All the components already exist for billions of tiny computers to be worn on our wrists, embedded in traffic lights, and scattered around factories. But Tyson Tuttle, the chief executive of Silicon Labs, predicts it will take "multiple decades" to take full advantage of the Internet of Things.
“The jet engine was invented in 1937 but it took fifty years for airplane networks to be deployed where we could jump on airplanes to get anywhere,” Tuttle said in an hour-long keynote last week at the Design Automation Conference in Austin, Texas.
Tuttle, who assumed the chief executive role in 2012, offered a peek inside Silicon Labs’ strategy over the last seven years. The company has increasingly moved its business toward connecting everything from factory equipment to home appliances, taking $314.6 million from its Internet of Things business last year, up from $209 million in 2014.
Silicon Labs is resting its future on chips that impart sensors and connectivity to “old school devices” like tennis rackets that analyze strokes or commercial drills whose torque can be calibrated from a smartphone to simplify delicate tasks like screwing down solar panels. Tuttle predicted that the largest markets would be industrial and medical devices.
“But how do you make this easy? How do you make complex software and hardware simple, so that an industrial company in Milwaukee can easily design products, not the RF and level-register,” said Tuttle. The plan is for Silicon Labs to integrate “everyone else’s devices, instead of the other way around.”
Executing on that plan starts with SoC platforms that are cheap enough to be manufactured in large volumes. Tuttle envisions that these IoT SoCs will contain power management, sensors and interfaces, embedded memory, mixed-signal blocks, secure processors, and radios that support standards from Bluetooth to Thread.
“You can’t design a product today without it being connected,” he added.
Shrinking and squeezing these components onto the same slabs of silicon has been the heart of Silicon Labs’ success over the last two decades. Silicon Labs, founded in 1996, made its name in tightly integrated RF synthesizers and transceivers. Tuttle was hired as a chip architect for the company in 1997.
That strategy also extends into the software stacks that let chips communicate wirelessly and development tools like Simplicity Studio, which helps developers to optimize chips for applications from smart lighting to asset tracking. The role of Silicon Labs is to make it easier for companies to profit from the data harvested from embedded devices, Tuttle said.
Tuttle has backstopped that strategy with small acquisitions. In 2012, Silicon Labs spent $72 million on Ember’s wireless sensor network chips for the 2.4 gigahertz band. In 2015, it paid $61 million for Bluegiga Technologies to expand its war chest of Bluetooth and Wi-Fi modules and later acquired Telegesis, a supplier of ZigBee modules, for $20 million.
The acquisitions have also added new dimensions to Silicon Labs’ software stack, which helps reduce the time spent connecting devices to each other and the cloud. When it acquired Zentri last year, for instance, it acquired a Wi-Fi software stack for applications like smart meters and household devices, making it easier to update the devices remotely.
This year, Silicon Labs introduced new boards under the brand Mighty Gecko, which contain specialized software that enables a single chip to efficiently swap between wireless protocols using little more than a smartphone and Bluetooth connection. Hundreds of devices can be configured to join a ZigBee or Thread network.
Today, half the engineers at Silicon Labs are writing software to help developers with these types of tasks, Tuttle said. Last year, it also bought Micrium with an eye toward customizing and optimizing its real-time operating systems, so that developers could avoid using a host processor to run the RTOS on Silicon Labs’ chips.
“The silicon is really the heart of the system,” said Tuttle. “But it’s not just about the chip.”