TLDRs
- XPeng begins robotaxi production using in-house AI chips for autonomy.
- Company shifts away from Nvidia to proprietary Turing chip architecture.
- Level 4 autonomous pilot operations planned for 2026 rollout phase.
- XPeng aims to remove safety officers by early 2027 deployment.
Guangzhou-based electric vehicle maker XPeng has officially entered a new phase in its autonomous driving ambitions, announcing the start of mass production for its robotaxi fleet powered by in-house Turing AI chips. The move marks a major strategic shift as the company accelerates its push into Level 4 autonomous mobility, where vehicles can operate without human intervention in designated environments.
The robotaxi rollout places XPeng in direct competition with global players racing toward fully autonomous ride-hailing services, including Tesla and leading Chinese EV peers. Each robotaxi is equipped with four of XPeng’s proprietary chips, designed to enhance real-time processing and decision-making capabilities.
Shift to In-House Silicon
XPeng’s decision to rely on its own Turing AI chips follows earlier plans that involved external semiconductor partners. The company initially explored using Nvidia’s next-generation Thor processor but ultimately abandoned the approach due to development delays and performance concerns. This shift came at a cost, including the termination of a chip partnership with Marvell valued at over US$100 million.
By bringing chip development in-house, XPeng has gained tighter control over both hardware and software integration. The Turing chip architecture is optimized for its vision-based autonomous driving system, enabling improved perception performance in complex driving conditions such as heavy rain, glare, and low visibility.
Level 4 Autonomy Ambitions
XPeng’s robotaxis are being developed under SAE Level 4 autonomy standards, meaning they are capable of fully autonomous operation within predefined geographic zones. The company has outlined a phased rollout strategy, with pilot operations expected in the second half of 2026.
A significant milestone is planned for early 2027, when XPeng intends to remove on-site safety officers from its robotaxi operations. If achieved, this would mark a major leap in commercial autonomous deployment and signal growing confidence in its AI stack’s reliability.
Expanding AI Ecosystem Strategy
Beyond transportation, XPeng is positioning its autonomous driving technology as part of a broader “physical AI” ecosystem. The same Vision-Language-Action (VLA) 2.0 foundation model used in its robotaxi program also powers other advanced projects, including its humanoid robot IRON and experimental flying-car initiatives.
This integrated approach allows XPeng to unify development across multiple AI-driven platforms, potentially accelerating innovation while reducing duplication of engineering effort. It also strengthens its long-term goal of building a full-stack AI mobility ecosystem.
Reducing Dependence on Foreign Chips
XPeng’s pivot toward in-house semiconductor design reflects a broader trend among Chinese automakers seeking to reduce reliance on U.S. technology suppliers. Companies like Nio and Li Auto are also investing in proprietary chip development as geopolitical tensions and export restrictions reshape global supply chains.
By designing its own silicon, XPeng gains strategic autonomy over key components of its autonomous driving system. This reduces exposure to external supply risks while enabling faster iteration cycles for AI model updates and hardware improvements.
As the global race toward autonomous transportation intensifies, XPeng’s robotaxi production signals not only technological progress but also a broader reshaping of the EV industry’s supply chain and competitive landscape.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
