What Emerging Technologies Are Up Next? From Early Signals to Commercial Reality

Chris Robinson: Hello and welcome to the webinar, What Emerging Technologies are Up Next from Early Signals to Commercial Reality. My name is Chris Robinson. I’m a Senior Director here at Lux and I’ll be moderating today’s session. Presenting today is my colleague Anthony Schiavo, Senior Director and Principal Analyst here at Lux. Throughout the webinar, you can type any questions you have in the questions box on your screen. Time permitting, we’ll answer all the questions that we can. If your question doesn’t get answered, please don’t hesitate to email us at [email protected]. We’ll make sure to respond. 

And lastly, if at any point you experience technical difficulties like a frozen screen, try refreshing your browser and just check that you have a strong internet connection. Lux Research helps organizations make more confident innovation decisions about what matters next. As an independent research and advisory firm, we work with a lot of the world’s leading companies to identify emerging opportunities, evaluate technologies, and ultimately make high-impact innovation decisions. So our scientists, engineers, analysts, industry experts combine original research and methodologies with practical decision-making frameworks to make sure organizations are able to separate signal from noise. So every day, we’re helping clients answer a few critical questions: Where do you focus your innovation efforts? Which technologies deserve investment? And which partners can help accelerate your innovation? Today’s webinar is an opportunity to share some of that thinking with you. 

So with that, we’ll jump in. Over to you, Anthony. 

Anthony Schiavo: Thanks, Chris. And thanks for everyone for joining this webinar today. Yeah, I’m really delighted to be talking about this. We’re going to be focusing on early-stage technologies and really looking at early-stage technology momentum. And I think we’re going to try and set up sort of two big questions, right? We want to dig into these and identify these early-stage technologies that have high momentum, identify the technologies that are up next. But there’s also a bigger question, which is we’re in the middle of this AI boom, right? What’s up next, right? What are people going to be talking about? First it was blockchain, now it’s AI. What’s that next big thing that’s really going to drive markets and valuations and hype? And so, I think we’ve heard from our clients that it can be very difficult to lift your head up from the news cycle, right? 

Whether it’s AI and everyone trying to focus on AI, whether it’s adopting AI or supplying data centers, there is a huge focus for companies to leverage this current boom. And at the same time, we have a ton of volatility. We have the war in Iran with the Strait of Hormuz closure. There’s this El Nino that has potential super weather effects, so we can have a very volatile year this year. All of this just makes it very hard for innovators to do their job. But it’s very critical that we still do think about long-term technology development. When you think about the current frenzy of AI research, it’s all based on this paper. Attention Is All You Need. This is one of the first papers really exploring transformer models that have formed the basis of the large language models that everyone is using today. This was published in 2017, just about a decade ago. And it didn’t really create a huge boom until 2023. 

So I think there’s real value in trying to understand what impactful work is happening today that could really dominate the headlines in five years or even a decade from now. And we’re conducting our innovation survey, something we do every year. And one of the questions we asked is just, hey, are things getting easier or harder? So we see pretty consistently from our clients that the folks saying somewhat harder or much harder are really outweighing the number of folks saying about the same and certainly easier or much easier. Getting to your innovation goals has been really difficult, in part because of this volatility, in part because of all these distractions, in part because of things like the pandemic. So getting an early lens into what’s up next, I think is even more critical in this context, right? If you can accurately identify these technologies that are going to impact you, you’re going to have a much better chance at navigating this more challenging innovation environment. 

And so that kind of sets up our two questions, right? Where are the early technologies that are showing strong momentum? And what’s going to dethrone AI as this sort of hot topic in innovation that is going to move markets? And as we dig into that, I want to just sort of reintroduce our three D’s framework of discover, dabble, deploy. This is how we kind of think about technology maturity at a large scale. Discovery is really primarily academic or involves modeling mechanisms for identifying technologies. And at Dabble, we’re moving into thinking about applications, right? And we’re really testing the limits. We’re trying a bunch of different things. And we’re developing the knowledge to deploy the technology successfully, right, to actually scale those applications. And so for us, in this presentation, we are really thinking about what are technologies that are just beginning to move from this discover phase into this Dabble phase. 

