DAC Opens with an Enlightened Keynote

The opening keynote of DAC was by Joe Costello. Joe, of course, was the first CEO of Cadence (created when SDA was merged into ECAD) and held that position for over a decade. Over ten years ago, he gave a keynote at DAC and gave his rules for bringing a product to market. Now he is CEO of Enlighted, which does a lot more than lighting, as he explained. In some ways, this keynote was his rules for success in the IoT market, titled IoT: Tales from the Frontline. Joe said that the keynote came about when Mac (this year's DAC chair) called him and said "Can you do a keynote?" and, without thinking much about it, he said "Sure."

2000 in Emeryville

Joe told about his IoT experience in 2000, long before the words "internet of things" had passed anyone's lips. He started an incubator in 2000 focused on technologies for education. But 20% of the time, engineers could work on what they wanted. This was headed up by Luke Julia (now at Samsung), who started to think about "things". They spent a lot of time in that era working out how to connect things to the Internet. Remember, this was 2001. The iPhone was still six years in the future. Even WiFi was not yet ubiquitous. In the end, they put it into suspended animation because they couldn't see a path to market.

Meanwhile, for a decade, Joe watched all the hype about IoT and now he had an idea of what would work and what wouldn't. A few years ago he joined Enlighted as CEO.

Joe thinks IoT will be bigger than anything that has come before. The Internet created a lot of consolidation with only a few suppliers of desktop, servers, and smartphones. But he thinks IoT will be the opposite, with tremendous proliferation but without the same consolidation, since everything has different requirements. It will be way bigger than the Internet.

Focus on the High-Order Bit

So what has he learned in the last 15 years? He admits it's not perfect, but he can tell you a few things that didn't work. He reminded us of Jim Moyer, his first boss at National Semiconductor (pre-Cadence, pre-SDA even): "the high-order bit...focus on the high-order bit." That is an analog designers way of saying that you must focus on the big-picture direction, not the details. If you get that right, you can correct other things.

So the high-order bit is what thing you pick. People screw up right there at the high-order bit. For example, there are half a dozen companies doing WiFi-connected dog bowls. There are seven companies doing sprinkler systems for your phone. These are not inherently bad, but they will eventually get subsumed by IoT platforms that change the world. IoT will be a feature of products and it will touch everything. But it doesn't make for a separate company or a good market. There will just be a few defining new companies, the IoT equivalents of Facebook, Google, and co. That is the end game of Enlighted, not lighting.

Another mistake people make is that IoT is about intelligence. For example, Joe remodeled his house and put Nest thermostats everywhere. But his wife wants to get rid of them. "But I'm an IoT guy, they are great," Joe protested. "But I have to find my phone, find the app, remember how to use it. By then I could have walked through the whole house and adjusted the old thermostats." IoT needs to be more about data plus processing to create intelligent action, what Joe calls IQoT.

You presmably know that GE build lots of big "things" like jet engines and locomotives. They sensored them up and turned a reactive prophylactic service business into a business where they do proactive service. In the jet engine business alone that is saving $1B per year and the companies are getting better service. The money is not in dog bowls, or even light bulbs.

Enlighted

Finally to Enlighted. If you have heard anything about them, then you probably have heard that they come into commercial buildings and replace all the lights, and use the energy saved to fund their business. But actually they do much more than that.

The definition of their business is IoT for commercial real estate. Commercial real estate may be the largest business in the world, certainly in the top three. Enlighted is the answer to the question "What if we made the building into a connected object that is intelligent?"

So they create sensory systems for previous inanimate objects, known as buildings. They put a sensor module in every light fixture. This turns out to have a lot of advantages since any building has enough lights to light everywhere, fairly evenly distributed throughout the building, so everywhere is visible from at least one light. The processing power in each module is about the same as the original iPhone, so every 100 square feet or so there is a network of processors. Each module has seven sensors (not including video, mainly for privacy reasons). They compress all the data in the sensors since it would otherwise overwhelm the wireless network. For example, their largest installation has 22,000 of these modules. The volume is huge even after compression. Another advantage of light fixtures is that they have power so there is no need for batteries to be replaced.

Another advantage is that they put a sensor in every light fixture, so there isn't any need to "design" the network, there are no choices. After all, these are no PhD engineers installing these fixtures, they are journeyman electricians and the like. So the first application is saving energy from lighting by building the most efficient lighting system you have ever seen. More by luck than anything, it turns out that the money from saving on lighting is just about the right amount to pay for the sensor system. Everything else on top of that is already paid for, and they can produce much richer data.

