Rapid Changes Add New Concerns For IC Industry

Experts at the Table: Semiconductor Engineering sat down to discuss the impact of leading edge technologies such as generative AI in data centers, AR/VR, and security architectures for connected devices, with Michael Kurniawan, business strategy manager at Accenture; Kaushal Vora, senior director and head of business acceleration and ecosystem at Renesas Electronics; Paul Karazuba, vice president of marketing at Expedera; and Chowdary Yanamadala, technology strategist at Arm. What follows are excerpts of that conversation. Panelists were chosen by GSA’s EMTECH Interest Group. To view part one of this discussion, click here. Part two is here.

L-R: Accenture’s Kurniawan; Renesas’ Vora; Expedera’s Karazuba; Arm’s Yanamadala.

SE: We have more devices connected to other devices, and not all of those devices are secure. In addition to that, we have data that’s being collected today that probably will be decipherable 10 years down the road with quantum computing or better techniques. People are banking data today in preparation for that. But both of these are step functions when it comes to security. How do we plan for that now? And are we barely keeping pace, or are we actually making progress here?

Kurniawan: Quantum computing is capable of cracking any codes you have. But what is the maturity level of the field right now? And when and where are they going to hit the market in the future? That can impact your business, and it can change the way you secure your data and how you operate. This is imperative for any company operating in the world today. We need to have a continuous scanning capability to understand what new technologies are coming. It’s not just quantum computing. It’s also genAI. How does that impact my business? We are now in an environment where the technology landscape is evolving really fast, and that is something everybody needs to understand.

Karazuba: If quantum computers existed today that were available to the masses, there wouldn’t be any security. But when quantum computers become a reality, initially they’re going to be the domain of nation states. Those nation states are unlikely to be attacking your Alexa and trying to get it to order stuff for you. They’re going to be going after defense positions, nuclear launchers, and financial institutions like the stock market. That’s scary, but that’s what the attack vector is going to be. Quantum computing is absolutely the biggest fundamental risk to security as it exists today, just because of the capabilities of what a quantum computer is. And in some ways, we’re probably deluding ourselves right now thinking that we’re ready. I’ve seen quantum encryption-proof algorithms. I’m not sure how effective those can really be, but it’s potentially extremely scary.

Vora: Security is getting more complicated. Quantum computing is going to have to be a very controlled technology if it can break through most traditional security infrastructures. It’s not going to be available to you and me anytime soon. This is all about ecosystems. There has to be an ecosystem built around what is coming next, and to compensate for what’s coming next. A big part of that is going to be security. One of the challenges with security always has been that nobody cares about it until they get attacked. The medical device industry wants security now because hackers have been able to gain access to insulin pumps. The constant challenge with security is that everybody loves to talk about it, but nobody wants to pay for it until they realize how important security is and that bad things can happen. It’s important to remember that in most cases, the bad guys are five steps ahead because the incentives are higher for them than the good guys.

Kurniawan: There are still a lot of challenges in bringing that technology to the masses. The majority of quantum technology implementations developed so far still need very low temperatures, which narrows down the stakeholders, and the people or entities that are able to develop these technologies. This is an emerging technology that you should have some basic understanding of, and monitor where the technology’s going, to think through what this means to your business. How is it going to be beneficial? Or how is it going to disrupt my business? This is a technology with huge potential for disruption.

SE: We’re starting to see advanced chips for automotive that are supposed to last 10 or even 20 years. That means you have to start thinking about hardware architectures very differently in the past. How do you see this playing out? Is it going to be different partitioning? Or is it going to be a series of replaceable parts?

Karazuba: With devices that are in the market longer, like in cars, there certainly could be some type of swap-out of electronics, and maybe that that’s part of a new paradigm of car ownership where with certain cars you’re buying the car but renting the battery. There may be a service model in the future where you rent your electronics, and there’s a certain upgrade cycle where you bring in the latest and the greatest self-driving, and that would likely involve later-generation chips. It has potential from a semiconductor point of view, where it’s interesting from a revenue perspective. And from a consumer point of view, it’s extremely interesting from a feature perspective. It would be a sea change in the way that people consume vehicles, but it could happen. Cars today look very different than they did 10 years ago, and the car of 10 years from now is going to look radically different than what we have today. New business models are likely going to be needed to keep up with technology.

SE: So instead of a tire rotation, you take your car in for a board swap-out, as well?

Karazuba: Yes, and that is a distinct possibility. You don’t lubricate the chassis anymore in cars. That’s an older thing. And with electric cars, you’re not changing the oil. With a lot of cars these days, you’re not even really rotating the tires, because many of them have directional tires that are built in. But upgrading the electronics package could be extremely interesting, especially if the car is built with the intention of having that happen.

Vora: My car gets a software update almost every month now. It’s no longer about taking cars in for oil changes and things like that. We’re currently living in the software era for cars. But platform-as-a-service type of approaches potentially could include hardware. Either we have chips that are over-designed with features that are enabled with software on demand, based on certain services that consumers might want, or there can be modularization where you do hot swaps on hardware — boards, batteries, things like that, as a service. A lot of these things make sense.