They’re just now at the start of that second generation. And the way we’re going to find that is with our Lux TechSignal, right? This is a tool that we’ve used before. It is data-driven. We have data on government funding, academic research, patents, VC funding, all sorts of other proprietary data that we pull in. And we use this to build a unitless measure of innovation interest. Basically, you get a comparable benchmark of how much innovation activity there is and also how much momentum there is, how fast that activity is changing. And so we are going to use this as a tool to kind of uncover the innovation interest and changes in these trends. We want to go beyond the headlines. We want to look at early signals. Now, if you saw our recent webinar, kind of the mid-year Tech Check webinar, we used the same framework there in that webinar to identify technologies. But we deployed it differently. 

We looked at the technologies that were really showing a lot of high levels of innovation interest. So very high up on this vertical sort of axis here. And here we’re going to be taking a slightly different tack. We’re going to be looking for technologies that really have strong year-over-year change, really strong momentum, but are actually below, for the most part, below that 50 mark on our innovation interest benchmark. And that’s an important benchmark because right around 45 to 50 is typically where we see sort of the average technology. When we think about all these different technologies we cover, we have this long list of 3,000 different technologies. On average, you’re looking at an innovation interest score of around 45 to 50. So what we really want is technologies that have a lot of momentum. And in the last year or two, we’re well below that benchmark and are maybe just now starting to pass that 45 to 50% benchmark. 

Or excuse me, that’s that 45 to 50 innovation interest score. And those are the technologies where we’re really seeing that change from most of the academic discovery work to that initial application development and dabbling. 

So let’s get into this first question, which technologies are up next? We scanned over 3,000 technologies, and we looked for technologies that sort of fit those criteria I laid out. Technologies where they have this really high momentum, 60, 70, 80 in some cases, 100% doubling innovation activity for something like crop stress management. But we’re still relatively low in terms of total innovation interest. There were a lot of different technologies that fit this criteria that we could have highlighted. But I think that what we try to do here with these technologies, because there is a little bit of human selection involved, it’s not a pure data exercise. We wanted to highlight technologies across a range of different industries and innovation areas. And we also wanted to highlight technologies that we felt were sort of important platforms and had a lot of wide-ranging impact across different spaces. 

And so those technologies are humanoid robotics, optical computing, materials informatics with a specific discovery focus, distributed energy simulation, crop stress management, and AI-driven diagnostics. There’s a little bit of an outlier that we wanted to include after some discussions with our teams. So let’s go through each of these six technologies, talk about them, talk about the signal that we’re seeing and where we see that technology going and get into a case study for each one. 

So starting with humanoid robotics, I think this is something that is really beginning to benefit from this sort of physical AI hype wave. We’re in AI, it’s a lot of LLMs, but a lot of the leading lights of the AI industry have been consistent in saying that large language models are not a path to AGI. What we really need is world models or physics models, AI models that can understand the world, understand things like gravity on an intuitive level. And this is going to benefit every type of robotics by making more powerful AI models to run those systems. But it’s really in humanoid robotics that we’ve seen a huge uptick in innovation activity, especially patenting. This is driven a lot by that. And VC funding is a very strong increase in that signal. And you can see that right around ’23, that’s where we really see that momentum. There was sort of a wave of interest in the sort of first robotics boom. 

From 2015 through 2017, there was a lot of focus on cheap robots, cobots, if you remember that area. And you see a little bit of that in the tech signal. But there’s very clearly a substantive momentum shift in the last two years that really is driving this type of innovation. But this is still early stage, right? You’re still looking at innovation interest that’s ultimately below average. And I think that reflects two things. I think, one, that reflects a genuine immaturity in the technology. You look at a company like Figure, they have made some impressive strides. They have a partnership with BMW to launch these factory humanoids. They are loading parts on assembly lines and they’re actually contributing to production of X3 vehicles. And they’re improving their workflows. Things requiring perception, things requiring on-the-fly correction, it’s getting easier. But the real value of humanoids is, I think, not in the factory. 

