Investors Want To Jump Aboard a Self-driving Car but the Tech Still Has Some Way to Go

By November 29, 2018 April 3rd, 2019 The Weekly
GM Cruise Self Driving Car

This week, General Motors announced its plans to sack more than 14,000 employees and shutter seven factories in the US and Canada by end of next year.

The culling is intended to make room for GM’s significant investments into electric vehicles (EV) and autonomous driving – the car industry’s current darlings.

Reaction has been mixed – the stock markets were joyful, with GM’s stock jumping almost 8 percent on the day of the announcement. Meanwhile, politicians from Justin Trudeau to Donald Trump were frustrated about the loss of jobs.

But the market might be getting too excited about an industry that still has a lot of uphill battles to wage.

Bumps Ahead

Investors are jazzed about GM embracing its EV and self-driving destiny. Its highlight acquisition, Cruise, has positioned the company as a frontrunner in the self-driving space. Analysts even see it as farther ahead than pioneers like Waymo or tech dazzlers like Uber and Tesla.

In a new Insight, analyst Henry Kwon notes that the industry is still stuck in what’s known as “Phase 3” of self-driving technology, with significant progress more than a decade away.

The Society of Automotive Engineers (SAE) splits progress in self-driving in five phases, from driver-monitored and assisted cruising all the way up to fully autonomous driving in any kind of weather or road conditions.

In Phase 3, a driver is still needed but is able to pass on several functions to the car as long as traffic and environmental conditions are ideal. It’s the one that’s giving companies the most trouble at the moment, as Uber found out the hard way.

Kwon’s concerns about autonomous driving technology include challenges still faced by engineers and developers – the sheer amount of data that needs to be processed or the need for artificial intelligence to respond to unpredictable situations.

Want an example? Jerusalem-based computer vision powerhouse Mobileye says the data processing needs of a fully-equipped self-driving car generates about 5TB per hour. That’s a lot of USB sticks.

Regulations are also getting in the way. Several automakers around the world are trying to solve ethics problems like what should an autonomous vehicle do when faced with an impossible dilemma.

Kwon cites Germany’s example, where it is illegal for machines to decide whether to kill one person in order to save another – making it impossible for currently developed autonomous cars to function there.

Pit Stop

Tesla has been quietly pushing its own full self-driving (FSD) option to the background, after facing several hurdles with chipmakers from Mobileye to NVidia, according to Insight Provider William Keating.

While nowhere near the capabilities of self-driving tech the likes of Waymo, Tesla’s system has arguably been one of the more visible in the mainstream. Cars outfitted with its (Phase 2) Autopilot system have been on the roads for a while.

As development costs mount and delays get longer, “Tesla’s removal of FSD as an optional extra on its website reflects a long-overdue reality check,” says Keating.

Want to dig deeper in the themes discussed above? Explore some of the investment research network’s Insights listed below.

More Insights by Smartkarma

Sign up or log in to Smartkarma to read more.

General Motors (GM): Trading in ICE for EV and AI?
Henry Kwon weighs in on the market’s reaction to GM’s shift to EV and self-driving cars.

Notes from Detroit: The AI in Automotive Conference – The Suppliers’ Perspective
In an Insight from earlier this year, Kwon discusses the challenges that still plague autonomous driving.

Tesla Quietly Drops Its Delusional Full Self-Driving Option As Reality May Finally Be Dawning
William Keating analyses Tesla’s self-driving reality check.

Tesla’s Autopilot 3.0 Hardware – The Delusion May Be Dawning
Keating takes a look at Tesla’s self-driving hardware troubles.

Get a Different Perspective

Sometimes, it helps to take a step back and take a wider look.

If Waymo Is Having Difficulty, Everyone Else Must Be In A Right Autonomous Driving Pickle
The difficulty of developing AI robust enough to drive a car anywhere, anytime.

The Self-Driving Car Timeline – Predictions from the Top 11 Global Automakers
An explainer on the five levels of self-driving tech and automakers’ plans for it.

Subscribe to Our Blog

Leave a Reply