What's in store in Electronics for the next 5 years?
What's in store in Electronics for the next 5 years?
In this episode of Global Electronics Hub Podcast, we interview with Dr Dishant Mahendru, Director of Honeycomb Global UK & India. We discuss the current state-of-the-art in technologies, upcoming disruptions to watch out for, how to break into the market through building a strong competitive technological advantage, how should one adapt towards technological disruptions and what new breakthroughs can we expect over the next 5 years. Please note an edited transcript of this podcast follows.
“Creativity is fantastic, and it pushes innovation, but my only advice is that technologies that are opening up digital business opportunities specifically, it is fantastic, but in order to really and realistically benefit from these business opportunities it is really important to confirm what are those customer demands, what are those white spaces, to showcase the value of the innovation instead of still conducting exceptional amounts of R&D and push ideas to the public. Because at the end of the day it is only successful if your actual customer, whether it’s B2B or B2C, sees value in it, for it to become a viable business.” - Dr. Dishant Mahendru.
[00:00:00] Ranjana: Hi, good afternoon. This is Ranjana Pathi from Honeycomb Global India. And today on our podcast, I have Dr. Dishant Mahendru, who is my partner in Honeycomb Global. Dishant's got a very technology heavy background and always leaves a mark wherever he goes, with his knowledge expertise, so to say. Dishant has a PhD in nanotechnology from the University of Cambridge, and he also does technology and innovation consulting in his other firm in London.
[00:00:29] So, welcome.
[00:00:30] Dishant: Thank you Ranjana.
[00:00:32] Ranjana: Today we're going to speak about what's in store for electronics for the next five years and the various new technologies that are coming up. Disruptions to look out for and how there's a lot of disruptions coming from outside of the industry as well, and how anyone within of the industry should adapt to these disruptions and ride the wave; and what you're most excited about.
[00:00:58] Dishant: Really excited about today's conversation. I think, you know, some of the ideas from the call today will not just be exploring technologies that everyone's been hearing about such as on the digital end with machine learning or various different types of sensing frameworks like IoT, but also more on the materials and nanofabrication side of things, which is what, you know, companies like Intel are currently pioneering as well, in terms of Moore's Law at some point stagnating.
[00:01:26] Ranjana: Yes. So according to you, what are the various new technologies coming up in the fairly medium term? So say five years .
[00:01:36] Dishant: I think what's happened with the electronics industry over the last five years, five to 10 years is that a lot of the solutions have been exceptionally application driven. And that's primarily because a lot more day to day products now are becoming smarter and therefore require dedicated chip sets as well. Now, one of the main areas of technology over the next five years, which will continuously grow in terms of importance is the adoption of higher computing, algorithmic processing systems like machine learning, but within the ability to apply that into a low power chip set.
[00:02:13] And that's actually an area of innovation that both academia and industry are currently exploring as well, because even though everyone's exploring the sort of improvement of density for battery storage for smaller devices. It is also down to having more efficient electronic systems around that, to not just have a much more energy efficient system with the current batteries that are being used, but at the same time, being able to allow faster, computing power to open up more use cases for your average sort of consumer product. That's one area. But then that's looking more on the digital side of things, but then if we're looking at, say on a semiconductor level, what's becoming a growing issue in the semiconductor sector is the fact that at some point Moore's Law will hit a plateau.
[00:03:07] And that's an issue generally because a majority of electronics or a majority of our sort of standard products that use electronic systems, has been very much dependent upon the performance from that nano fabrication level. However, there has been some innovation going on now in terms of how to potentially improve Moore's Law even further.
[00:03:30] And a lot of the sort of primary transistor sets at the moment have been, say, fin FET based on a mass scale, but what's happening now also is we're also investigating monolithicstacking, which is looking more towards making transistors architectures at a three-dimensional level. So pretty much stacking to transistors on top of each other as well and being able to identify which transistor at what level to activate and how to transfer data throughout it, say a nano wire at the end of the day. So it was really exploring the third dimension now, but of course the fabrication techniques at the moment is still at a very much on a prototype level. And whilst this is still happening in terms of innovation, the semiconductor companies, as I think we've all seen over the last five years, have actually expanded their business now towards providing more digital services surrounding chip sets that they've designed themselves for specialized components, such as wearable technologies.