A factory is a very controlled environment and one that is already very amenable to a lot of automation. We have a ton of automation in our automotive factories. What makes humanoids different from every other type of robot is that they can navigate human spaces, a home or a hospital, places where you have to go upstairs or navigate open doors, navigate these complex environments that a regular robot might struggle with. And that type of human mobility, I think, is still relatively immature. So when you see a demonstration like this, it’s impressive. It shows the capabilities are improving quickly. But it’s also very far from the potential of the technology as we see it. So I think that reflects that immaturity that is still very much present in the technology. 

Secondly, optical computing. This is a use of essentially photonic processors to transmit light at very, very fast speeds. I think what’s happening here is we’re seeing a shift in both the commercial focus of the technology. We’re really beginning to move more into a commercial testing phase as opposed to a pure academic testing phase. We’re also beginning to see the development of more complex hardware configurations. Coprocessors and optical interconnects that handle very specific tasks and can sit alongside conventional hardware. We’re not yet at the point of having a pure optical computing system. This is quite, I think, valuable in the context of this AI wave again that we’re having here. When we look at the data signal, this is one that’s been a lot more dominated by academic work. It’s been a lot more dominated by patenting. 

We’re really seeing more and more of that type of fundamental research still driving the amount of interest or driving the overall activity within that signal. But there is commercial work happening. There are startups and companies like Q-Ant. They’re working on trying to bring this to commercial applications, right? And they have essentially developed a thin film lithium niobate photonic circuit. And there’s some very specific AI operations and simulation operations that you can execute now in an optical domain. And this has primarily been used in academic supercomputing configurations, the Leibniz Supercomputing Center. But now they have a first customer they signed in just May of this year. But this is still very early stage again. I think the challenge here that you’re going to have to look out for is really that challenge of integration. The compute capabilities that optical computing brings right now are highly, highly specialized. 

Now, there is a lot of interest in AI, and there is a lot of demand for AI infrastructure. So that needs to, that is not a small market, I should say. But you’re going to need to develop more general purpose compute capabilities. And you’re going to need to develop a more robust supply chain to package them with other types of chips for this to really grow into a mature commercial market. 

Third, I want to talk about materials informatics with a particular focus on discovery. I mean, materials informatics, if you’re in the chemicals industry, is not a new topic. But I think what has happened is that over the last decade, since we saw the first startups, there’s really been a shift towards more optimization, leveraging materials, right? These tools, they link computation, they link machine learning, they link experimental data. You can search across larger and larger chemical spaces. But where that’s really been useful is optimizing, whether it’s a formulation for enhanced properties or optimizing a material so that there is a process change that you can make, right? Maybe reducing the temperature of a process or eliminating some type of process bottleneck. But what’s happening now is a couple different things. 

One is we’re seeing, certainly since ’23, a renewed interest on all things AI, we’re seeing more investments into automated data collection with a very discovery-focused orientation. We’re seeing new advancements in the development of synthetic data to power these potential materials informatics discovery models. And we’re seeing a lot of funding. There’s been a handful of companies in this space who have raised hundreds of millions of dollars. Whereas previously, before the last year or two, companies were really raising in that single digit millions to 10s of millions of dollars range. So we’ve really seen a step change in terms of the momentum, especially on the startup side. And a lot of this is focusing on areas like critical minerals, right? 

I think there’s been a real increase that’s been growing over the years, but certainly in the last two years has hit a sort of a fever pitch or the highest level of intensity around demand for things like rare-earth-free magnets. critical minerals alternatives for data centers, chips, for military applications, right? So Magnex, excuse me, Magnex, the company, has developed a discovery platform. It screened about 100 million candidate compositions, and it identified a Magnex, which is this rare-earth-free permanent magnet. Again, This is still quite an academic test. They were working with the University of Sheffield and the Henry Royce Institute to synthesize and test the material. Discovery has been limited historically by a lack of data and the lack of manufacturing infrastructure for these novel materials. The data side is increasingly being solved. We have these newly enhanced simulation capabilities with a lot of compute capabilities. 