I said that there are no cameras but there are heat sensors that can see people as heat blobs so they know where people are. They can use this for space planning since they know how many people are in which part of the building and have a perfect picture of how a building is being used. This is worth something like $3/square foot compared with the lighting which is just a few cents. Joe met the CFO of Apple and their facilities manager. The facilities manager said "We don't need this, we are running hot, we just need new buildings, we are running out of space." So they took two buildings of 40-50,000 square feet and added the sensors. Showed them the blobs moving around. They then pulled back and showed that the building is used just 40% of the time. So the CFO turned to the facilities guy and said "i told you we are not running hot. i want this in every building." Real estate is a huge cost, and even Apple had almost no idea how it was being used.

In a different market, Joe talked to the CEO of a hospital. His response was “I don’t much care about your energy stuff, but it could do asset tracking and revolutionize hospitals." It turns out that nurses spend an average of one hour per day looking for equipment. The hospital needs a way to see where it is in real time. It's actually worse, since nurses hate looking for equipment, they want to do true nursing. Of course they can put a Bluetooth LE tag on equipment and run it for five years. The CEO told them to start with wheelchairs. Joe figured that this was a test, they would try it somewhere that didn't matter, and later they'd get to tag MRI machines and portable X-rays. But it turns out that wheelchairs are a huge issue. In general, you can't leave a hospital except in a wheelchair. No wheelchair means a patient has to stay in their bed, sometimes overnight. It turns out that this is a 9% problem for this hospital, that their occupancy rate is 9% higher than it would be if wheelchairs are always available, and in most cases that is a pure cost. The CEO didn't give them wheelchairs because they were a trivial test case, but because it was his biggest problem.

Another meeting, this time a bricks-and-mortar retailer. "Have you looked at our website?" he said. "It's terrible. We are a top retailer, but our website is in the bottom 20%. But even so I can do all the tracking I want. In the store? I know nothing until you check out and then only if you use a credit card." Bricks-and-mortar retailing may be dying but the other side is warehousing, and that is not going away, and there is even more need there to track all stuff all the time. 

So the high-order bit is to pick the right thing and make sure you are gathering the data. There is nothing without the data. It is the crude oil of the IoT revolution. So whatever you are doing, get the data, and protect it like crazy since that's where the value is. Several big AI machine-learning companies came to Enlighted. They need the data for their AI, but they are not going to just hand it over. If you don't have the data, it doesn't matter how fancy your tools are.

However, data is only as powerful as what you do with it. The next generation will be data and machine learning. Dense data. It will have a profound impact. A lot of blue collar people have been put out of work by robots, but next will be white collar people. But the combination of human and machine is the best. Computers can beat humans at go and chess. But humans plus machines are even more powerful and can beat both other humans and other machines.

Going to Market

Going to market is not simple in IoT. Enlighted's goal is to enlighten every building. There will only be a handful of platforms that survive. Joe talked about jet engines earlier. How many platforms will there be for handling jet engine data? Maybe a couple.

The big goal is to be one of the last five people left standing. You will need to be gigantic to be a player in the end, scale is the biggest thing by far. That's why WiFi dog feeders are not that important—it's not a good place to start.

Since Enlighted is in buildings, Joe said he would use that as an example. But it is the same in all markets. You have old technology, existing stuff, very slow moving. On the other hand, you have IT, Moore's Law, fast, innovative.

But you have to close that gap. There is a huge gap between Enlighted sensors and the lighting and HVAC guys. But those are the guys you have to work with, you can't use AI engineers to install air conditioning. There are dozens of companies who supply lighting and HVAC, but they don’t have DSP, etc. in their DNA. The IT guys (and venture capitalists) get excited until they hear there are guys on ladders putting stuff in ceilings. Then they want you to come back once it is all installed and give them the data. It is a culture clash between existing guys with the thing, and 100 years background, and on the other side, the IT guys with intelligence but who don’t understand the old stuff.

The go-to-market challenge is to build a bridge and understand the old technology side well enough to build a robust connection. In effect, you have to give up everything for the first five years and learn how to build bridges. Then you can truly scale, and if you don't scale, someone else will and you will be crushed.

The Opportunity Is Bigger Than the Internet

Joe said he never planned to run a company again. But the opportunity is bigger than he has ever seen. Joe started life as a technology guy, like most of us, and will die a technology guy. This is the golden age of technology, fueled by people inventing and pushing the boundaries. This all started 17 years ago in a little lab in Emeryville that they had to shut down because they were too early.

But now it is real.