[00:04:30] So we're looking at, say for example, CASE vehicle industry as well. So really kind of building chip sets for specific applications of industries where convergent patterns are really growing. So, as I said last two, but also say smart cities, for example. So a lot of these sort of smart solutions that popping up. Semiconductor organizations are getting into that sort of industry of providing a full package of not just a chip set, but also a digital component on top of that as well and operating that chip set as well for that targeted application. And that's something that will consistently grow as well, because there's a lot more smarter devices that are coming up and that's pretty much something that's going to happen for a variety of different industries based upon trends, such as automation and the adoption of various sort of IoT sensing frameworks as well.
[00:05:18] Ranjana: Okay, great. So what are the disruptions to watch out for?
[00:05:25] Dishant: I would say that what has been quite interesting is everyone's being eyeing on quantum computing for the moment. But quantum computing, for now has been great for a couple of interesting applications, but over the next five years, that's not where the realistic disruption will be.
[00:05:44] Potentially on an investment side, fine. But in terms of replacing supercomputing capabilities or replacing human machine learning or high processing machine learning capabilities, and we just won't get there just because of the price point of course how expensive the architectures are at the moment.
[00:06:00] However, in terms of disruption, disruption still primarily will be through making more energy efficient electronic systems and not at the semiconductor level. I think on the materials front and nano fabrication side of five years, there'll still be a lot of innovation going on. But it won't be disrupting the sector for, for electronics. It'll pretty much be looking towards more energy efficient systems, how that ties in with high density battery systems as well for consumer products. And at the same time how to integrate better algorithmic systems for low power electronic designs as well. And that's primarily, another area of interest that has been growing significantly is how to look at traditional electronic components like for example, sensing electronics like MEMS and how to change their primary purpose of being a sensor and actually being more of a computing transiting system as well. So looking at smarter ways of looking at day-to-day electronic components and how to repurpose them for other applications. All of a sudden there's a lot of innovation going on in that front as well, because at the end of the day, that comes down to really the whole sort of chip set design and not really the sensor itself, as well as the signal processing algorithms that come into play as well as the machine learning algorithms that can go into that also.
[00:07:26] Ranjana: Great. Okay.
[00:07:28] When we also said about, we talk about all the disruptions that are coming from outside of the electronics industry, say the IT sector or there's MedTech and FinTech, which are also going to be quite big and there's some sort of interplay, right? What's your advice for players who are trying to break the entry barrier within of the electronics sector as well as for players outside of the electronics industry who need this for their business to thrive?
[00:07:55] Dishant: I think partnerships are a must. I think, all the industries now are converging. And two, they're overlapping with each other opening up way more larger pockets of competition. So an example is if you look at the mobility sector, that's a convergence of say six, seven different industries, and that gives the players and those individual industries, their purpose for that converging trend but because it overlaps it also, and you know, these industries are performing well they have capital for further expansion to new industrial and technology areas outside their comfort zone, it opens up more room for competition. Now that's fine. But at the end of the day competition, isn't just really about who gets into the market first, it's about what their product or service will be because they are competing with many other companies, right? So therefore one of the big solutions to this would be to formulate partnerships. So it's clear example is what Apple has been doing very recently. So the business sort of relationship they have with their semiconductor companies have been literally, you know, they buy their chips, they subcontract their designs to these companies, these semiconductor firms. And that's it. But now what they're doing is they're actually partnering up there with the semiconductor companies to further innovate as well. So recently as well, you know, I think Apple's and one chip now officially has gone into mass production with the, Taiwan Semiconductor Manufacturing Co. and what's interesting about that relationship at the moment is it was a partnership, but also as a partnership where both organizations innovation skills. So Apple's sort of strong pedigree of low power electronic systems, but also Taiwan Semiconductor Manufacturing's innovation and around materials, chip design, and chip fabrication, allowed them to easily formulate a very strong relationship to automatically accelerate, chip manufacturing or better chip designs for Apple products as well. I think this is a really easy way of non-electronics companies getting into the space and instead of actually disrupting it, it makes more sense to partner up with a large player there.