The physical infrastructure to actually manufacture these, though, is more challenging. Something like an electronic compound is maybe an easier target, I think, because you can scale that manufacturing a bit more quickly. But there are still, I think, meaningful challenges. So just speeding discovery by itself is good, but you actually have to be able to put this in production. And this one’s really interesting, crop stress management. This is in some ways a very old technology, right? People have been worried about crop stress for a long time. It’s not a new concept in terms of the agricultural space. But we are seeing, I think, a real sort of step change in the technology side. We’re seeing a lot of convergence of different approaches that have been developed. And we are really seeing that agribusinesses are kind of turning back towards innovation a little bit. There’s a lot of M&A in the space. There’s a lot of consolidation. 

And of course, this year in particular, we’re seeing that crop stressors are becoming a lot more prominent, a lot more common. Climate change is driving this increased crop stress. So you have this combined set of technological maturities across different, I would say, technological areas that are now being combined into new startups and new targeted approaches for crop stress management. And that’s really, it’s really emerging in a lot of the patenting data, a little bit of the funding data, but we really see this as quite commercial. And what’s interesting here is that you have companies that are a little bit more mature actually developing new capabilities to target crop stress. So Sentara is a company that has used drone-based scanning and drone-based imaging of crops for a while. This is not a brand new company. They were actually just acquired already by John Deere. 

But what’s new is that they have added increased capabilities to these existing drone platforms for crop stress management, with new hyperspectral imaging, those new sensors, and new imaging capabilities, again, driven by AI that allow leaf level detection of these crop stress indicators. So it’s a very interesting case because we’re not seeing, for example, a brand new wave of approaches or technologies. We’re really seeing all of a sudden new capabilities are possible now, and that’s being detected in the tech signal, and we’re seeing that in some of the product launches in this space. 

Our fifth technology is distributed energy simulation. These are simulation tests, particularly for low voltage networks, distributed assets, and we’re really trying to create digital twins of these more complex assets. Part of what’s happening here is we’re seeing more and more data sources. Smart meter data is becoming more common. AI-driven estimation of data, synthetic data, can help fill some of those gaps. And of course, we’re seeing that congestion and connection capacity have become major, major concerns. As the AI data center buildout has been happening, especially here in the US, this has become more and more of a challenge. And you see that in the data where funding for these technologies and also patenting interest in these technologies has increased a lot. And a company like Thinkit Labs is a good example of what we’re really seeing in this space. 

They develop models that combine machine learning with some of the more classical equations to help those planning and operations support decisions for the grid. The goal here is to really shrink the time. AI capabilities can be much faster to run and faster to produce results than some of the classical equations, which can be quite compute heavy. And the goal is to really enable more utilization and more prediction in the same quantum of time, which really makes it a lot more valuable for grid operators, I should say. And the challenge here, I think, will be establishing trust, right? These are still early-stage technologies, you can see the tech signal is still relatively young, grid operators, a very high demand for reliability and for consistency. It needs to work, right? There’s a lot at stake when it comes to managing the grid. 

But the growing complexity of the grid and the growing interconnection queue in particular really make these types of technologies important going forward as we look to enable the continued growth of our electricity infrastructure, ideally without raising prices for everyone. 

And lastly, this is an interesting one, AI-driven diagnostics. This is a bit of a combination. You’ll see the signal looks a little different from the others. We do have an uptick in maturity or in momentum in the last couple of years. And what we decided is there were a number of consumer health and wearable tech signals that really clearly fit our pattern in terms of that very high momentum in the last few years, relatively low tech signal. And in my conversations with Nardev, our colleague who leads our medical devices and consumer health sector, we decided we wanted to combine a few of these into one technology, this AI-driven diagnostics area, which we think is really reflective of the broader trend that we’re seeing in the space. This is a little bit of a different one. But these diagnostics really enhance detection accuracy, speed, reproducibility around biosignals. 

