NVIDIA Locks Hyundai, BYD Into Autonomous Chip Ecosystem as Auto Competition Narrows
Fresh partnerships with Korea's Hyundai Motor Group and China's BYD cement NVIDIA's position in the $191 billion autonomous driving market while automotive revenue remains under 2% of total sales.
NVIDIA announced partnerships with Hyundai Motor, BYD, Nissan, Geely, and Isuzu for its DRIVE Hyperion Level 4 autonomous platform at its GTC conference Monday, extending its reach into both Korean and Chinese electric vehicle markets as the autonomous chip race consolidates around three suppliers.
The agreements, announced at GTC 2026 in San Jose, bring Hyundai Motor, Nissan Motor, Isuzu, and Chinese automakers BYD and Geely into Nvidia’s robotaxi-ready platform. The timing underscores automotive’s role as a diversification hedge: while NVIDIA generated $2.35 billion in automotive revenue for fiscal 2026, the segment represents just 1.09% of the company’s $215.9 billion total revenue—a rounding error compared to the data center business that delivered $193.74 billion in the same period.
Hyundai Deploys Blackwell Infrastructure Across Manufacturing
Hyundai Motor Group’s agreement extends beyond in-vehicle chips. The Korean conglomerate will build an NVIDIA AI supercomputer to accelerate model training, validation and deployment for in-vehicle AI, autonomous driving, smart factories and robotics, according to NVIDIA. The partnership includes NVIDIA Blackwell-based AI infrastructure to power everything from vehicle software to factory automation.
Hyundai has worked with NVIDIA since 2015, but Monday’s announcement formalizes the deployment of DRIVE Hyperion platform for an integrated autonomous architecture scalable from Level 2 through Level 4. Hyundai’s autonomous vehicle joint venture Motional will advance discussions with NVIDIA to leverage new technologies in support of further advancements to its Level 4 robotaxi capabilities.
“Together with Hyundai Motor Group — Korea’s industrial powerhouse and one of the world’s top mobility solutions providers — we’re building intelligent cars and factories that will shape the future of the multitrillion-dollar mobility industry.”
— Jensen Huang, CEO, NVIDIA
BYD Deal Extends NVIDIA Reach Into China Amid Export Restrictions
BYD’s inclusion is geopolitically significant. The world’s largest electric vehicle manufacturer is expanding its collaboration with NVIDIA from the car to the cloud, and plans to use NVIDIA’s AI infrastructure for cloud-based AI development and training technologies, per NVIDIA. BYD had already committed to DRIVE Thor for next-generation vehicles; Monday’s announcement expands that to the full DRIVE Hyperion stack for Level 4 capability.
The partnership comes as U.S. export controls restrict NVIDIA’s advanced AI chips to China. While NVIDIA’s H20 data center chip faced $4 billion sequential drag in Q2 FY26 with zero shipments to China, automotive chips remain unrestricted. There is no indication so far that the Chinese government is looking to ban this NVIDIA system, and BYD has not been told to stop using any NVIDIA products, according to CNBC.
BYD’s 2025 production reached 4.5 million passenger vehicles, with the company targeting 1.3 million overseas shipments in 2026. NVIDIA’s chips power the company’s intelligent driving features as it pushes into European and Latin American markets.
NVIDIA held about 52% share of the global navigate-on-autopilot market and cooperation agreements with 25 automotive OEMs globally as of mid-2024. In China specifically, NVIDIA topped the computing solutions market with 53% share in the first half of 2023, followed by Horizon Robotics at 31%.
Supply Chain Consolidation Around Three Platforms
The new partnerships reinforce a market structure where Mobileye, NVIDIA, Qualcomm, Horizon Robotics and Huawei will achieve a 78% market share by 2035 up from 69% in 2025, per Counterpoint Research. NVIDIA’s advantage lies in its full-stack approach: DRIVE Hyperion is part of an end-to-end AV platform that includes data center training, large-scale simulations and in-vehicle computing.
