Finding the product-market fit for any new technology always involves trial and error: autonomous driving has so far has swallowed an estimated $50 billion in investment with relatively little to show in return. In 2021 we will see autonomous driving finally find its fit – and it’s not in personal travel.
Billed as the first everyday consumer application of AI, self-driving has simultaneously inspired its champions and proved a worry for its investors. Aiming high in terms of functionality has delivered plenty of exciting business cases, but the science and engineering problems they entail are not easily soluble. Aiming low has not delivered business cases that make sense and deliver a return. This conundrum has sent technology companies in a number of exciting directions, whereas automotive firms have ended up favouring simpler driving-assistance features that they can monetise more easily.
In 2021, we’ll see the first fruits of precise product targeting by well-funded tech firms.
In the trucking sector, companies such as San Diego-based TuSimple, Alphabet’s Waymo and Aurora in San Francisco will start to sell their “drivers” to logistics operators. These will carry cargo, fully autonomously, from hub to hub over US highways, promising real savings in cost and time. At the other end of deliveries, robotics company Nuro, based in Mountain View, will move from last-mile trials with US retailers Kroger and CVS, into fully fledged, paid-for autonomous delivery services. We’ll see similar use cases in China using Baidu’s self-driving software.
There are two reasons why these applications of self-driving technology will find product-market fit.
The first is the pandemic, which has squeezed ten years of e-commerce growth into a few months. Consumer demand for home deliveries, groceries and other errands has been established and business cases predicated on home deliveries – and moving cargo around to support them – have been transformed.
The second is that the technology now exists to speed up the development and safety assurance of autonomous-driving technology. Virtual development and testing won’t fully replace real-world testing, but, in 2021, it will make it easier for developers to find and fix bugs, improve performance and explore how software interacts with complex conditions such as environment, road layouts and human behaviours. My own company, Five, uses cloud-based software that can be hyper-scaled, making it possible to advocate for the safety of specific self-driving applications.
In 2020, the new services that launch will attract capital from across the autonomous-driving sector, and will spur automotive companies to move faster towards “full” self-driving applications to keep up. We will see a combination of proven technology and real-world market fit take a step forward into environments where self-driving vehicles are finally ubiquitous.
Stan Boland is CEO and co-founder of Five