[00:10:05] Another way of looking at it is that a lot of companies sometimes feel that, oh, we might as well just invest in R and D, invest in innovation, invest in business development to access this new market. But because you're not the only sort of competitor out there in terms of whether you're an SME or multinational corporation, it is much easier to formulate a strategic acquisition. Or potentially merger, with a partner who is very strong in that space to quickly get into that space as well. So for example, IT companies who really want to expand their business into electronic sectors, that there'll be a lot more M&A work going on in this space over the next five years. And that's whether they're hiring a startup or potentially merging with an SME or a larger corporation.
[00:10:52] Ranjana: Okay, this leaves me with another question.
[00:10:54] What are you most excited about for the next five years?
[00:10:58] Dishant: I think what I'm excited about specifically is the adoption of high computing, algorithmic systems in low power chip sets. Again, I know I've kind of been repeating this quite extensively. I think one of the biggest issues right now with a lot of products at the moment that are non-electronic, but they can be smarter. And that's pretty much where a lot of innovation is going on in the consumer electronics space. The opportunity of doing this specifically will crack open multiple use cases. So an example is if you look at, say a company like Whirlpool, so they design a lot of sort of household kitchen appliances. But if you look at their innovation roadmap, they are now trying to make a lot of the appliances smarter and more interactive as well.
[00:11:47] These elements had very basic electronic systems at the end of the day. Trying to make them smarter will require a lot more complex electronic systems on top of that as well. If they're able to do that with a low power functionality, because everyone's following ESG trends, the whole sustainability trend, reducing their carbon footprint, generally. And doing that with energy efficient systems, but also integrating smart, you know, integrating smarter machine learning algorithm systems to open up new use cases and say, for example, if you look at an oven and, you know, oven with a smart screen and the smart screen would start to tell you, say the internal temperature of say the food that's inside, then it should provide other sort of suggestions or the temperature can itself just regulate consistently on its own and so forth. That itself is going to be a useful element that will support the demands of the consumer, for example. So I think just to kind of summarize that making day-to-day items that have had premature electronic systems and making them smarter through machine learning algorithms will be something really exciting at the end of the day.
[00:13:00] And that's pretty much where a lot of the road, that's what you see in terms of the roadmaps. A lot of sort of basic product service-based organizations today is to make a lot of their sort of premature electronic systems, a lot more complex, but cracking open new use cases.
[00:13:15] Ranjana: Okay. Great. So I have a, a question with, you know, how every time there's a smart device or a smart system, there's also a lot of analytics and everyone wants to acquire data and make things further smarter. So with this increase in the data mining and analytics, and even 5G computing, 5G connectivity, one of the very relevant factors is the increasing need for data centers and also the the data centers that exist, as well as say take the Bitcoin mining, all of these poses such a heavy burden on the world's electricity output.
[00:13:52] Do you think there's enough being done in the green energy space to match this increasing energy hungry growth in technology?
[00:14:02] Dishant: That's a great point. So I want to, I want to highlight that there's actually two major issues going on right now. One is, the abundance of data that's being gathered for the sake of data.
[00:14:13] And I would like to highlight these are data lakes that are unstructured. This is a general issue that's happening consistently in any industry where someone has for, they made an executive decision of just gathering data for the sake of it. Right. That's one issue. The other issue is because there's a huge sort of demand now for cloud based computing systems, therefore, you know, because it's not really in the cloud it is actually technically in data centers that are owned by a lot of the sort of tech giants. Of course you are right. The energy demand for such data centers drastically increases not just because of computing power, because a lot of people are online on the cloud and using their systems on the cloud. On top of that, just the electronic systems generate a lot of heat and of course the energy prices skyrockets. So, a couple of steps that a lot of these companies are doing are the following: (a) instead of cooling towers or cooling tanks, they're using, you know, they've kind of moved along their server farms into colder climate areas, such as parts or Russia as well, where naturally the temperature around it will automatically cool the data center itself. So Facebook, Google, Microsoft. A lot of them have a lot of their data centers now in those sorts of regions. Another interesting test that, has been looking in terms of natural cooling has been underwater data centers. And a lot of the tech giants have been exploring this as well. And it's also quite interesting because another additional benefits they found is that it's also less susceptible to other factors such as vibrations, noises, other types of pollution surrounding, a data center, if it was above ground. So that's one interesting thing.
