How has the pandemic changed the game for semicon majors like Renesas? How do they view the acquisition of Arm by Nvidia? To get answers to these questions, Rahul Chopra spoke to Dr Sailesh Chittipeddi, EVP, GM of Renesas’ IoT and Infrastructure Business Unit. Here are the extracts of their discussion.
Q. Can you share some of the key strategies that have been outlined after Covid-19 and how things have changed?
A. One of the biggest trends that we see is that there is a big push towards more bandwidth consumption. For example, Netflix or web consumption, Facebook or even text messages—most of these have gone up by over double-digit percentages in a matter of months. That has driven up quite a bit of data centre spend for major players like Google and Amazon.
There has been an increased focus on healthcare and the environment, whether it is for lowering the SpO2 level or measuring it. There is a correlation between the SpO2 level and Covid, if your oxygen levels drop alarmingly. So, that has driven interest especially for bio-sensing and ventilator systems.
With the move towards electric vehicles, less polluting activities have taken hold. On the industrial side, certainly smart metering brings more efficient energy consumption. Surveillance and voice user interface have been big and will grow as people are moving away from touch buttons. So, these are some major trends we have seen. Fundamentally, the push is towards data-centric, all the way from front-end to back-end.
Q. Has that resulted in launch of new products also, as in, some products being pushed and others being put on the bench, or has the usual production increased as per greater demand?
A. A little bit of both. The demand has certainly increased. The need for laptops and personal computing has really gone up dramatically. Even in the USA, there are a lot of regions where school children cannot afford a laptop or a notebook. So, there has been a big push for Chromebooks, for which districts have been placing big orders.
In terms of investment, we have emphasised a little bit more on things like voice and surveillance and are trying to get some of the products out a little bit quicker. For example, we have pushed our 48V scooter and e-bike solutions a little more aggressively than we would have in the past. Same for the ventilator. So, I would not say that we have slowed down on anything per se, instead we accelerated somewhat to take advantage of some of the six scenarios that we are seeing.
Q. How did Renesas manage to launch products so rapidly during Covid-19, while most brands have been silent? What best practices did you follow to ensure new innovative products kept coming out?
A. We were trying to make up for the lost time. We were late to launch the Renesas Advanced (RA) Arm-based microcontroller family in 2018. So, we launched that MCU family in October 2019 and followed it up rapidly with a whole set of RA MCU products to significantly expand our portfolio. And then we went out and addressed the gaps that we had in our portfolio early-on in a hurry.
One of my messages to the team has been that customers will always remember how you have served them during challenging times. Because challenging times always allow for the formation of best relationships, whether it’s between people, or between organisations. Another message has been that this was the time to out-serve the competitors, whether through new products or a better supply chain.
Coming to Arm being acquired by Nvidia, how is that affecting the Renesas strategy now?
First, I would say there is a finite possibility against the deal occurring as the regulatory authorities may have concerns. But assuming it happens, and based on the discussions we’ve had with Arm, they have been pretty clear that they are still going to be on an open-licensing model.
We recently announced our first implementation of RISC-V-based solutions, which will begin sampling to customers in the second half of 2021. These MCUs are going to be application-specific solutions through which we will attack motor control and voice markets. Beyond that, we have proprietary 8-, 16-, and 32-bit MCUs as well. We have the RX and RL78 families, which have been around for a long time. So, our customers do not have anything to fear major disruption.
On the other side, there are other GPU providers, such as Imagination Technologies, who do provide cores. So, if we need to pivot, we could on the MPU side. But we don’t feel a compelling need to do so at this point of time because Arm as well as Nvidia have been very clear that both will support an open-licensing model. But, obviously, you don’t want to be caught working with someone with whom you are competing in another area. That brings a bit of a disadvantage. So, we need to be careful. I must add that we had already taken up some steps to come up with some RISC-V-based products, even before the Nvidia news was announced.
Q. That brings up the question, how do you see the future of RISC-V?
A. I think people wanted an open-source architecture anyway, primarily because the licensing fees were going up. So, there was a natural move towards that direction. And I think that the RISC-V movement, thanks to such a deal, is going to accelerate. The advantage that the Arm ecosystem has and will continue to have for the next several years is really a huge ecosystem of support that has been established around it. No matter what, it will still take time for any organisation to develop that kind of infrastructure support. It will take a while for anybody to replicate it. If Arm is true to its word of being an open-system architecture, then that is no big deal. But it’s always good to be careful since we as companies are at the receiving end of things.
Q. Where all are your R&D teams spread? Is there a team in India too?
A. Earlier there wasn’t but now there is an R&D team in India. We are establishing a footprint in Bengaluru primarily for RF and timing devices that have very good capability. And we have leveraged our partnership with a company called Steradian. We have always used India-based design outfits and worked with companies like Tata Elxsi, HCL, Sankalp, and TCS. We have decided to establish a good footprint in India.
