COMMENTARY: Autonomous driving technology — why I’m still a believer

PALO ALTO, Calif.  - 

As a long-time car enthusiast, I can’t help but get excited by all the attention being lavished on the automotive industry right now.

There are times as a venture capitalist, however, when I wonder if the spike in financings and associated high valuations are signs that we are in a bubble that’s about to burst.

And I’m not alone. Wired recently published a with the cautionary headline: “After Peak Hype, Self-Driving Cars Enter the Trough of Disillusionment.”

The piece, which I read on my way to (whose auto-glitz is beginning to give the Detroit Auto Show a run for its money), posits that we have moved into the “hard phase” of making autonomous driving real. For the most part, I agree. But it’s because of the very real challenges that are highlighted that make it worthwhile to be in the space. 

The technology demands with Level 4 and 5 automation, in which vehicles will need to sense, interpret, judge and make critical decisions about dynamic and static objects on the road substantially better than humans, are massive.

The perception-and-decision process for autonomous vehicles in unbounded environments is one of the most challenging artificial intelligence problems in existence.  Autonomy software and hardware stacks must function as mission-critical systems, with millisecond-speed and ultra-high accuracy. Development of such “no-room-for-error” technology requires not only great foundational ideas but also extensive trial-and error in both real and virtual worlds.  

Because autonomous driving comes with non-negligible regulatory and policy implications, the pressure to deliver the right solutions is that much higher.  This mix of inherently complex technology and high performance thresholds is, from a venture perspective, a perfect storm. 

In such an environment, the strongest startup companies will find room to run away from the pack, create defensible barriers and have a shot at capturing big pieces of giant markets.

But even though it seems like we’ve been talking about it forever, autonomous driving technology is still in its infancy, reminiscent to where the computing industry was in the ‘60s and ‘70s.  Just as the evolution of computing was fertile ground for startups to keep fueling the innovation engine from mainframes to PCs to the Internet to the Cloud, so is true for the evolution of personal transportation.

As with nearly every tech revolution in the last few decades, startups have played an essential role in transforming what seemed in early days to be abstract visions, to reality. While the incumbents in the automotive space — vehicle manufacturers and systems suppliers — are jumping in to not only ward off, but keep pace with, increasingly serious threats from upstarts (Tesla), fleet operators (Uber, Lyft) and the new technology establishment (Waymo, Apple), their ability to get to market quickly will depend heavily on the innovations coming out of the startup ecosystem.   

The flow of venture capital to autonomous driving startups, then, is hardly a surprise. 

That the underpinnings of the autonomous future depend on emergent technologies — such as artificial intelligence, high performance edge computing and the latest in software development tools — means that demands are not organically addressed by traditional automotive talent.  Most car companies are not heavily-stocked with, for example, deep neural net experts or hardcore LiDAR engineers.

Major auto players will be on the hunt not only for great technology and IP but also amazing teams, in order to build out pools of new-gen expertise they don’t currently have.  And unlike many enterprise or industrial spaces we have followed, timelines here are not open ended: nearly every major automaker has committed to some form of fully autonomous product or service by 2022.  With the performance bars set and time-to-market sketched out, startups are arguably facing less uncertainty here than in many other segments.

That said, with hefty dollars invested and much hype in the rear view mirror, the pressure to deliver has arrived.  I believe that 2018 will kick off a two- to three-year period that will begin to define the winners.  In this crucial stretch, major design-in decisions will be made for Gen 1 platforms. 

Startups will need to move beyond the headline grabbing “big idea” to real deployments.  And as new markets take shape, layers in the product stack will begin to flatten or swell in value.  Startups will need to adapt swiftly in this shifting landscape, dancing around commoditization and margin pressure. 

As a Valley investor, I have engaged with more than my fair share of inspiring autonomous driving startups in the past year (more on that in a future post) but even the strongest among them are not guaranteed victory on exceptional technology and impressive venture funding alone. 

Success will depend heavily on their flexibility and agility, in combination with collaborations and partnerships across the vast spectrum of participants on the autonomous playing field. 

This is where a big part of next47’s value proposition comes in.  Our objective is not only to invest in the most promising startups, but meaningfully connect them with decision makers and experts within Siemens businesses and also its vast global ecosystem of customers and partners. 

In so doing, we strive to confer a winning advantage in this complex and competitive autonomous game, where many pieces have to fall into place for startups to succeed — but winners will be richly rewarded. 


Debjit Mukerji is a director at next47, the global venture capital firm of Siemens. The VC firm recently led an investment round in, which provides flexible used-car leasing. 

Поможем вам подобрать и в любом городе.

Today's top headlines