San Jose, CA. Does your sensor functionality represent a feature or a discriminator? That was the question posed by Dr. Kaigham J. “Ken” Gabriel, president and CEO of The Charles Stark Draper Laboratory in his keynote address today at Sensors Expo & Conference. He said his comments in the address would be informed by his earlier experience with CMOS MEMS-based microphones.
He began his presentation with a bar chart of the market capitalizations of two groups of companies. Group 1 on the left of his chart included companies like Intel, Qualcomm, and Texas Instruments. On the right he presented companies including Apple, Alphabet, Microsoft, and Facebook. The companies in the latter group, he said, have market capitalizations reaching three to 50 times that of those on the left. That, he said, is because the companies on the right are closer to the users—a vantage point that allows them to identify a “sensor discriminator instead of just feature.” As the world operates today, he said, the semiconductor and sensor makers (who offer features) on the left need design wins from the companies on the right (who can identify discriminators) to survive.
End customers, he said, don’t understand the value of components. They don’t know how a cellphone decides whether to present an image in portrait or landscape mode—if they give it any thought, they probably think it has to do with software. As for cellphone and other mobile-device makers, they are looking for consistent reductions in component cost.
“The destiny of the semiconductor industry is in the hands of others,” he said, adding, “Is that ok?”
It’s not, he said: “We need to change the game.” He noted that right now you can buy an IoS or Android device and choose from any number of apps—throwing away the ones you don’t like after a few hours, then selecting others. If you went into a cellphone store right now and were told the device you propose to buy would be stuck with the same software for three years, “…you wouldn’t put up with it. But that’s exactly what we put up with in hardware.”
Hardware has a history of modularity, he said. He recalled his work with DARPA in the ‘90s upgrading the B-52. “That aircraft was first built when I was born,” he said, but every single part of the original has since been swapped out. And, he said, original IBM mainframes were more modular than people remember.
With mobile devices, however, people value hardware that’s compact and monolithic—“black boxes that cannot be violated,” Gabriel said. With the emergence of the IoT, he said, it’s time to challenge that view.
How might a modular hardware concept work? He cited the bizarre example of a cellphone with integrated cigarette lighter—the Chinese market absorbed about a million of those devices, he said. There might not be much of a market for such a device now, but what if you substituted a blood-oxygen sensor? Such are the goals of initiatives such as Alphabet’s Project Ara, the modular phone.
He noted that today consumers should be able to buy a cellphone with twice the battery life that now comes standard, but that simply isn’t an option for you. There would be a price—the recharge cycle life would fall to 150 cycles from 300. But some people might want to make that tradeoff. But because the industry has settled on 300 cycles, the long-live-battery life is not available. A modular phone would solve that problem
Such choice would be good for hardware makers as well as consumers, he said. Now, component makers face pressure for 3% or 4% quarter-on-quarter price reductions on parts costing a few cents. With modular hardware, consumers—less price sensitive than OEMs—might pay a few dollars for the right hardware module.
“I’ll take the price of an app for my sensor any day,” he concluded. “A buck is a tremendous margin for sensor products.”