[00:15:50] And of course, what's great is that a lot of the sort of general verticals, all of the sort of industries that are out there have set very sort of ambitious ESG goals for the organization. And I think, what's been really good is that a lot of the tech giants specifically will be carbon neutral very soon over the next five to 10 years. One of the major reasons why they're doing this is because they're supplementing energy demand with, renewable energy, whether that comes from a third party source or they're setting up their own sort of solar farms.
[00:16:21] But yeah, I think, you know, just to kind of summarize that in terms of data centers there at the moment. Yes, energy from data center is currently being driven by renewable sources. That makes sense. But also what's really exciting is that they're looking at other ways of naturally reducing energy consumption through as a highlighted, moving data centers, into colder climates, such as in Russia. But also looking to underwater data centers as well, which is also quite exciting. So that's one issue. The other issue, which is huge at the moment is the gathering of unstructured data from any industry, whether it is the automotive sector, whether it is the energy sector, the utility sector, whether it's the infrastructure service sector, the, you know, every everyone's gathering data because they've just been told they need to gather data. The biggest issue right now is how to structure that data and how to apply some really exciting machine learning algorithms, which provides really good analytics, whether it's current analytics or predictive analytics or historic analytics that provides benefit to the organization internally to improve internal operations or to provide new products and services to their current customers who really would benefit from such analytics as well.
[00:17:38] And that's something that you know, the last 10 years everyone's been kind of building up data assets is always fun and super exciting, but the biggest issue has been, what do they do with it? What are the use cases? Now? The innovation of the next five years will be to really understand what are clear-cut use cases there'll be beneficial to the, to the company internally, but also to their customers at the end of the day. And that's a really exciting area of innovation generally.
[00:18:05] Ranjana: That sounds great. Another point, which, you know, there's commercial, nuclear power has also started gaining steam. Well barring the regulatory hurdles, right, that's going to be different in different regions. And the speed of the technology coming to market will be different based on that. But where do you see this technology steering towards? Do you think it's going to be ready in the next five years, 10 years. When do you see us using it?
[00:18:31] Dishant: I think, small nuclear reactor projects have already kicked started in, in the Western world. For sure. We can already see Rolls Royce launching their small nuclear reactors around the UK, for example. So I think on a small scale, yes. I think large scale systems are still not as energy efficient as we wanted them to be, and of course the CapEx for such entities are exceptionally expensive.
[00:18:57] That's what you mean with the small nuclear reactors. It only works if you get some subsidies to support that as well. I would say that in order for it to be an alternative to other renewable sources, especially with the CapEx issue, I don't think it's going to be something that's going to be disruptive or it's still over the next five years, we will see a lot more smaller nuclear reactor systems pop up. But to be really honest with what's happening with wind energy, and solar at the moment, where these industries have gone through this process already, they have received lots of government subsidies for quite some time. And the industries are now becoming self-sustainable. There'll be a lot more disruption involved in those spaces to allow a quick sort of transition of dependency to those types of renewable sources where nuclear will hopefully in the future follow up and it'll be a lot more appropriate over the next 10 years, especially in larger nuclear reactors.
[00:19:57] Ranjana: Right. Okay. So another question I had was with regards to wearables, and lab-on-a chip or Bio-MEMS. Well this is going to be quite disruptive as well, obviously, because we've not had anything of the likes so far. How do you see material science catching up with this? Do you think it's on power, or would we be stopped by that?
