Wayve's Approach to Autonomous Driving and AI Integration

Wayve revolutionizes autonomous driving by focusing on AI's ability to learn in diverse environments.

### Wayve's Innovative Approach to Autonomous Driving
* 🚗 Wayve's strategy emphasizes a focus on the **foundational model**, or the "**robotic brain**," which enables cars to **learn how to drive themselves** without being tied to specific hardware or cores.
* 🌍 The decision to test in California, particularly in San Francisco, is part of Wayve's ambition to **tap into the diverse U.S. market**, enhancing their capacity for **generalization across different driving environments**.
* 🛣 This approach supports a spectrum of automation, from **driver assistance to full autonomy**, utilizing various vehicle types and training data from **over 15 countries**.
* 🤖 Wayve is developing an "**embodied AI**" that integrates artificial intelligence into the physical world, capable of **understanding and navigating diverse environments** with different sensors.
* 📈 The initial model will be a software-based driver assistance system, which will evolve through **real-world exposure** to improve the overall quality and pave the way for a **fully autonomous future**.
* 💰 With over **$200 million raised** from prominent investors like SoftBank, Nvidia, and Uber, Wayve is focused on maintaining a **disciplined approach** while efficiently utilizing funds to support their **end-to-end autonomous driving model**.
* 🔍 The company aims to leverage partnerships with key players in the tech industry to **enhance data efficiency and fleet learning**, which are vital for **advancing their autonomous driving capabilities**.

What's so interesting about Wayve is no focus on the hardware, no focus on the cores, just the foundation model, the robotic brain that helps a car learn how to drive itself. Let's stop by asking why you felt you needed to bring your testing to California, to the streets of San Francisco. Hey. Well, we're excited to bring a fresh approach to the US market. I think in general, being able to operate an air and powder test in such a diverse market as the US is going to be awesome for us. We really want to push the boundaries of generalization. The neat thing about this approach is that it can allow all ranges of automation, from driver assistance up to full autonomy, with different types of vehicles in different driving environments with different coaches. And, you know, look, we're now training from data from over 15 countries. But to be able to actually have a testing fleet on the ground and show

that I can generalize and drive in a new environment like the US and California. This is a really exciting moment for our business. Alexa, out the pathway from here. You're going to start with basically level two plus kind of ADAS functionality. So you don't need a DMV permit, but what are the next stages? How will you expand the testing and what's your ultimate goal for the California market? You know, fundamentally, Wayve is building an embodied AI. This is an AI that, you know, brings it out of the Internet and into the physical world to be able to drive and operate vehicles. And this is an AI that can understand the physical world, can drive with different types of sensors. And what's really exciting for us is that while we want to deploy it in a range of automotive products, the latest vehicles that have been produced by the industry have used surround cameras and radar.

sensors. And I actually are the perfect deployment environment to reap the benefits of this. Right. So the first model we want to bring is a software model where we can see this air deployed in vehicles, deployed as driver assistance and use that experience and exposure to improve the quality of the system over time and ultimately bring us to this autonomous future we all dream of. You've raised more than ,000,000,000 from the likes of SoftBank and Nvidia, Uber recently topped up the round. How quickly burning through that and where's the money going? Yeah. Look, we've been a fairly disciplined company for the last seven years. Seven years ago, we. We had the thesis of going all in on an end-to-end approach for autonomous driving. And it's something that we've been building out and pioneering some really interesting approaches that allow us to be more data efficient and to learn from simulation and generative AI.

And so while we're seeing a lot of expansion and growth in the amount of data we're training on, we want to continue to make a lean and cost-effective solution possible to come to market. And of course, the partners that we've got in the US that also make us really interesting for us to be able to operate in that in the same market, closer to our partners like Uber and Nvidia and Microsoft that have been really key parts of this journey to power compute and the fleet learning of these systems is going to be so critical.

Wayve Begins US Test Rollout

Welcome to the Bloomberg Technology YouTube channel. Subscribe for original series and mini-docs on game-changing technologies, profiles on innovators, live conversations with experts, and breaking technology news & analysis from Bloomberg TV.

Bloomberg Technology

Wayve's Unique Approach to Autonomous Driving

Wayve focuses on a foundation model for its robotic brain, aiming to teach cars how to drive themselves without emphasizing hardware.

The company is excited to test its approach in the diverse US market, particularly in San Francisco, to push the boundaries of generalization in automation.

With data from over 15 countries, Wayve aims to deploy its AI in various automotive products, using different sensors to understand and navigate the physical world.

The company is deploying a software model in vehicles for driver assistance, aiming to improve the system's quality over time and achieve an autonomous future.

They have raised over $1 billion from investors like SoftBank, Nvidia, and Uber.

The company has been disciplined in its spending over the past seven years, focusing on an end-to-end approach for autonomous driving and using data efficiency, simulation, and generative AI for learning.

Expanding Data Training with Strategic Partnerships

The focus is on expanding data training while maintaining a lean and cost-effective solution.

Key partners in the US, including Uber, Nvidia, and Microsoft, are crucial for operating in the same market and powering compute.

Fleet learning of these systems is emphasized as a critical aspect of the journey.