Written by Srini Chinamilli | CEO | Tessolve
On Deck for the Semiconductor Industry: Managing the Conflicts of Global Disruption vs Time To Market
The Covid-19 pandemic caused havoc with the world economy as it ravaged the world population. While the semiconductor industry survived reasonably well, the industry was not spared the impact of the virus. The semiconductor ecosystem must now contend with disruptions to supply chain and labor, in the form of both shortages and remote workers. However, there’s been no interruption in the clock running behind the scenes that drives demand for uninterrupted innovation and aggressive time to market goals. Regardless of any COVID-related disruptions, chips keep increasing in capability and complexity. How the industry responds to these push-pull conflicting forces will determine its economic growth rate during the coming years.
A Current Position of Strength
Despite the Covid-19 virus, the semiconductor industry grew by more than 6% in 2020. In 2021, analysts predict a banner year with nearly 20% growth. While the industrial and automotive markets suffered sharp declines in 2020, the consumer market supported the semiconductor industry with increased demand for PCs and smartphones to meet the needs of remote connectivity for both work, student education, entertainment and purchases of products and services via e-commerce. The interconnected and IoT-enabled home has become a self-sufficient entity.
This increase in semiconductor content, as everything connected to work and home becomes digitized, continues to act as the major growth driver for the semiconductor industry. Semiconductor content growth seems to be all-consuming and insatiable:
- Smart home electronics with lighting control, environmental control and security devices. Smartphones have multiple cameras and more sensors.
- All types of automobiles, including internal combustion vehicles as well as hybrid and electric vehicles. The latest vehicles can have as many as 4000-5000 semiconductor devices. The electronic systems powered by semiconductors represent about 35% of the total cost of the vehicle. For example, Tesla has around 70 processors in its electric vehicles. In the next 10 years, the percentage of semiconductor based electronics could represent up to 50% of the vehicle cost. Autonomous driving, safety, efficiency, infotainment, and connectivity are triggering the greater electrification of vehicles. Future vehicles will employ gesture recognition and augmented reality displays. We are already seeing collision avoidance technology. We will soon see vehicle-to-vehicle communication and vehicle-to-infrastructure communication for traffic congestion management, weather information, and road safety.
- Beyond consumer demands, infrastructure requirements for semiconductor will fuel double-digital growth for our industry, as seen in new wireless telecommunication technology, 5G, Industrial IoT devices, medtech, artificial intelligence (AI) and big data processing. These technologies will impact a wide range of industries such as pharmaceuticals, healthcare, agriculture, energy/climate control, transportation and cloud computing. These technologies will enable the smart city which will facilitate such services as safety and traffic control, for example. The smart factory will evolve as AI permits machine learning and IoT devices monitor all aspects of machinery performance. AI will overcome the limitations of humans to process large amounts of data. AI combined with big data will drive greater efficiency and help companies reduce costs and downtime. For example, systems will be capable of determining a machine’s root cause of failures or defects in real-time. Big data will permit companies to identify customer buying characteristics and lead to improved product development decisions.
These advances are what we can project today. We can only contemplate what future technology might bring us.
Meeting the Call for Accelerating Semiconductor Innovation
Implementing these advances will require more complex chips along with process advancements to make them. As Moore’s Law reaches the end of its long, illustrious life, improvements in computing power, reduction in power consumption and size reductions can no longer occur simultaneously. Semiconductor R&D will focus on 3D structures, novel materials, new ways to shrink features, new multi-core technologies, dynamic power management and new packaging technology. The additional R&D challenge we all face is the need to develop new technologies that can work with existing processes and can work with existing upstream and downstream operations.
New computer chips for applications, such as AI in particular, will need to be capable of processing massive amounts of data and handling the large amounts of energy they’ll require. The industry will need to significantly improve computing capacity-per-Watt for these new AI processing chips. In addition, faster workload-specific hardware will facilitate such processing-intense activities such as computer vision, pattern detection, facial and voice recognition and language processing. Above the chip level, new computer technology and new materials will be needed to process these massive amounts of data. The need will exist for faster application-specific hardware built from customized and new materials technologies.
One approach being adopted is mixing circuit blocks with different process node technologies into a single package. This strategy, the chiplet strategy, gives designers maximum flexibility to optimize a chip design without having to scale all components to the smallest node technology. This new type of chip design challenges R&D engineers to revise design-for-test techniques. Process engineers need to develop new manufacturing technologies for fabrication, and the packaging engineers have to adapt packaging procedures for the chiplets. In addition, test engineers must adopt their test technologies for wafer test and performance verification.
Another developing technology is chip-on-flex in which a die is mounted directly on a flexible substrate. The technology reduces the existing steps required for IC packaging and creates opportunities for making smaller products such as sensors, implantable medical products, and RFID devices. This technology will be employed in creating smart textiles.
