Google Alphabet: Is it good for robocars?


Everybody has heard about Google's restructuring. In the restructuring, Google [x], which includes the self-driving car division, will be a subsidiary of the new Alphabet holding company, and no longer part of Google.

Having been a consultant on that team, I have some perspective to offer on how the restructuring might affect the companies that become Alphabet subsidiaries and leave the Google umbrella.

The biggest positive is that Google has become a large corporation, and as a large corporation, it suffers from many of the problems that large companies have. Google is perhaps one of the most unusual large companies, so it suffers most of them less, but it is not immune. As small subsidiaries of Alphabet, the various companies will be able to escape this and act a bit more like startups. They won't get to be entirely like startups -- they will have a rich sugar daddy and not have to raise money in the venture funding world, and it's yet to be seen if they will get any resources from their cousin at Google. Even so, this change can't be understated. There are just ways of thinking at big companies that seem entirely rational when looked at up close, but which doom so many projects inside big companies.

Here let me list some of the factors that will be positives:

  • While Alphabet has said nothing about the structure of the Google [X] companies, it seems likely that they will be able to give options and equity to their employees; options that might have a big upside. Google stock options have lost the big upside. Due to the structure, however, the equity packages will probably be smaller, with nobody getting the large chunks founders get -- and nobody taking the risks founders do.
  • It will be easier, of course, for Alphabet to sell off these subsidiaries, or even take them public or do other unusual things normally not done with corporate divisions. (It's not impossible with corporate divisions, but it's rare and it rarely is a bonanza for the staff.)
  • The subsidiaries will be freed from the matrix management of large companies. They will get their own legal departments, be able to set their own benefit structures and culture to some extent. Don't underestimate the value of not having to work within a corporate legal or HR department when you're trying to be a startup.
  • The companies can take risks that Google can't take. For example, consider Uber, which simply violated local laws in some area to kickstart ride service. It's much harder for a division of a large company to even try a stunt like that. For Uber, it worked -- but it doesn't always work.
  • The companies can also do things that would otherwise tarnish the Google brand. Huge as it is, the public has a natural distrust of Google, particularly on issues like privacy. While I think all robocar companies should work hard to protect privacy, being inside Google creates a whole new amount of scrutiny and established principles. In the case of making robocars, they might one day injure somebody, and that is a scary thing for the big brands. If you live in fear of that all the time, you won't win the race, either.
  • The CEOs of the new companies should have a lot more autonomy than they had before.
  • They still will have access to vast financial resources. If the new car company needs ten billion dollars to build a fleet of 400,000 taxis, or even needs to buy an existing car company, it's not out of the question.
  • Being inside Google conveys a certain arrogance to people because it's one of the world's leading companies in many different ways. But sometimes it's good not to be so cocky.
  • Out of fear, there are companies that won't do business with Google. I once asked the folks at WAZE if we might get their data on accidents. I was told, "The one company we would be afraid to sell our data to is Google." Of course, Google got WAZE's data, but at a much higher price!
  • People will finally stop wondering if they are building a car just to show advertising to you while you ride.

Of course there are some negatives:

  • Google brings with it vast, vast resources, not just in money. Google is also the world's #1 mapping company, and in fact many of the early members of the Chauffeur team were people who worked on maps and streetview. Google's world-leading computing resources also are useful for the big data projects and simulation a robocar team has to do.
  • There is also a giant talent pool at Google, though of course in all big companies, poaching top employees from within the company comes with risks of internal strife. The ability to even borrow top-notch people and resources is immensely valuable.
  • Google has fantastic benefits that are hard to duplicate in a small company. One suspects Alphabet's subsidiaries will probably mirror a lot of Google's policies, but there is a limit to what they can do. A google badge does not just get you dozens of restaurants and a large commuter bus system, it gets you things like a great series of internal talks from technical and world leaders and many other events. Google spends a lot on keeping its people happy. A lot.
  • Google has a fun company culture, with lots of cool people. People make a lot of friendships with very smart friends there, even outside their groups.
  • Inside Google there is always the opportunity to switch to different projects, many of which are grand and sure to affect a lot of people, without getting a new job.
  • The projects at Google [x] have the personal interest of Larry Page and Sergey Brin. That's been very useful to them within Google, but it also threatens the necessary independence of the CEOs of the new subsidiaries, who will still report to Alphabet. It remains to be seen if the founders can be sufficiently hands off.
  • Google is perhaps the world's top brand. It is able to get things done. When you call companies and say, "I'm calling from the Google car team" they return your calls right away and jump at the chance to talk with you. Doors are opened that are closed to most startups. (Admittedly, the project at Google is now so famous that it might overcome this.)
  • Google's power has allowed it to also do things like get laws made and changed around robocars; in fact this kickstarted the legal changes around the world. A small company will have a harder time.
  • Google's power gives it a strange upper hand in negotiations with other players like big car companies. Big car companies are very used to being in charge of any talks they have with partners and suppliers.

