Earlier this week, Waymo announced that anyone in Los Angeles can take fully autonomous rides 24/7 with Waymo One. The company began commercial operations in LA earlier this year, with nearly 300,000 people joining our waitlist, and riders have taken hundreds of thousands of paid trips across the city and highly rated them at 4.7/5 stars on average.

“Now is an exciting time to welcome everyone in Los Angeles along for the ride,” said Tekedra Mawakana, Co-CEO of Waymo. “Our service has matured quickly and our riders are embracing the many benefits of fully autonomous driving. We’re so grateful to all of our first riders in LA, and we can’t wait to serve more riders soon.”

To ride in LA, and in other focus cities of San Francisco and Phoenix, riders can download the Waymo One app and ride.

 

Over 100,000 paid weekly trips

Waymo’s continued leadership in autonomous vehicle technology is being recognized with more support announced by company Co-CEOs Mawakana and Dmitri Dolgov in October. The company closed an oversubscribed funding round of $5.6 billion led by parent Alphabet, with continued participation from Andreessen Horowitz, Fidelity, Perry Creek, Silver Lake, Tiger Global, and T. Rowe Price.

With the latest investment, the company will be able to welcome more riders into the Waymo One ride-hailing service. It will also continue advancing the Waymo Driver AI-powered autonomous driving system to support a variety of business applications over time.

“While AI is only just beginning to capture the public imagination, Waymo has been working to bring its endless possibilities to the world of physical transportation for years,” said Egon Durban, Co-CEO of Silver Lake. “The Waymo Driver leads in earning trust by safely actualizing the value and potential of AI through cutting-edge research, practical solutions, and a vastly expanding scope and scale of real-world experience.”

This year brought continued commercial progress toward Waymo’s mission to offer the world’s most trusted driver. The company expanded Waymo One service areas in San Francisco, Los Angeles, and Phoenix including providing curbside service at Sky Harbor International Airport. Through an expanded partnership with Uber in Austin and Atlanta, it will begin serving riders in those cities in 2025. To create a more useful experience, the company began fully autonomous freeway operations in Phoenix and San Francisco.

These efforts have enabled it to provide over 100,000 paid weekly trips—a tenfold increase from last year.

 

Hyundai strategic partnership

One of Waymo’s most significant announcements of the year came in October when the company said it had entered into a multi-year strategic partnership with Hyundai Motor Co. In the first phase, the companies will integrate the sixth-generation Waymo Driver into Hyundai’s all-electric Ioniq 5 SUV, which will be added to the Waymo One fleet over time.

The vehicles destined for the Waymo fleet will be assembled at the new Hyundai Motor Group Metaplant America EV manufacturing facility in Georgia and then integrated with Waymo’s autonomous technology. The plan is to produce an Ioniq 5s fleet equipped with Waymo’s technology in significant volume over multiple years to support Waymo One’s growing scale.

“Hyundai’s focus on sustainability and strong electric vehicle roadmap makes them a great partner for us as we bring our fully autonomous service to more riders in more places,” said Mawakana.

Initial on-road testing will begin by late 2025 and become available to Waymo One riders in the years to follow. The Ioniq 5s for Waymo will have specific autonomous-ready modifications like redundant hardware and power doors.

“The team at our new manufacturing facility is ready to allocate a significant number of vehicles for the Waymo One fleet as it continues to expand,” said José Muñoz, President and Global COO of Hyundai Motor Co., and President and CEO of Hyundai Motor North America. “Importantly, this is the first step in the partnership between the two companies and we are actively exploring additional opportunities for collaboration.”

Hyundai recently announced the launch of an autonomous vehicle foundry business to provide global autonomous driving companies with vehicles capable of implementing SAE Level 4 or higher autonomous driving technology.

“There is no better partner for our first agreement in this initiative than industry leader Waymo,” said Chang Song, President and Head of Hyundai Motor Group’s Advanced Vehicle Platform Division.

No word on how HMG’s Motional unit fits into its AV plan.

