As electronics OEMs shift their focus toward wearable devices, so too are semiconductor vendors. As with any major industry shift, there is a fresh batch of start-ups looking to capitalize on the emerging opportunity. One such startup, Ineda, is coming out of stealth mode with a new round of funding and the announcement of the Dhanush family SoCs for wearables, or as the company refers to them Wearable Processing Units (WPUs).
The company boasts board members, executive staff, and engineering talent from other leading SoC and IP vendors. In addition, Ineda counts Imagination, Qualcomm, and Samsung among its investors. The goal of the company is to provide SoCs or WPUs that meet the extremely low-power and long battery-life requirements of wearable electronic devices. This includes some very high goals, such as 30 days of always-on battery life, with always-on speech recognition, sensor analysis, and contextual computing, and also Bluetooth Low Energy (BLE) connectivity.
To accomplish the task, Ineda plans to introduce a family of Dhanush SoCs, including the Nano, Micro, Optima, and Advanced. They aim to meet the demanded system prices of wearables that start at US$299 on the high-end and extend to below $99 on the low-end.

At the heart of the Dhanush WPUs are three CPUs that range in performance and power consumption combined with shared memory, I/O, display, storage and file system, and accelerators in a heterogeneous computing configuration. The Dhanush WPUs will feature MIPS cores and other licensed IP, but the configuration of the SoC will require new software development tools that Ineda has developed. The development kits are available now, and product will be available for shipping in the fourth quarter.
The Dhanush product family is positioned against other low-power MIPS products, as well as the ARM Cortex-M class MPUs being used in current wearable devices. Unfortunately, without power and performance specifications, we will have to wait to compare the Dhanush WPUs to other SoCs. The major challenge for Ineda is convincing OEMs to use yet another platform. To convince OEMs, the Dhanush products will have to meet the power and performance goals Ineda has established and leapfrog the other solutions available in the market.
There is definitely room for other hardware architectures and platforms in wearables and other IoT applications, especially with the use of other software platforms. Samsung has already announced plans to use Tizen and other real-time OSs (RTOSs) in such applications. It is also likely that other silicon vendors and OEMs will experiment with all the available silicon solutions, but supporting multiple platforms and ensuring compatibility is challenging.
The more interesting aspect of the announcement, however, is the investment by Imagination — the company that acquired MIPS — and the investments by Samsung and Qualcomm, two major ARM vendors. If Ineda is successful, one of these companies is likely to acquire Ineda and/or the company’s IP.
Jim McGregor is Principal Analyst at Tirias Research

Taking the example of the Google Glass rated battery capacity of 570mAh, 30 days or 720h of standby time would mean a total device leakage including SoC, PMIC, display, wireless and other links of less than 800uA which sounds like a very challenging goal and in order to hit it I'd expect the wearable would be doing nothing i.e. in standby most of the time.