Why is Google shutting down lots of robots companies?

In 2013 Google bought seven robotics companies. One of them was Boston Dynamics. Now in 2016 Google wants to sell out Boston Dynamics. Why? They realized that it will take at least a decade to turn the operation into a profitable revenue stream and they realized that their cultures do not fit. So... Here is a mini guide to how to invest in robotics companies. I write it for venture capital and for friends and for foes who ask me for advice.

Rule #1 Invest in software (paradigm) oriented startups. Not in hardware paradigm oriented ones. For example. Jibo Corp. sells the Jibo robot. The robot has two components: the hardware and the software. To succeed they have to excel at both. I predict they won't. Being good at one thing alone is hard enough already. So how can they possibly be good at two?

I have seen so many robot startups do this very same mistake that I will not go through the list. On the other hand you have software oriented robotic startups. The ones that focus on AI only and use other's hardware (mobile phones) as brains. An example is Loujee, (which I advise). It is an app that runs inside a phone that is stuffed inside a pillow. And it is profitable.

Rule #2 West Coast vs. Japan / East coast. Now that we established that hardware is not the way to go then don't unless you are a vacuum cleaner manufacturer that requires one neuron to zig and another to zag.

Rule #3 Founder selection.Founder is an academic? red flag. Founder is not a software guy? red flag. Founder is enamored with hardware? red flag. Founder is not UX expert? red flag.

Rule #4 Robotics is about software not hardware. Hardware is a commodity. Therefore pick someone who understands software and its cycles.

Rule #5 Pick the right platform. ROS, Apple, Arduino...

Rule #6 Price. I have seen so many engineers (Honda Asimo, P..., ) build +$1M price tag robots that are too expensive for anyone to use. Make sure your team is price sensitive.

Rule #7 Focus on the toy/entertainment market first. Think who is the Atari for robotics. Think Mac 1987. Stop thinking Aldebaran 2017 because that is the "empty box" paradigm. A toy that comes with no games.

Rule #8 Don't do it. At the beginning of the century there were +400 car companies in the world. Today there are only 3. Today there are about 400 robotic companies around the world. In 50 years time there will be only 3 alive.

Rule #9 If you are going to do it anyway, ask yourself... Can this scale up? No? Then let Apple distribute the hardware for you. Piggyback.

Rule #10 8 /2 rule. A balanced robot cost is: 80% software cost. 20% hardware cost. If it is the other way around, ask yourself - Am I investing in a team under the influence of empty the box syndrome?

Rule #11 Amazon echo uses far field mic tech. That is a core UI you want to have in your product. Humans live in noisy environments.

Rule #12 AWS. Is your AI running on AWS? if not, red flag.

Rule #13 It is more sensible to invest in a company that makes a Tesla car smarter rather than in a company that wants to replicate Tesla. It is better to invest in companies that make robots smarter than in robot makers.

Think Windows vs. Taiwanese PC makers. Who made more money? And who are those companies today? In fact, they might not look like robotic companies at all.