Kurniawan: We’ve seen a trend across industries, but particularly in automotive, where there is a move toward customization. You are fine-tuning the features to what the customers are looking for. That translates to a future where you can potentially swap hardware or unlock new features at any time with over-the-air software updates. This could happen fast, just like the hyper-customization of electronic gadgets and phones. Now it seems to be moving to cars, as well. There are tons of things that can be done when it comes to customization and feature enablement via software. Your hardware needs to be able support many different use cases, including some you are not even thinking about today. If you can create an architecture where you have the hardware that is easily swappable and compatible with the next generation through a plug-in that adds more features, that definitely has a future.

Yanamadala: Industry rules and regulations around vehicle compliance, homologation, and legislation for spare parts are extremely complex. Requirements to hold legacy components post-vehicle production have been in place for many years. We don’t see this as the catalyst for changing the hardware architecture. But what is moving the needle is the growing demand for digitalization, leading to a software-defined vehicle approach. A single vehicle is predicted to require over 1 billion lines of code in the not-too-distant future. That’s just not possible with a linear hardware and software approach. Key to delivering multiple digital features for drivers is the ability to decouple links between hardware and software, making each element independent. So yes, hardware architectures will be radically different in the future. We are seeing that happen now, and they will continue to move toward zonal and centralized compute set ups. The independent software, using an open-sourced architecture like SOAFEE, can then be built in a far more simplified way, using cloud-to-edge technology and virtual prototyping methods. Getting that foundational platform of efficient hardware, overlaid by easy-to-use, quick-to-market software will unlock multiple opportunities for OEMs to differentiate.

SE: Where are we in terms of ensuring resilience in the supply chain, particularly in light of current geopolitics.

Kurniawan: Supply chain disruption creates a lot of pain for a lot of semiconductor players, but that pain is not uniform across the value chain. And it changes in time, as well, as businesses evolve. COVID was very disruptive across different industries, and suddenly the world realized that everybody needed semiconductors. Now, to dampen the oscillation from disruptions, we’re seeing a lot of co-investments and collaborations. One example is automakers and foundries working together to ensure there’s a strong and more predictable supply of chips. At the same time, some automakers are starting to get into the chip design. They want to ensure they have more control over the supply, that they understand the industry more deeply, and that they have a direct relationship with the chip manufacturers so they can get the products they need. That’s one example. Another example involves fabless companies. The foundries used to put in CapEx from their own pockets into building their fabs, and once the facilities were running, customers would come to them. What we are seeing now is co-investments. So instead of just the foundry CapEx, there are multiple players chipping in to reduce the cost and risk to the foundries.

Vora: The good thing that came out of COVID was that it exposed how fragile the semiconductor supply chain was. The industry is now more experienced. At Renesas, we have experienced earthquakes and tsunamis, which impacted our factory operations and how to regulate supply. As an industry, all the players have learned a lot. We’ll do a better job next time to not go from a demand bubble into a supply bubble, because everybody wanted chips and nobody knew how to forecast — and everybody was sandbagging to some extent. Things will smooth out. With all the capital investment that’s going on across the globe, building out more fabs, more factories, hopefully over the next 5 to 10 years that should help normalize some of these supply constraints and reduce our reliance on one or two companies for fabrication. It’s not going to be something we can solve in a short period of time because of how complex things are, whether that’s implementing a factory or geopolitics.

SE: What’s next for the GSA EmTech Interest Group?

Karazuba: There are so many exciting things happening in the semiconductor world right now. We’ve only begun to understand the overarching impact that AI and quantum computing are going to have on our industry. At the same time, chiplets are beginning to finally realize the promise of application-tailored heterogenous computing chips. Certainly, we also need to talk about how we are going to manufacture – and power – all of this. The GSA EmTech Interest Group will continue to cover these and emerging topics and technologies.

Vora: The only thing I’ll add is that the GSA EmTech interest group is a great forum to engage the extensive community that GSA has brought together. The semiconductor ecosystem is center stage to the unprecedented disruption we are seeing across numerous technologies. It’s a great place to discuss, learn and grow together.

Kurniawan: The tremendous demand for semiconductors, along with the increasing interdependencies, complexity, and the rapid technological shifts – they’re happening at a pace that we’ve never seen before. At the heart of the industry’s effort to sustain technological progress far into the future is ecosystem collaboration. At EmTech interest group, we truly value collaborations — idea sharing, workshops, co-writing white papers, and other activities that enable us to see new opportunities on the horizon and tackle them together. We’d love for others to join and get involved in this growing interest group.

Read the rest of the discussion:
Preparing For An AI-Driven Future In Chips (Part 1)
Designs need to be flexible enough to handle an onslaught of continuous and rapid changes, but secure enough to protect data.
Broad Impact From Accelerating Tech Cycles (Part 2)
How disruptive new technologies affect the infrastructure that will leverage them.