Getting to these digital biomarkers, these measurable things within the human body that can tell you something about health is really important and it’s really powerful, but it’s not yet clinically validated. However, I think we are seeing more and more new types of capabilities and more and more combinations of capabilities that are providing more value. And Hinge Health is a good example of this. They combine monitoring via video with AI-driven coaching for rehabilitation. So you have wearable sensors, you have your phone camera, and you have essentially an AI agent assisting you in completing these physical therapy routines. And that’s why we wanted to, instead of highlighting maybe one of the specific technologies that had that very sort of specific shape we were looking for in a TechSignal, we wanted to highlight this broader collection of technologies because they’re really working in concert to create new capabilities for the space here, the digital wearable space. 

And I think what’s interesting is that there is a theme that emerged here, which is AI. All of these technologies benefit from or benefit the current AI boom. And this was not our intention. When we put this together and we looked at the math, we wanted to pick the emerging technologies. And we picked them purely based on the numbers with a little bit of filtering, just based on interest areas and then things we thought we’d have something interesting to say about, but we really didn’t try and pick AI-related technologies. It, I think, emerged naturally from the data that where we’re seeing the strongest momentum is around these AI-driven technologies. 

But that doesn’t quite answer both of our questions, right? Because we’ve highlighted a bunch of technologies, we’ve talked about what technologies are up next, whether it’s digital health or whether it’s discovery or even something like optical computing. But there’s still a bigger question, which is, okay, we’re in the middle of this AI boom. What’s the next big thing? What is the next hot topic innovation? What’s going to drive billion-dollar valuations? 

And I think what’s interesting is that when you look at this, you see actually a lot of the same infrastructure and companies even in these booms for the last few years. 

Starting with crypto, you had the crypto and NFT boom, and then you had companies like CoreWeave and Crusoe start in that era as crypto mining companies, or really cloud providers for crypto mining companies. Then in ’23, you have the ChatGPT launch, and you have the chip stock run-up. In ’25, you have the data center frenzy, the memory stock run-up this year. All during this time, these companies who had really started, whether it’s Nvidia, providing the chips for Bitcoin mining or companies like CoreWeave acting as the cloud providers, they have sort of pivoted through each of these phases and have continued to remain hot as we transition from AI to, or excuse me, as we transition from crypto into AI. 

And so I think this thread is going to keep going. And that’s really why we see robotics and particularly humanoid robotics as the beneficiary of this compute wave. There’s really this one continuous thread from Bitcoin through to now where we are with LLMs, and that’s this compute change. And robotics has the ability to benefit from that. It requires a lot of compute for new models, these physical AI models we talked about. And frankly speaking, humanoid robotics have that it factor. Part of blockchain and AI and even the metaverse, what makes these technologies common is that they can tell a story about radically changing every single thing in the world. If you remember the Bitcoin boom, we were talking about, we’re going to put medical records on Bitcoin. Your driver’s license is going to be on a blockchain. You’re going to buy pizza with blockchain. There was a story about how this would change everything. 

And we see this story with AI now, how it’s going to be integrated into everything or change every job. And something like materials informatics, I think it’s a great technology, doesn’t have that type of storytelling capability. But humanoids really have that. They have that hype factor. They have that storytelling capability. And they can actually move valuations and move markets. And so I think that’s, if we’re going to plant a flag here or I’m going to plant a flag, we can see if Chris agrees with me. But I really think in the 2030s, we’ll be talking a lot about humanoid robotics, and we’ll have this humanoid robotics wave in the same way we’re having an AI and an LLM wave today. 

So with that, I want to leave you just with a few takeaways. 