The automotive market and the medical market, particularly implantable products, will require zero defect design goals. Safety is critical for successful adoption of autonomous vehicles. Innovative technology will need to be developed to achieve ultra-high chip reliability.
This level of innovation doesn’t come without cost. Certainly, increasing chip density and complexity raises R&D costs. For example, developing a 7 nm node increases chip development costs by 25% to 30% compared with development of a 10 nm node wafer. With new technologies, mask and layout costs have increased dramatically for new tape outs. Greater complexity requires more tape outs.
Battling Time-to-Market Pressures and Workforce Issues
While every industry has time-to-market pressures, time-to-market is acute in the semiconductor industry. Missing a technology node milestone, for example, can result in a tremendous loss of return-on-R&D investment. To achieve the development timetable goal, companies are increasing the sizes of their design teams, along with relying on outsourced design-and-innovation providers such as Tessolve.
Another way to reduce time-to-market is to reduce inefficiencies in the development process by expanding automation to decrease the number of manual tasks. It’s not practical to deploy endless staff resources to drive launch timetables; companies will have to increase automation in the coming years to contain the rise in R&D costs.
Another major challenge is having enough engineers to address product development plans. Currently there is a finite and even insufficient number of engineers available to the semiconductor industry. This is a worldwide problem as companies are not able to fill their ranks with the number of engineers that they need. Semiconductor companies are competing with system companies such as Facebook and Google for talent. The market is currently a seller’s market, as engineers are looking for the most challenging work, top culture/work environment and of course compensation. Thus, retention is a critical factor in maintaining and achieving needed employment levels.
The Covid-19 pandemic has accelerated changes in workforce dynamics and increased technology advancement in companies. Changes have been occurring at a rate of three-to-four times the normal pace of changes in company evolution. Whereas companies felt that all work had to be done at the company facilities, management has developed ways to allow employees to work remotely. Managers are more open to work within the locality of their R&D facilities as well as across geographic regions. Engineers have adapted to working remotely and are probably more efficient working outside the company due to fewer interruptions. Most engineers, however, want a hybrid arrangement where they can work at the company facility, typically, two days a week, with the remainder of the work week conducted at home.
Remote or hybrid work poses challenges for company leadership and management with regards to sustaining a company culture. A company culture forms when employees directly observe how supervisors and other leaders work and manage. A positive culture improves employee morale and improves retention rates. In the near term, company management will need to find methods to integrate their culture with their remote work force.
Restoring the Supply Chain
The Covid-19 pandemic has created discontinuities in the semiconductor supply chain. The automotive industry was hurt badly by the Covid-19 pandemic as sales dropped by 15%. Automotive companies responded by cutting costs, which resulted in substantially shrinking inventories.
Then, in the latter part of 2020 and in 2021, the automotive market started recovering with an expectation of 8 – 10% growth. The recovery has created the well-documented shortage of semiconductors as the semiconductor industry attempted to quickly pivot production schedules back to meeting the demands of the automotive production re-start. However, the shortage of components will likely impact automotive market growth in 2021.
The automotive industry is one portion of discontinuity in the semiconductor supply chain. The other major factor is the current state of geopolitical tensions, particularly China vs the US and the rest of the world. That is causing companies to increase their raw material inventories and finished goods inventories. Companies do not like uncertainty. Their risk assessment analyses dictate a higher-than planned inventory position under the conditions of global tensions and low cash borrowing rates.
The semiconductor industry initially responded to low automotive production volumes by switching capacity to meet the rising consumer demand for connectivity devices and for the 5G market roll out. Facing the uncertainty of the pandemic, semiconductor companies, like most other companies, delayed expanding capacity even though the semiconductor market experienced growth in 2020. In 2021, with analysts expecting solid market growth of as much as 20%, supply chains will continue to be stressed until expanded capacity comes online to address product shortages.
Fabless companies are hedging their bets on finding available production capacity by designing product processes for multiple contract fab houses. This work contributes to both higher R&D costs and to requirements for more engineers putting further strain on companies to meet time-to-market targets.
Reason for Optimism
While the industry grapples with the effects of the recent economic downturn and a pandemic that has not yet ended, the outlook is still good. Chip innovation, new market technologies, and the increase of semiconductor content in products will continue to propel our industry. The challenges will be overcome, as they always are, and even the innovation promised by Moore’s Law will continue to live on.
___________________________________________
About Srini Chinamilli
Srini has over 25 years of experience in Semiconductor Engineering and Management. He held technical and management positions at Cirrus Logic and Centillium Communications prior to joining Tessolve as Co-Founder. He has extensive experience in Silicon Validation, Product Engineering and has managed high volume productization of several complex System on Chip and Mixed Signal devices. As a Co-founder at Tessolve, he helped grow the company from a startup to a world class engineering solutions company with over 2200 employees worldwide. Srini got his Masters in Electrical Engineering from University of Southern California and Bachelors in Electronics Engineering from Birla Institute of Technology.