Is there a roadmap for India that you would like to share now, or is it under progress?
For RF, we certainly have a very compelling roadmap. A lot of our millimetre wave activities are currently done in India. By millimetre wave I don’t mean sub-6GHz, but millimetre wave that is 28GHz and above. We are also working on some of the timing activities in that area and looking at setting up a stronger presence in Bengaluru for that activity. Radar activity also goes on there.
Q. Are e-bikes and surveillance the two key markets or are there other markets also wherein you see some growth?
A. There are other markets too where we see a growth in. Bio-sensing is one. Due to Covid-19, getting low-cost healthcare to people and air-quality monitoring is going to become much more important. Our team is going after a range of customers in IoT, mobility, medical, consumer, infrastructure, and industrial sectors. It is quite exciting.
My view is that India is an untapped market in terms of electronics reach. And the potential is just enormous. So, in addition to the R&D, for which we have established a footprint, we have plans to work with emerging companies like Ather, besides the established where we have already established footprint. The reach is going to be quite significant and, if India sets its mind to build the white boxes themselves, I think it will be a massive gamechanger for us. If it starts to do some of the things that are done in Taiwan, including the product design specs, then it will propel us to the next level.
Q. Do you see IoT as a buzzword or a real market opener?
A. That’s a good question. To me, any edge or endpoint device is fundamentally IoT. Are more devices and more sensors getting connected across the globe? Yes, absolutely. Without a doubt, we are marching towards one trillion connected devices in a not too distant future. It doesn’t matter to me whether you call it IoT or connected sensors. The trend is already there.
For us to capitalise on that market, it boils down to having all the intellectual property that is needed for it. Companies will not be able to go very broad in this area, and that is the fundamental challenge in IoT. There’s no single company that is going to dominate every vertical, because not only we need to have the sensor solution, we also have to be able to interface it to something else in order to get it to operate. If it’s a smart home, whether it’s a Google Nest thermostat or something else, it needs to interface with another device and the software. So, you have to decide the verticals you are going to go after. Like in India, we are going after smart metering. Motor control is another great vertical that we are pursuing.
Any particular geographies like Europe or the USA that have taken lead in industrial IoT?
For industrial IoT, the areas that are doing well are Japan, Europe, and then the USA. China has an interesting growth story for a variety of reasons, including the fact that they came out of Covid-19 crisis much earlier than a lot of other regions in the world. One of the areas that they do continue to surprise a lot of people is of AI. The dialects in China (certainly not as complex as in India) are vastly different. They have AI companies like iFLYTEK that are really advanced when it comes to voice recognition-based AI technologies. So, China is probably far ahead of India when it comes to industrial IoT, although India is starting to pick up.
Most of the advanced robotics is largely happening in regions such as China, Japan, and the USA. The traditional base of industrial customers has largely been European, Japanese and the American to some extent. However, China has been increasingly very aggressive in terms of industrial companies that adopt new products quickly. The advantage that they have is that they move extremely fast and don’t require the 10/15/20-year lifetime reliability cycles, which the traditional players require. They understand that most of these products would not last more than 7-8 years before you will have to upgrade them to the next version. So, watch that space as there is going to be a big set of disruptions including digitisation, which will get hold of that market over the next 5-10 years and major new players will emerge.
Q. Having seen the entire silicon ecosystem from pretty close, do you think India should go ahead and invest in a semiconductor fab or remain fabless?
A. India has tried investing in fab earlier as well and this isn’t the first time. If we take a look at the history of countries that have been successful in introducing fabs, such as Taiwan who started investing in the late 70s and early 80s, it took thirty years of learning to become what they are now. There is no short-circuiting that process. They have the companies and supplier base, despite being located on a small island, which is a fraction of a vast country like India.
China has repeatedly tried to establish fabs, with far more resources than India, to be successful in the wafer fabrication area. And they are still several decades away from getting there. So, the question is: Do we have the long-term vision to make US$70-80 billion worth of investment? It is critical to identify the vertical after which you want to go, such as communications and enterprise infrastructure, and then start establishing a company that will build the systems.
If it’s for 5G, then there is a perfect opportunity with Open RAN systems to establish a company in India that could set up their own telecom infrastructure. Open RAN is hugely disruptive versus traditional telcos. If you do that, then you will naturally start establishing fabless players and others that will support that ecosystem and set up the entire manufacturing system for it. But, without identifying the vertical to pursue, you won’t establish a strong footprint. The semiconductor industry in India will be challenged. Huawei became what it is today because they built the boxes, and that set up Hi Silicon, which led to set up other companies.
To read the full interview, click here.