I think for the vast majority of Americans it is not yet. And I. That is the big turn. Right. When can a wayo work in the snow? When they're going to deploy them in Chicago? I'm as somebody in Chicago for a long time. Yeah, I'm very curious how that goes in Chicago. New York City. City. Right. The, the question I have is the mainstream experience feels like you just buy a car and just like level two ADAs is kind of a commodity in cars now level four will be a mainstream commodity in cars. You push the button and starts driving itself. How far away do you think we are from that?

[Xinju Wu]

Well, first of all, that's exactly my mission, you know, trying to help the industry to get there. I would say if I need to give a time, I would say five years, less than five years.

[Neil C. Patel]

Well, that is a bold prediction. I think we're going to leave it there because we're at time. You've been really great, Xinzhou. I'm excited to talk to you again. We'll have you back before five years to check in on that prediction. But what should we be looking for next from Nvidia?

[Xinju Wu]

There's quite a few things we are planning. So first of all, by I think end of this year, we are rolling out our technology on the basically ADAS side in all Mercedes vehicles and some other partners as well to all over United States. And also basically, you know, studying for the next few years this technology we're trying to roll out to the rest of the world. And meanwhile, basically we are also working closely with NUR, for example Uber. We announced that in GTC, try to basically roll out our L4 basically kind of service in the next few years. It's super exciting. And on top of that, obviously we are again an ecosystem player. We are working, you know, with all, almost like all OEMs right now. I would say 80% of the mass production OEMs are in Nvidia's Hyperion basically ecosystem for L4. So we are really building with this future with everybody. So this is, hopefully you'll see more exciting announcements from us somewhere down the road.

[Neil C. Patel]

Yeah, well, like I said, we'll have to have you back soon. Thank you so much for being on Decoder.

[Xinju Wu]

Thanks for having me, Nele. It's a very nice chat.

[Neil C. Patel]

I'd like to thank Xinji Wu for taking the time to speak with me and thank you for listening. I hope you enjoyed it. Definitely. Let us know what you thought about this episode or really anything else at all. Drop us a line. You can email us atdecoder the verge.com we really do read all the emails. Or you can hit me up directly on Threads or bluesky. We're also on YouTube. You can watch full episodes at Decoder Pod. It's the same handle on Instagram and TikTok. Check out those platforms, they're a lot of fun. If you like Decoder, please share it with your friends and subscribe over at your podcast. Decoder is a production of the Verge and part of the Vox Media Podcast Network show is produced by Kate Cox and Nick Statt. This episode was edited by Xander Adams. Our editorial director is Kevin McShane and the decoder music is by Breakmaster Cylinder. We'll see you next time.

[Xinju Wu]

Hi, Ryan Reynolds here for Mint Mobile. Are you looking for a beach read this summer? May I suggest your big wireless bill? It's got suspense, mystery, a slightly flat emotional arc, and a shocking twist where you realize you've been overpaying the entire time. Fortunately, the so Mint's story is better. Every plan $15 a month, even unlimited. That's it. Happy ending, zero tears. Give it a [e-mail] Switch upfront payment of $45 for three months, $90 for six months or $180 for a 12 month plan. Required $15 per month equivalent taxes and fees. Extra initial plan term only greater than 50 gigabytes. Me slow when network is busy.

[Sponsor Announcer]

See terms. Your team just added its 67th AI tool and also your 67th security blind spot.

[Neil C. Patel]

The good news?

[Xinju Wu]

The Vanta agent works like a GRC engineer.

[Neil C. Patel]

The background finding every app your team

[Sponsor Announcer]

uses, scoring the risk and drafting fixes for you. Vanta is the platform used by over 16,000 fast moving companies like Ramp, Cursor

[Neil C. Patel]

and Harvey who are shaping the future

[Sponsor Announcer]

with AI and staying ahead of AI risk.

[Neil C. Patel]

Get [e-mail] evening buyer's remorse buy a new car.

[Xinju Wu]

I'll be moving in.

[Sponsor Announcer]

Let's get started.

[Xinju Wu]

Sorry, I think there's been a mistake.

[Neil C. Patel]

I bought it from Carvana.

[Xinju Wu]

You what?

[Neil C. Patel]

Yeah, great price.

[Xinju Wu]

I even have seven days to love

[Neil C. Patel]

it or return it.

[Xinju Wu]

So there's no. No buyer's remorse. More like buyers rejoice. I guess I'll let myself out.

[Sponsor Announcer]

Congratulations.

[Neil C. Patel]

I mean it.

[Xinju Wu]

Buyers rejoice. Buy your car today on Carvana. Limitations and exclusions may apply. See our seven day return [e-mail].