siliconindia | | November 20179More companies will continue to do design work in India given the availability of good engineers in digital design and embedded systemsand IP development, to developing derivative SOCs and now to full-blown complex SOC IC development. Given the complexity of FinFET based SOCs, the large staffing needs continue to cut across international boundaries. India continues to focus on its current strengths in IC design and development, embedded software, and supporting MNCs. Analog will continue to be a growth area in semiconductors, to power these SOCs. To develop an indigenous semiconductor manufacturing ecosystem, there would be a need for government subsidies to develop manufacturing, packaging, test, and production capabilities. Companies established in the West leverage the East (Taiwan and China) for these capabilities. All though electronics and hardware industry in India is expected to be around $130 billion (ASSOCHAM), the indigenous market for semiconductor consumption is not large as Printed Circuit boards are currently populated in China or Taiwan. What are the latest technologies and trends emerging in the field of semiconductor?New applications in IoT, cloud, and AI will continue to push the need for performance and power consumption. Technologies (FinFET) are fast moving to 7nm and FABs are starting to work on 5nm. Analog continues to morph to smart analog with the advent of simpler MCU's integrated with Analog; GaN will continue to drive new efficiencies for Power applications and some major advances in packaging like the Intel's 17-qubit superconducting test chip. Some of the emerging technologies that will have a high impact include:IoT Edge: The growth of IoT Edge devices will continue to drive the need for smart devices that are more mixed signal in nature. Powering these rapidly growing IoT Connected cars/transportation will be major areas that will drive Edge analytics with ADAS-based systems in the short run and tie in with cloud analytics based on V2x.Cloud Computing & Storage: We expect a push to support petabytes of data used by servers and storage devices. System chip architectures will continue to morph to keep the need for low latency search and indexing for databases and may include accelerators that are more hardwired to the end application. Machine Learning (ML) & AI: Digital architectures in silicon are undergoing a phase of rapid evolution with the advent of neuromorphic microprocessors configured more like our brains than today's von Neumann architectures. Intel, Nvidia, Qualcomm, and the likes continue to push the boundaries here.Security: Hardware security continues to be a major point of concern in IoT devices and requires deeper, integrated silicon support to prevent unauthorized access.We see a trend for custom ASICs, both digital and mixed signal from several non-semiconductor companies, which can now leverage Cyient's end to end ASIC and embedded software services. The primary reason is to ASSP's again is to fine-tune the performance for their end application also, the pricing has become more affordable for doing this today as compared to 15 years ago.IoT is one of the major technologies disrupting almost all industry verticals in recent times. Tell us about the role of the semiconductor industry in this regard.With over 50 billion devices, connected by 2020, IoT presents a huge growth for semiconductor companies, and for many companies becomes a critical component for asset management especially around predictive maintenance and asset health monitoring. Every end application is different in IoT and has different requirements. It will become imperative for semiconductor companies to move up the stack and offer generic solutions that can be quickly customized by partners like Cyient. The semiconductor industry will need to continue to produce faster and cheaper components to enable IoT growth on Edge devices where we expect rapid commoditization. Differentiation will come from system solutions and new areas such as ML and cloud computing.How can semiconductor technologies be integrated with AI, SMAC, Robotic Process Automation, or other niche industries to utilize their maximum potential?Semiconductors (SOC's MCU, analog, memory, and connected products) will play a big role in enabling IoT for industrial automation, transportation, and smart cities--utilities and other services. However, in areas such as AI and ML, semiconductor architectures will need to be re-imagined to solve these problems given the advent of new neural network processors and neuromorphic processor roadmaps.
< Page 8 | Page 10 >