I have no inside information on this deal -- this is all based on lots of observation of public information about Google and non-confidential impressions from having been there. Some of this could be wrong. Alphabet might have Google re-sell some of its perks like the bus system to the other companies. It will certainly lend a hand where it makes a lot of sense. There is a fine line, though -- the more "help" you give, the more "perfectly reasonable" conditions the help comes with and soon you're like a division again.

There have been no specific announcements about Chauffeur either. Will Google [x] be a subsidiary with Astro Teller as CEO, including the car? Will the car have its own company? Will Chris Urmson be CEO if so? Or will [x] continue as the research lab of Alphabet, while other "graduated" portions of it go off into their own companies? Specific mention was made of "Wing" which is doing drones, but not of other [x] projects. More news will surely come.

Overall, I think this is a strong decision. If Google was to fail in the race to robocars, I always felt that that failure would come from one of two fronts -- either the mistakes that big companies make because they are big, or from the special hubris of Google and its #1 position. Now these two dangers are dimmed.


Recently Google created Google Auto LLC.

The Google Auto is beta-testing in Mountain View.

It seems Google is less than 3 years away from launching Auto on a wide-scale.

And you cannot do that under Google Inc.

Autos are a business on a scale which rivals search.
They require very different expertise and focus. As we see with the Google kuala car, autos look very different than regular cars. Many of the things that make a car good (handling, power, etc) are irrelevant. Usage cases are also very different (short range (< 100 miles), 1 or 2 people). Clients are mostly fleets (self-owned by Google, or purchased by Ubers). They are most probably electric. Someone needs to design them and build them. And legacy car companies have too much momentum in the opposite direction (all-purpose, ICE, driver-focused, privately owned) to do it.

So either Google will license the software, or it'll make them (and possibly operate them as a mobility service). If Google is thinking the latter (which I believe they are), and they want to scale-up in 3 years, they need to start getting serious about manufacturing the auto bodies, batteries, etc. And to do that you need a separate company from Google Inc.

You don't transform Google into Alphabet without a triggering event. I believe the decision to prepare for the Auto scale-up is that trigger.

but it would allow them to IPO Google Auto which could then fund the car purchase/build. So if the cars cost $50,000 each then 25,000 would be $1.25 billion. Uber could also IPO and fund this. I suspect Google will not manufacturer the cars themselves but outsource that to a contract manufacturer like Magna Steyr. It is also possible that the manufacturer will partner with Google to have an ownership stake in the service.

Apple doesn't really make any hardware. FoxConn does it for them. But it is still very much an Apple product.

So even if we get "Google Autos" for fleets, I agree they will probably not be made by Google factories.

But on the other hand: who can make things like these? There is no good answer.

My gut feeling is that legacy car companies are not well suited to make a small 2-seater electric car. It's simply too different than what they usually do. No expertise in electric drive train, batteries, light-weight construction, etc. And they don't do anything significant on a 2-3 year timescale.

Tesla has the know-how to do it. But I doubt they can scale-up even faster than they are already doing.

So I see no obviously well-suited manufacturing company to do make a million Koalas in the next 5 years. Just the batteries will require a second Gigafactory.

Exciting times.

Add new comment