First, we’re seeing that this AI boom is driving innovation and early-stage innovation really across all industries, whether it’s materials, chemicals, energy capabilities, consumer health, crops, you name it, it’s driving across every area. And I think that was a really good time to dabble, right? There’s a lot of technologies that are emerging from academia. Optical computing is a great example. Some of the crop stress work is a great example. We’re really in that early application development stage. It’s a good time to figure out if these technologies can work for you or your business, because it’s still a relatively cheap time to get involved. There have been a ton of huge raises across these sectors as a whole. It’s mostly been papers and patents and academic work that’s driving that uptick in interest. So it’s early, it’s forward-looking, and it’s a little bit more cost-effective to get involved right now. 

And as I said, I’m planting my flag. I think humanoids are going to be the next big innovation hype cycle, especially if the AI bubble bursts. Where is that interest going to go? I think it’ll go towards robots. And that’s an area where, just like all these others, we’re seeing a lot of activity out of China as well. So it’d be interesting to see if that continues to increase these sort of tensions. With that, Chris, I think we can, maybe we can open up for questions. 

Chris Robinson: Yeah, thanks very much, Anthony. We’ll be taking questions you might have in the presentation. So as a reminder, pop those in the questions box on the screen. And if we don’t get to your question on the call, we’ll make sure to reach out afterward. 

Yeah, I think maybe related to humanoids, but one of the first questions we have here is really about some of the other, another driver, which is defense tech. So the question is, what about defense tech? Is that something we considered in our analysis? 

Anthony Schiavo: Yeah, it’s certainly something we considered. It’s certainly something when we were reviewing the data and kind of scanning the data that we were thinking about. I think that defense tech has a bit of a different character, a bit of a different flavor when you look at the data. I think the momentum there has actually been a bit stronger over the last few years, even starting in the Biden administration and certainly in ’22, kicking off with the war in Ukraine. But even before then, we were seeing an uptick in dual-use technologies. We were seeing stronger momentum in defense tech. These concerns about resource availability, competition with China, even going back to Trump one, certainly through that Biden administration. 

And now I think it’s just a trend that we’ve seen that has been on more of a slow and steady increase, if you will, or maybe a fast and steady increase these days. 

But it’s not something that’s just, even though I think it’s entered the mainstream consciousness in a big way in the last few years, when you look at the innovation work and the innovation signals, you don’t see that low-then-big-rise pattern; you see something that looks much more steadily upwards. So I think there’s been a bit of a disconnect between the sort of public perception or the amount of interest that’s being paid to it and the actual momentum of the work there. 

Chris Robinson: Yeah, I think we’ve seen it be a driver, right? I mean, you talked about humanoids. I like that you called it a kind of hype bubble. But I think others like underwater robotics we’ve looked at, right? A lot of support and interest that’s defense-motivated. But if you’re any company operating offshore assets, you’re going to have a lot more tools in the toolbox in the 2030s, right, from a robotics perspective. 

Anthony Schiavo: Yeah. And I feel I should clarify a little bit here because we’re going to see a big boom in robotics. I think everyone kind of understands that the potential capabilities of robots are really maturing quickly even today. I think the hype, maybe the overhype, will be toward humanoid robotics, right? 

But there will be a lot of utility and value and real effort and work and genuine improvement in technology across the whole spectrum of robotics, whether that’s drones, whether that’s underwater robotics, whether that’s things like the quadruped robotics for industrial applications, even just general automation for industrial use cases. 

I want to make it clear that there’s going to be a lot of work that’s going on outside of the humanoid space that’s valuable, but the next bubble, I think, is going to be humanoids. 

Chris Robinson: Yeah. And I think with that question that we turned into a couple of questions, I think we’ll call it there. So thank you, Anthony. That concludes our webinar for today. We’ll send the slide presentation and the webinar recording via email to all attendees later today. After leaving, you’ll be prompted to complete a survey on today’s presentation. We do appreciate any feedback you might have. It helps inform and improve future webinars. So lastly, take a moment to check out our upcoming webinars. They’re posted on our website. So thank you for joining us and have a great day.