[00:20:21] Dishant: That's a good question. I don't think materials are a limitation, in my opinion. We're not going to replace Silicon anytime soon as a fundamental chip layer. There is a lot that really has to do with on the price point. If you look at electronics right now, it is technically a volume based industry. Right. I'm making a very small margin on each chip that you manufacture. So, you know, you wouldn't really change any of the materials unless on a price point, it was identical and performance was better or it was cheaper and therefore it make, and the performance was at least the same. So in terms of the materials front, I don't see much going on. I think a lot of the sort of innovation to improve the price point and performance will be down to the architecture and the fabrication of the architecture. So therefore highlighting what I said earlier, the 3D stacking will be what will push Moore's Law for sure over the next five years instead of materials.
[00:21:14] Now, if we go back to the concept of wearables, generally the innovation there specifically, won't be really around the materials and the stacking side, but also the formation of chiplets. So, you know, right now we have monolithic chips, which is literally like, you know, you have a single piece of Silicon and then you construct all your transistors or your processing systems on that one piece. Chiplets will be literally a chip that has all the major components that's required for a computer to work. So you'd have your CPU cores in one area, you have the memory in another area, then you have your graphic units in another area and comms and another area. So all the major components required for computer will be on that chip. That's in a really exciting area which will push. Again, more applications out there into the market. And therefore also looking into wearables where certain chips are absolutely a must. It's literally a mini computer at the end of the day for a specific application purpose. And that's mostly the reason why wearables hasn't really blown up really as a business. A couple of reasons why. One is that (a) the electronic systems weren't strong enough or they, you did not have these chiplets before. Two was, you know, people did not really understand what the users want in terms of the use cases and therefore creating really dedicated application and solutions for those use cases. But I think what's happening now is that people are realizing that on the healthcare side of things, when you're looking at, individuals that unfortunately have a lifelong disease, having some level of support and real-time support, i.e., and this is where this is where the wearable comes in, will be exceptionally beneficial for patient management, patient engagement as well, especially seeing that people don't have time to always go to the doctors on a daily basis. It's more important to bring the doctor to them at their own convenience, but also to have a wearable that pretty much does the basic monitoring that not just supports your needs, but also the doctor's needs as well in terms of gathering data in a real-time fashion. So I think with more wearables in the medical field, like that's where the adoption will really blow up now because (a) the chiplets are popping up and wearables are becoming a lot more computer heavy, which is great for that application.
[00:23:35] And also the use cases are better understood where better applications are being developed for that. Another interesting area, which will definitely grow, is the ability to reduce the price of medical devices generally. So as we see today, large medical device companies like Phillips, like Abbott, for example, have developed medical devices that are still quite expensive, but that's because their business model and their clientele have been more on the professional side of things than the consumer when it comes to medical devices specifically. And now what's happening is a lot more companies are now trying to disrupt the medical device space by reducing the cost and looking at home-based medical devices as well for the general consumer, which will be great for not just Western markets, but also, Eastern markets. And of course, developing countries where they'll be they'll have access to medical devices at a price point that is affordable as well, and that's also another interesting area for disruption, because at the end of the day, the innovation, the space is becoming really exciting where you're looking at various types of nanotechnologies that are replacing your sort of standard state of the art sensing mechanisms for your basic medical devices, such as your average pregnancy test or your average, diabetes test, for example.
[00:24:58] Ranjana: Okay. So another question I had was with regards to 5G and you know, how well, technically it would impact every area that we've spoken about so far and specifically, maybe industry 4.0 and internet of everything. Do you see the impacts to be realistic in the say, five year roadmap?
[00:25:23] Dishant: I would say so, I think 5G technologically has confirmed that it does what it promises: it increases the bandwidth for large data applications. It definitely has low latency communications, and also, you know, it is much more energy efficient as well. At the end of the day, in terms of availability at the moment, a lot of products have been popping up of course, due to political reasons as well. The adoption of 5G in various Western markets has significantly slowed down also, but I think this will definitely improve over the next couple of years or so I would definitely say that 5G does crack open a lot of applications that requires high speed connectivity as well. The mobility sector is the number one area for sure which not only incorporates your autonomous vehicles that everyone's been talking about, but actually also the logistics side of things as well.
