Tesla’s Still Testing Its Vision-Only Robotaxi, Xpeng Just Started Building One

• 7 min read • EV, autonomous, Tesla, SUV, manufacturer

# Tesla’s Still Testing Its Vision-Only Robotaxi, Xpeng Just Started Building One

Tesla’s Still Testing Its Vision-Only Robotaxi, Xpeng Just Started Building One

The race to develop fully autonomous vehicles continues to accelerate, with major players making significant strides—and sometimes setbacks—in their quest to bring driverless technology to the masses. Two notable developments stand out: Tesla’s ongoing testing of its ambitious vision-only robotaxi, and Xpeng’s recent announcement that it has begun building its own autonomous vehicle designed to operate without human intervention. Here’s a comprehensive look at where these giants stand, what their efforts mean for the industry, and what enthusiasts and consumers should watch for in the near future.

The Autonomous Vehicle Landscape: A Quick Overview

Before diving into the specifics of Tesla and Xpeng, it’s important to understand the broader context. Self-driving technology is typically categorized into levels, from Level 0 (no automation) to Level 5 (full automation). Currently, most commercially available systems—like Tesla’s Autopilot and GM’s Super Cruise—fall into Level 2 or Level 3, offering driver assistance rather than full autonomy. The goal for many automakers and tech companies is to achieve Level 4 or 5, where a vehicle can operate entirely on its own, without human intervention.

The Challenges of Vision-Only Autonomy

While many companies initially relied on a combination of lidar, radar, and cameras, Tesla has taken a different route. Its approach emphasizes a “vision-only” system, relying solely on cameras and neural networks to interpret the environment. This strategy simplifies hardware and could reduce costs, but it also presents considerable technical challenges—especially in complex urban environments or adverse weather conditions.

Tesla’s Vision-Only Robotaxi: Progress and Challenges

The Vision-Only Strategy

Tesla’s CEO Elon Musk has long championed the idea that “vision is all you need,” asserting that lidar and radar are unnecessary. Tesla’s Full Self-Driving (FSD) and Autopilot systems are built around a suite of cameras, ultrasonic sensors, and powerful neural networks trained on vast amounts of data collected from Tesla’s fleet.

Current Testing Status

Tesla’s approach has been iterative. The company has been testing its vision-only robotaxi prototype—often referred to internally as the “Tesla Bot” or through its FSD Beta program—in various locations across the United States. As of late 2023, Tesla has not officially launched a commercial robotaxi service, but the company is actively refining its system.

Recent updates include:

  • FSD Beta 11.4 and beyond: Tesla has incrementally improved its neural network capabilities, with some beta testers reporting more natural and reliable autonomous driving in urban environments.
  • Public Road Testing: Tesla’s fleet continues to operate in select cities, including San Francisco, Los Angeles, and Austin, with drivers still required to supervise the vehicle at all times.
  • Limitations: Despite progress, Tesla’s system still occasionally encounters difficulties, such as misinterpreting complex intersections or reacting poorly to unpredictable pedestrian behavior. Musk has publicly acknowledged that full, safe autonomy at Level 4 or 5 remains a work in progress.

Practical Implications for Buyers and Fans

For interested consumers, Tesla’s vision-only approach signals a future where hardware complexity and costs could decrease—potentially lowering the price of autonomous services or vehicles. However, the timeline remains uncertain. Tesla’s aggressive testing underscores a desire to reach full autonomy, but regulatory approval and real-world safety benchmarks are hurdles yet to be cleared.

Things to watch:

  • Continued improvements in neural network performance.
  • Tesla’s ability to demonstrate safe, consistent operation in diverse environments.
  • Regulatory developments that could enable commercial robotaxi services.

Xpeng’s Entry into Vision-Only Autonomous Vehicles

A New Player in the Field

Xpeng Motors, a Chinese EV manufacturer known for its smart, connected vehicles, has recently taken a significant step by starting physical construction on a new autonomous vehicle designed specifically to operate without a human driver—marking a bold move into the fully autonomous robotaxi space.

