Segars and Son

When I was growing up, there was a TV show in the UK called Steptoe and Son, and renamed Sanford and Son when it was remade for the US market. Yesterday was the Segars and Son show. Son's company Softbank acquired Simon Segars' company ARM about six weeks ago. Actually, Son's name is not pronounced like "son", more like "song" without the "g". And nobody is a junk dealer.

Simon Segars, CEO ARM

To open ARM TechCon last week they did a joint keynote. Simon kicked off and then passed the baton to Masayoshi Son. He is Japanese, but if his name doesn't look very Japanese that is because he is of Korean descent.

Simon started off pointing out that ARM had processors in $50B of silicon shipped last year, which is a 20% increase YoY. The entire semiconductor industry is $340B but $70B of that is memory so ARM is in nearly 20% of all non-memory products. They are in supercomputers, and in barely visible sensor-based IoT things.

He said that the deal closed very fast, in just seven weeks. That mean that they didn't have to spend a lot of time focused on closing the deal (a deal that size normally takes months) but getting their heads down to deliver the benefits. As Simon said, "It was a new experience for me" since it was his first experience of going through an acquisition. Simon said he was constantly being asked why Softbank acquired ARM, so rather than giving his own explanation, he invited Masa to open ARM TechCon 2016. Simon first met Masa in Japan about 10 years ago along with Warren East (Simon's predecessor as CEO of ARM). Softbank had just acquired Vodaphone Japan and they were exchanging notes on mobile in Japan. Son has not been sittiing idle since the ARM acquisition. He just created a vision fund worth $100B. With that, Simon got off the stage.

Masayoshi Son, CEO Softbank

The first thing Masa said was to emphasize "the business model of ARM does not change. I am committed to provide neutrality for all of its partners. Don't worry, I will make everybody happier."

Masa started with an analogy, the Cambrian explosion. Before the Cambrian explosion 542 million years ago, there were just a few species on earth. Afterwards, there were tens of thousands. Many are now extinct. (If you want to know more about it, I recommend Steven Jay Gould's book Wonderful Life: The Burgess Shale and the Nature of History.) According to Masa, the first species with eyes was the trilobyte, but actually it seems nobody knows. Before the Cambrian explosion there were no eyes, afterwards, many species had eyes of one form or another.

Masa thinks that we are on the cusp of a Cambrian explosion being accelerated by IoT and artificial intelligence (AI). This won't just affect the semiconductor industry, it will accelerate human evolution. The center of gravity of the internet has moved from PC to mobile, and now is moving to IoT.

When people ask “Why are you interested in acquiring ARM? And why now? And why 40% premium?” he answers, "Now is the time, this is the Cambrian explosion. IoT devices will exceed mobile in by 2018."

Son had data predicting a cumulative one trilliion IoT devices will be created between 2016 and 2035. That's a lot. Data traffic will increase by 2500 times from 1 exabyte per month to 2.3 zettabytes per month.

One big challenge is security (for a recent post by me on the subject of IoT security, see Video Cameras: No Service for You) with somebody being hacked almost every day. Masa reckons secuirty is more important than things like doubling the clock cycle or increasing the memory. In a new car, there are hundreds of chips but no encryption, no security. In fact, the Thursday keynote at TechCon is by Charlie Miller on Automotive Security: A Hacker's Eye View, which I will be writing about separately (Charlie was one of the two people who took over a Wired journalist's vehicle, which you can read about in Hackers Remotely Kill a Jeep on the Highway—With Me in It.)

The other important trend is AI. Speech recognition and visual recognition are starting to surpass humans doing the same task. Son believes "the singularity is near." This is a phrase created (or perhaps just popularized) by Ray Kurzweil in his book The Singularity Is Near: When Humans Transcend Biology.

Son wrapped up by going back to the development of eyes. "Do you want to eat or be eaten?” The ones who can evolve with the new technology can live a better life. He believes that 1T IoT devices will jump-start human evolution. It will be a revolution requiring the collective effort of mankind. "This is coming."

Q&A

These presentations were followed by a Q&A with Simon and Masayoshi, moderated by ARM's VP marketing and bizdev, Ian Ferguson. Son emphasized that ARM just provides the core (and some other IP), which has to be integrated with all sort of other functionality, which will only happen with partners. Simon said that we were only at the beginning of seeing how connected devices can provide benefits (for example, most cars are not "connected" yet). WIth Softbank's backing, ARM would now be able to make bigger, longer-term bets, and get things done more quickly.

Son was asked about getting involved in ARM's strategy. "If you have a great ship, and great direction, and a great captain, there is no need to micromanage. I'm very excited about talking strategy, next 10 years, next 20 years. I'm not that interested in the next quarter."

Simon was asked about China. "It's a big market for us, a ton of innovation. A lot of focus on smart cities. Lots of experimentation. People are taking low-cost technology and seeing if it works or not. ARM is investing there, they have launched three or four accelerators there and are putting an investment fund together with a local private equity company.

Talking of funds, Son talked a bit about the recently launched vision fund. "The singularity is definitely coming, the explosion of IoT is definitely coming, That will need to the re-invention of many things. First the media has been reinvented, then commerce, now transportation. In the future, finance and healthcare. There is a huge range of opportunities, $100B is not enough but a good start."

Ian said that was impossible to follow and the Q&A wrapped.

Next: Wave Computing: a Dataflow Processor for Deep Learning

Previous: Automotive at Linley: Intelligent Vehicles and Intelligent Intersections