 

Sixth-generation Driver

As ridership has grown, Waymo has focused on expanding the safety and mobility benefits of its driver to more places while enhancing operational capabilities. To accomplish its goals, the company recently introduced the sixth-generation Waymo Driver optimized for cost and enhanced capabilities. It is also testing in more complex environments through road trips to cities like Buffalo, NY, and Washington, D.C.

In August, Satish Jeyachandran, Waymo’s Vice President of Engineering, revealed how the company’s approach of designing both hardware and software from the ground up has been crucial to its success and continues to pay off with the latest generation. It has significantly reduced the cost of the sixth-generation system while delivering more resolution, range, and compute power, and enabling more capabilities.

Since redundancies are essential in an autonomous driving system to provide safe backup functions for assured reliability, the Waymo Driver has a surround view of the world from three complementary sensing modalities. An enhanced camera-radar surround view and a more capable system of lidars enable the new sensor suite to safely navigate an even larger set of road conditions.

Advancements in sensor technology and strategic placement have enabled engineers to reduce the number of sensors while maintaining safety-critical redundancies. They can adjust sensing components to match the specific conditions of each operating environment, like adjusting sensor cleaning for vehicles in colder climates.

With 13 camera, four lidar, and six radar sensors, and an array of external audio receivers, the new sensor suite is optimized for greater performance at a significantly reduced cost. It provides the Waymo Driver with overlapping fields of view all around the vehicle at up to 500 m (1640 ft) away, day and night, and in a range of weather conditions like extreme heat, fog, rain, and hail.

Testing in some of the company’s newer cities has deepened engineers’ understanding of winter weather’s impact on technology and operations, which has been applied to the new system. Since vehicles are exposed to the elements for long periods without manual intervention, they implemented preventive measures for each sensor to maintain a clear view of its surroundings.

The six-generation sensor suite has already undergone thousands of miles of real-world driving experience and millions more in simulation. The Waymo Driver learns from the collective experiences gathered across the fleet including from previous hardware generations. The system’s performance in simulation shows promising indications that the company is on track to begin operating without a human behind the wheel in about half the time.

 

End-to-end multimodal model research

Waymo has been at the forefront of AI and ML in autonomous driving and continuously contributing to advancing research in the field for over 15 years. In late October, the company shared its latest research paper on its end-to-end multimodal model for autonomous driving (EMMA).

Powered by Gemini, a multimodal large language model developed by Google, EMMA employs a unified, end-to-end trained model to generate future trajectories for AVs directly from sensor data. Trained and fine-tuned specifically for autonomous driving, EMMA leverages Gemini’s extensive world knowledge to better understand complex scenarios on the road.

Notably, EMMA demonstrates positive task transfer across several key autonomous driving scenarios: training it jointly on planner trajectory prediction, object detection, and road graph understanding leads to improved performance compared to training individual models for each task. This suggests a promising avenue of future research, where even more core autonomous driving tasks could be combined in a similar, scaled-up setup.

“EMMA is research that demonstrates the power and relevance of multimodal models for autonomous driving,” said Drago Anguelov, Waymo VP and Head of Research. “We are excited to continue exploring how multimodal methods and components can contribute towards building an even more generalizable and adaptable driving stack.”

One of the most promising aspects of EMMA is its ability to improve through co-training. A single co-trained EMMA can jointly produce the outputs for multiple tasks while matching or even surpassing the performance of individually trained models, highlighting its potential as a generalist model for many autonomous driving applications.

While EMMA shows great promise, its current limitations in processing long-term video sequences restrict its ability to reason about real-time driving scenarios; long-term memory would be crucial in enabling EMMA to anticipate and respond in complex evolving situations. Other key challenges to ensure safe driving behavior include EMMA not leveraging lidar and radar inputs, which requires the fusion of more sophisticated 3D sensing encoders; the challenge of efficient simulation methods for evaluation; the need for optimized model inference time; and verification of intermediate decision-making steps.

Despite the challenges of EMMA as a standalone model for driving, the research work highlights the benefits of enhancing AV system performance and generalizability with multimodal techniques.