The Xpeng G9-Based Robotaxi

While Xpeng has produced several models like the G3 compact SUV and the G9 flagship SUV, the company’s current focus is on developing a dedicated autonomous platform. Reports indicate that Xpeng has begun assembling prototype units based on the G9 architecture, equipped with advanced sensors and AI systems suited for Level 4 autonomy.

Technical Specs and Features

Although official technical details are still emerging, industry sources suggest:

  • Sensor Suite: The vehicle will feature a combination of high-resolution cameras, lidar (possibly solid-state lidar), and radar—though Xpeng emphasizes a more sensor-fusion approach compared to Tesla’s vision-only method.
  • Computing Hardware: Xpeng is reportedly deploying custom AI chips capable of real-time processing of vast sensor data and neural networks.
  • Software Platform: The company is leveraging its Xpilot 3.0 software stack, which already offers driver-assist features, and will evolve into a fully autonomous system.

Building, Not Just Testing

Unlike Tesla’s incremental testing in the field, Xpeng’s recent announcement indicates that it has moved beyond simulation and is actively manufacturing prototypes for real-world testing and eventual commercial deployment. The company aims to launch pilot programs in Chinese cities within the next year.

Strategic Significance

Xpeng’s move signals a shift in the industry, from mere software and hardware development to actual vehicle assembly dedicated solely to autonomous operation. It also reflects China’s broader push to lead in autonomous mobility, with government support and an expanding urban ecosystem for robotaxis.

Is Vision-Only the Future?

Tesla’s unwavering focus on vision-only autonomy represents a bold vision (pun intended) that, if successful, could simplify autonomous vehicle design and reduce costs. However, many experts remain cautious, citing the complexity of real-world environments where lidar and radar continue to provide valuable redundancy and reliability.

Xpeng’s approach, integrating multiple sensor types, suggests that most manufacturers favor sensor fusion as the most practical path toward full autonomy—at least in the near term.

Who Will Lead the Next Phase?

Tesla’s large data fleet and aggressive testing could give it an advantage in refining vision-only systems, but regulatory hurdles and safety concerns remain. Meanwhile, traditional automakers and Chinese tech firms like Xpeng, NIO, and BYD are adopting more sensor-diverse strategies, sometimes combining lidar, radar, and cameras.

What Buyers and Enthusiasts Should Watch

  • Regulatory approvals: Autonomous taxi services require clear safety certifications—watch for pilot programs and legal frameworks emerging in different regions.
  • Technical milestones: Improvements in AI perception, decision-making, and safety redundancy will be key indicators of progress.
  • Market offerings: While fully autonomous vehicles are not yet mainstream, partnerships and pilot programs could lead to early robotaxi services in select cities within the next few years.

Concluding Thoughts: A Dynamic and Uncertain Future

The development of true autonomous robotaxis remains one of the most exciting—and challenging—frontiers in the automotive industry. Tesla’s ongoing efforts with its vision-only system exemplify a disruptive approach that could lower costs and simplify hardware, but faces significant hurdles before commercialization.

In contrast, Xpeng’s recent move to build a dedicated autonomous vehicle indicates a more cautious, sensor-diverse strategy that might better navigate current technical and regulatory landscapes.

For auto enthusiasts and consumers, the key takeaway is to stay informed about these technological advances, as they could reshape urban mobility, car ownership, and even the broader automotive ecosystem in the coming decade.

Things to Watch:

  • The progression of Tesla’s FSD Beta and real-world safety performance.
  • Xpeng’s pilot programs and how their vehicles perform in urban environments.
  • Regulatory developments that could accelerate or hinder autonomous ride-hailing services.
  • Innovations in sensor technology and AI that could tip the balance toward one approach or another.

As the race unfolds, one thing is clear: the future of autonomous robotaxis is shaping up to be as innovative and unpredictable as the vehicles themselves.