Existing DRIVE Hyperion customers include Aurora, Nuro, Sony Group, Uber Technologies, Stellantis and Lucid Group. Monday’s announcement adds five major OEMs with combined annual production in the tens of millions. The partnerships also include Uber’s robotaxi rollout, with NVIDIA full-stack robotaxis launching across 28 markets by 2028, beginning with Los Angeles and San Francisco in the first half of 2027.
| Supplier | Key Platform | Primary Market | Notable Customers |
|---|---|---|---|
| NVIDIA | DRIVE Hyperion/Thor | Global L2-L4 | BYD, Hyundai, Mercedes, Volvo |
| Mobileye | EyeQ6/EyeQ Ultra | Mass-market ADAS | Ford, VW, Nissan, Geely |
| Qualcomm | Snapdragon Ride | Cost-effective L2+ | BMW, Hyundai Mobis, Mahindra |
| Horizon Robotics | Journey 5/6 | China domestic | XPeng, Li Auto, SAIC |
Competitive Pressure on Alternative Stacks
The consolidation around NVIDIA’s platform creates switching costs that disadvantage rivals. Tesla develops its own FSD chip, achieving vertical integration but limiting economies of scale. Mobileye pursues vertical integration, bundling EyeQ SoCs with proprietary perception software, but its closed ecosystem alienates OEMs seeking flexible software development. Qualcomm emphasizes cost-effectiveness and power efficiency over absolute performance, targeting mass-market Level 2/2+ deployments.
Chinese competitors face different constraints. While Horizon’s Journey 5 powers XPeng, Li Auto, and SAIC vehicles, geopolitical considerations drive Chinese OEMs toward indigenous solutions. Huawei’s ADS 2.0 platform, powered by the MDC 610 chipset, competes directly with NVIDIA in the domestic market but lacks global reach.
For automakers, the calculation hinges on control versus capability. NVIDIA allows customers to customize the product and develop their own software, while Mobileye’s product has been a ‘black box’—when you use Mobileye, you are using their software, noted Ali Kani, NVIDIA’s VP of automotive, in an interview with Globes.
- NVIDIA added five major OEMs to DRIVE Hyperion, including Hyundai Motor Group and BYD, expanding presence in Korea and China
- Automotive remains sub-2% of NVIDIA revenue but grew 39% year-over-year to $2.35 billion in fiscal 2026
- Top five chip suppliers (NVIDIA, Mobileye, Qualcomm, Horizon, Huawei) will control 78% of autonomous chip market by 2035
- BYD partnership extends NVIDIA reach into China’s EV market despite export controls on advanced AI chips
- Hyundai’s Blackwell-based infrastructure deployment shows NVIDIA expanding beyond in-vehicle chips to full manufacturing AI
What to Watch
Track whether NVIDIA’s automotive design wins translate to material revenue growth beyond 2027, when Uber robotaxis begin deployment and Thor-equipped consumer vehicles reach production. The current $2.35 billion run rate represents less than a third of the roughly $5 billion in automotive revenue NVIDIA targeted for FY 2026—the company is behind internal projections.
Monitor Chinese regulatory decisions on automotive chip sourcing. While BYD faces no current restrictions, Beijing has reportedly been discouraging local tech firms from buying NVIDIA’s AI chips. Extension of that policy to automotive Semiconductors would eliminate NVIDIA’s fastest-growing customer base overnight. BYD executive Stella Li confirmed the company has a backup plan if supply is cut.
Watch for Hyundai’s chip architecture decisions. The automaker is in serious discussions with Samsung Electronics to mass-produce autonomous driving chips using the 5-nanometer SF5A process, per KED Global. If Hyundai shifts toward in-house silicon manufactured by Samsung—as Tesla did with its FSD chip—NVIDIA’s partnership becomes a training-wheels arrangement rather than a long-term dependency.
Finally, track autonomous chip ASPs as competition intensifies. Industry executives say NVIDIA is considered expensive compared to Qualcomm and Mobileye. If pricing pressure forces NVIDIA to cut margins to defend share, automotive’s contribution to overall profitability will remain negligible regardless of unit volume growth.