Xpeng’s AI-Powered Future: L4 Autonomy by 2026

Xpeng is gearing up to push the limits of smart in-car technology and autonomous driving as it enters its next phase of growth.

The company’s in-car system promises an intuitive experience, allowing users to give several voice commands in a row — the AI will handle them one by one. Right now, the voice assistant supports five languages, with more coming soon as Xpeng prepares for international markets.

The company admits its early focus was mainly on Chinese, but now it’s working to improve support for English and other global languages, though fine-tuning for slang and regional expressions will take extra time.

Xpeng’s smart system also gives drivers real-time weather updates and the ability to set favourite apps on a shortcut bar for easy access. The AI can chat naturally with users and handle questions beyond driving — like checking public holidays, identifying other cars, or offering travel advice.

User preferences, from music choices to driving settings, are saved in the cloud. This makes it easy to transfer personal settings to any Xpeng vehicle, allowing a consistent experience for every driver.

Looking ahead, Xpeng is building a network where vehicles share real-time information with each other. If one car spots a traffic jam or a pothole, other cars will adjust their route or even tweak their suspension automatically to adapt to road conditions ahead, improving both safety and comfort for everyone.

Powering this future is Xpeng’s new in-house AI chip, called “Turing.” Production will begin in the second quarter of the year. The chip was designed for the high demands of autonomous driving and AI-powered features, and it’s expected to deliver performance close to Nvidia’s Thor chip, handling around 700 TOPS (Tera Operations Per Second).

Turing uses two specially designed neural processing units and is built to run large AI models directly in the car — even handling up to 30 billion parameters. This boosts the car’s ability to understand its surroundings and make decisions, even in tough conditions like heavy rain or at night. Added safety systems, such as a dedicated safety chip and advanced image processors, help keep everything running smoothly.

The timing of Xpeng’s chip launch is important, as Nvidia’s Thor chip has reportedly faced delays and might not deliver its expected performance to automakers anytime soon. Xpeng’s Turing chip helps the company avoid supply chain risks and gives it more control over future development.

Xpeng’s AI plans go beyond hardware. The company is training huge AI models using its cloud computing platform and later compressing these models so they can run efficiently in the car.

This setup will allow Xpeng to handle more complex driving scenarios as it moves closer to launching Level 3 autonomous driving later this year, and it’s already working toward Level 4 by 2026.

Xpeng’s push into AI and custom-built chips shows its commitment to delivering smarter, safer, and more adaptive vehicles, keeping the brand competitive in the fast-moving world of smart mobility.

Source: CNC