Booming demand for inertial sensor functions in consumer gear like smart phones will generate robust 18.6% compound average annual growth in sensor units from 2011 to 2016, according to Yole Développement projections. But prices will continue to drop, holding revenue growth to 16.6% CAGR, for a $2.8 billion market opportunity by 2015. And the market is poised for restructuring, as a growing crowd of players scramble over the core function of integrating the silicon sensors into useful functions.
The basic fabrication of the MEMS die is only a part of this growing business. Despite the complexities of designing and fabricating MEMS devices, most of the value in these functions is not actually in the bare silicon sensor any more. Yole calculates the actual cost of front end manufacturing of the MEMS die will account for only about 17% of the total $2.6 billion total consumer and automotive inertial sensor revenue last year, while the ASIC, the packaging, the test & calibration and the software production costs make up some 41%, leaving roughly 42% total gross margin to be divided up among all the players in the value chain. Moreover, the MEMS die is likely to become an even smaller part of the total cost going forward as the sensors scale to smaller size and prices continue to fall. Meanwhile, the demands on the package, the testing and the software are growing more complex, and their relative value is likely to increase.
Increasing unit demand but eroding prices for discrete devices
Unit demand for all these sensors will continue to increase steadily, as more phones incorporate the established features of the leaders, and more applications are developed. But scaling sizes and volumes, and the growing competition from some 50 potential sensor suppliers, will continue to drive prices down, so revenue growth will level off.
Demand will remain strong for accelerometers, which are still only in 37% of cell phones, and still used predominantly only for switching between portrait and landscape orientation, in games and in pedometers. These features are becoming must-haves, and lower prices and more integration and application support from suppliers will extend the market into more and lower-cost handset units. Consumer accelerometer sales expanded 9% in 2011, but will begin to level off to essentially a flat market, with only 0.7% CAGR through 2016.
Demand for consumer gyroscopes started to take off in 2010 with the adoption in Apple’s popular products, and sales in 2011 jumped ~39%, but will also rapidly level off as the market matures and prices continue to drop, to see only 7.5% CAGR through 2016. So far demand seems driven largely by the potential for more precise sensing to enable possible new features that will distinguish Apple’s products, and are therefore also desired by all the phone makers who need to keep up with Apple. Close to 100% of tablets now have gyroscopes – only low-end tablets such as Nook Tablet or Kindle Fire do not integrate it. At this very early stage of market introduction, however, no widely used killer applications are as yet established.
The first gyro-enabled user interfaces are starting to appear, now that there’s some support for using gyro sensor data in the Google operating system software for Android phones. The LG Optimus Black incorporated the first InvenSense gyro in a phone for a tilt control interface--a major win for a startup and its MEMS foundry supply chain making it into the mega mobile volume world. A button on the side of the phone activates the gyro to allow moving from screen to screen with one hand by tilting the phone.
It also increasingly makes sense to use a gyro for image stabilization to improve the quality of photos and video, now that leading phone makers have raised the bar for camera quality. There’s even an image stabilization app to hold the gaming image steady for play in a moving vehicle—as some gaming motion controls jump around so much with external motion.
Sales of stand-alone magnetometers for consumer gear have also taken off, with an estimated 62% jump in 2011. But here too severe price pressure will limit revenue growth to 11.1% CAGR through 2016. There’s a scramble among magnetometer suppliers to challenge AKM’s dominant market share with alternative technologies they argue are more accurate, to improve accuracy from the current +/-5° down to the +/-1° or 2° likely required for accurate pointing for augmented reality applications expected in a couple of years.
Revenue growth to come from combo sensors
The coming growth opportunity in the consumer inertial sensor market will be combination motion sensor units, as the key applications for user interfaces and navigation require integration of sensor information, and integration offers a roadmap to reduce costs. Yole estimates sales of combo sensors will start to see market acceptance and strong growth starting this year, as the more accurate inertial-aided GPS becomes a strong differentiator for smart phones, starting the market on a ramp to close to $1 billion by 2016. First combo sensors to get traction will be accelerometers plus magnetometers. Each device can correct the other to provide more accurate location and navigation information, which is increasingly a must-have for smart handhelds, so some 400 million cell phones now already use both sensors. Combining the two sensors into one package with a single ASIC can potentially reduce both cost and footprint. Automotive sensor suppliers have recently had great success with their lower cost combo sensors for electronic stability control that combine two accelerometers and a gyro with one ASIC.
Accelerometer makers expect to combine magnetometers with their devices to supply the combo units. But magnetometer makers like AKM and Aichi MI are also adding others’ accelerometers to their magnetometers to offer combo sensors as well, though we don’t expect these mostly Japanese companies to drive the combo sensor market. The combo sensors will need very high yields to be economical, and systems makers will have to figure out a solution for a second supply source to get user acceptance.
Combinations of accelerometers and gyroscopes also make sense, as they are so often used together, particularly for controllers for gaming and remote controls. Nintendo’s 3DS handheld gaming unit, for example, adds accelerometers and gyros for control by tilt, as does a motion- sensing remote control for Internet TV that is currently a big success in France. The market for these combo units will develop a little more slowly, as the penetration of gyros in cell phones is still low. Power usage is also more of an issue, as any application that uses just the accelerometer, like the pedometer or landscape-portrait switching, will consume more power by having to activate the gyro in the combo too. On the other hand, since both sensors are made by similar MEMS processes, it should be easier to integrate them on the same die. The gyro makers will likely control the combo market, as the gyro is the higher value and less widely available device.
Who will capture the added value of these smart sensor systems?
The scramble is on for just who will add the intelligence to the system to turn the sensor data into useful functions. Early systems integrators did the job themselves, but MEMS device makers are also adding the software and functionality to make it easy for more systems makers to use the sensors without having to invest a lot of time or resources in development. Specialty software suppliers like Movea and Hillcrest Labs are also offering motion processing software that can do the task for any motion sensors. Core and specialty chipset providers also aim to take over more of this processing, so the sensor won’t need to add its own additional logic and memory chip components.