Why Google created a self-driving car and DuPont didn’t

Adam Hartung

Adam Hartung

This week a self-driving car built by Delphi of England completed a 9 day trip from San Francisco to New York City.  The car traveled 3,400 miles, and was fully automated for 99% of the trip.

Attention has again focused on self-driving cars.  There are a handful of players entering the market today, including Apple.  But the most famous company by far is Google, which has put over 700,000 autonomous miles on its vehicles since pioneering the concept after winning a DARPA challenge to build a functioning prototype in 2005.  In fact, we’re so used to hearing about the Google self-driving car that many of have stopped asking “Why Google?  They aren’t in the auto business.”

Of course the idea of a self-driving auto is as old as the Jetson’s (and if you don’t know who the Jetson’s are you are, that was a long time ago.)  And nobody should be surprised to hear that prototypes have been on the drawing board for 5 decades.  But I bet you didn’t know that DuPont was once seriously engaged in such development.

In 1986 DuPont was America’s largest and most noteworthy chemical company.  The company was a pioneer in petrochemicals, and was considered the company that brought the world plastic – at a time when plastic was considered a great, new invention.  A leader in films of all sorts, DuPont leadership saw the opportunity for electronics to replace film in applications such as printing (where films were used in high volume for platemaking and proofing) and healthcare (where xRays and MRIs were a large film users.)  They conceived of a future time when computers and monitors – digitial products – could replace analog film, and they chose to create a new business unit called Electronic Imaging to pioneer developing these applications.

As the team started they expanded the definition of Electronic Imaging to include all sorts of applications for digital imaging – and using all kinds of technologies.  The breadth of analysis, and product development, included non-destructive parts testing, infrared uses such as heads-up displays and inventory identification, and radar applications.  Which led the team to using a radar for automating an automobile.

In 1987 DuPont invested in a small company out of San Diego that accomplished something never done before.  Using a phased-array radar hooked up to the brakes of a van, they were able to have the car recognize objects in front of the van, calculate in real time the distance between the van and these forward objects, calculate the relative speed of both objects (whether one or both were stationary or moving) and then apply braking in order to maintain a safe distance.  If the forward object stopped, then the van would come to a complete stop.

This was all done with discreet componentry, and the team realized future success required developing more specific electronics, including specialized integrated chips that could operate faster and be more error-free.  So they drove the prototype from San Diego to Wilmington, DE with a person behind the wheel, but relying as much as possible on the automated system to do all braking.  The team collected data on location, speed, weather, traffic conditions, and many other items during the journey and prepared to take the project forward, planning to eventually build a module which could be installed in vehicles as small as cars or as large as 18-wheelers, with enough intelligence to adjust for different vehicle designs and applications (in order to calibrate for different braking distances.)

Net/net they had a working prototype.  The product was expected to reduce the number of accidents by assisting drivers with braking.  Multi-car pile-ups would become a thing of the past.  And this device could potentially allow for better traffic flow because automated braking would reduce – maybe eliminate! – rear-end collisions.  This wasn’t a self-driving car, but it was self-braking car, which would be a first step toward the sort of Jetson’s-esque vision the young team imagined.

What happened?

It didn’t take long for the older, “wiser” leadership to shut down the project.  Even though several executives participated in a controlled demonstration of the prototype in an enclosed DuPont parking lot, the conclusion was that this project demonstrated just how off-track the new Electronic Imaging Department had become, and that it was clear folks needed to be reigned in and budgets cut:

  • This clearly had nothing to do with film or replacing film.  DuPont was a chemical company, and to the extent it had any interest in electronics it was where they were applied to potentially cannibalize film sales.  Products which were not closely aligned with historical products were simply not to be pursued.
  • DuPont had no history in radar, analogue electronics or development of integrated circuits.  Yes, DuPont had an Electronics Department, but they sold film for solder masking and other applications of semiconductor and electronics manufacturing.  DuPont was a chemical company, not a computer company or electronics company and this division was not going to change this situation.
  • This product was seen as carrying too much liability risk.  What if it failed?  What if the car ran over a child?  The auto industry was seen as litigious, and DuPont had no interest in a product that could have the kind of liability this one would generate.  Yes, there was an Auto Department, but it sold films for safety glass, plastic sheets used for molding inside panels, and surface coatings which could be painted on the inside and/or outside of the vehicle.  But those did not have the kind of failure possibility of this active radar device.  [“By the way” the vice-Chairman asked “could that radar fry someone’s innards at a crosswalk?”]
  • The market is too limited.  Who would really want an automatic braking system?  Given what it might cost, only the most expensive cars could install it, and only the wealthiest customers could afford it.  This product was destined for niche use, at best, and would never have widespread installation.

Poof, away went the automatic automobile braking project.  Once this dagger had been thrown, within just a few months everything that wasn’t printing or medical – in fact anything that wasn’t tied to printing films, xRays and MRI – was gone.  Within 2 years leadership decided that for some variant of the 4 issues above the entire Electronic Imaging division was a bad idea.  DuPont would be better served if it stuck to its core business, and if it spent money defending and extending film sales rather than trying to cannibalize them.

DuPont liked competing in the oceans where it had long competed.  Venturing beyond those oceans was simply too risky. Today, 25 years later, DuPont is about 1/3 the size it was when its leaders launched the ill-fated Electronic Imaging division.

Google obviously has a different way of looking at the opportunity for automating automobile operation.  Since winning the DARPA competition Google has spent a goodly sum building and testing ways to automate driving.  And it has even gone so far as lobbying to make self-driving cars legal, which they now are in 4 states.  Pessimists remain, but every quarter more people are thinking that self-driving cars will be here sooner than we might have imagined.  This week’s cross-country achievement fuels speculation that the reality could be just around the corner.

Google seems happy to compete in new oceans.  It dominates search, where its share is attacked every day by the likes of Yahoo and Microsoft.  But simultaneously Google has invested far outside its core market, including software for PCs (Chrome) and mobile devices (Android), hardware (Nexus phones), media (Blogger, YouTube), payments (Wallet) and even self-driving cars.  To what extent these, and dozens of other non-core products/services, will pay off for investors is yet to be determined.  But at least Google’s leadership is able to overcome the desire to restrict the company’s options and look for future markets.

Which kind of organization is yours?  Do you find reasons to kill new projects, or are you willing to experiment at creating new markets which might create dramatic growth?

(Adam Hartung is the managing director at Spark Partners. He blogs here.)