[00:26:18] So one thing that's been a growing trend now is that because online retail has been blowing up and will considerably blow up consistently for the next five decades, at least. The ability to set up smart logistic hubs during highways and allowing autonomous vehicles to pretty much go to a logistic hub, collect the package, and then go to the destination for delivery. A lot of this will depend on 5G, for sure. And also if you look into highways as well, if you look at the infrastructure services sector as well, especially those who are developing construction companies, then we're also looking at 5G really helping them on improving, you know, traffic congestion, improving transportation needs of a consumer as well. A lot of this really hinges on a lot of the sort of vehicles being connected together to transfer data between each other in a faster sense. And of course, run our algorithms in the background, through a cloud computer to then provide insights to the customer, whether it's a professional customer or whether it's the, your average day to day user of that transport link.
[00:27:31] Again. Yeah, I think on terms of the transport sector as well and the right different sort of overlapping industries, it will have a significant impact for sure. You know, another interesting area has been remote stations, so say offshore oil and gas platforms. They're now setting up their own 5G towers as well, because even though having a fiber based communication network throughout the offshore platform, it is actually better and faster and safer to do so with a fast 5G tower there instead of, you know, tripping on wires through the, you know, throughout the off shore platform. So I think that's also quite exciting in terms of remote areas as well. Once these are set up and there you can set up, you know, smaller 5G towers, still using the same sort of communication framework, which is a smaller tower for that one offshore asset, for example.
[00:28:22] So that's also quite exciting is that 5G is not going to be a traditional way of setting up the telecoms infrastructure, there's very sort of flexible and agile business models for telecoms companies to adopt when setting up the infrastructure, but also the sort of communication, a business model as well for the client.
[00:28:41] Ranjana: Great. So I think we've covered a vast area of things with what we're expecting for the next five years. And it was really interesting, it's very exciting with everything that's coming up. Would you have any closing remarks as well?
[00:28:59] Dishant: Yeah. Yeah, for sure. I think what's really important here is that, you know, creativity is fantastic because that's what innovation is all about. And it's always great to push the boundaries and potentially become the next Apple. That's what every company hopes to achieve or from a startup to say an SME, even a multinational corporation that's smaller than Apple. My only advice is, is that the technologies that are opening up digital business opportunities specifically. It's fantastic. But in order to really, and realistically benefit from those business opportunities, it's really important to confirm what are those sort of customer demands? What are those realistic white spaces to really showcase the value of the innovation instead of still conducting exceptional amounts of R&D and just trying to push ideas to the public. Because at the end of the day, it's only successful if your actual customer, whether it's B2B, B2C sees value in it for it to become a viable business. So my only advice in a closing remark would be lots of great ideas out there. Really focus on the ones that are realistic, where there's a potential business and the best way to do this is to test out the use case with your B2B and B2C customer. And that's all.
[00:30:19] Ranjana: Great. So we've spoken about AI and ML and chip sets being able to compute more. We've spoken about materials. We've spoken about 5G and also about customer driven innovation. I think that's a lot to look forward to for anybody in the electronics space and we're all very excited. So thank you Dishant, for being with us.
[00:30:42] Dishant: Thank you Ranjana.
[00:30:44] Ranjana: We'll speak to you again on the next one.
[00:30:46] Dishant: Alright, great. Speak to you soon.
Key Takeaway Points
3D stacking will be the disruptive part of semiconductor manufacturing
All chipmakers are now looking to integrate digital as a layer on top of the chips they produce
Smart home appliances will be amongst the most exciting disruption we will see in the next 5 years
Data centres face 2 problems: one of abundance of unstructured data, and the other of energy requirements of analysing such massive amounts of data
Material science will not see much innovation in the short term, but all of the innovation will be from the architecture and the fabrication of the architecture
Wearables will see innovation in the formation of chiplets, going beyond the current monolithic chips, which will thus make the wearables a lot more computer heavy
Various nanotechnologies will disrupt the standard state-of-the art sensing systems for the medical devices industry, making these more accessible to a domestic consumer
Partnerships are a must for disruptive players in the electronics industry to thrive. Industries are now converging and overlapping with each other, opening up way more pockets of competition and co-opetition.
5G will be predominantly used in connected automated vehicles, in offshore oil & gas platforms and in remote areas