AWS Summit NYC 2026 | Jay M. Wong, Luminous Robotics & Alla Simoneau, AWS

Clips
More from AWS Summit NYC 2026

Jay M. Wong

CEO

Luminous Robotics

Alla Simoneau

Emerging Technology Physical AI Lead

AWS

play_circle_outline AWS, NVIDIA and MassRobotics Launch Physical AI Fellowship to Propel Robotics Startups into Real-World Applications

play_circle_outline GenAI Innovation Center developing visual-language-action and world models

play_circle_outline Data Flywheel for Edge-to-Cloud ML: Redeployment, Global Scaling, and AWS Go-to-Market Advantage

Info

Jay M. Wong, Luminous Robotics & Alla Simoneau, AWS

Jay M. Wong

CEO Luminous Robotics

Alla Simoneau

Emerging Technology Physical AI Lead AWS

search

Gemma Allen

>> Welcome back to theCUBE coverage here on the ground at the Java Center. It's AWS Summit 2026 and we are bringing you all the action on the ground, talking to some of the folks shaping this next phase of AWS's story. Joining me now are two folks who are going to talk about the world of physical AI, Jay Wong, CEO of Luminous Robotics and Alla Simoneau, Emerging Technology Physical AI Lead at AWS. Welcome, guys.

Alla Simoneau

>> Thank you.

Jay M. Wong

>> Thank you.

Gemma Allen

>> We're going to talk about solar panels, which I think are something, Jay, that most of us can at least visualize and understand. You have made a bet though that it's not really about the cost of the panel itself. It's about the human element of having to install this. Talk to me a little bit about that. Maybe start there and break down what it is that Luminous Robotics does and what the before and after looks like for your customers.

Jay M. Wong

>> Sure. So, what we do is we build fleets of industrial robots and we deploy those robots onto active solar construction sites. And essentially, we go and try to slot into the existing workforce, and help increase the speed and safety of building out these large scale solar power plants. We're in a very interesting timeframe right now within decarbonization, electrification efforts, as well as the AI data center demand for energy. No other time in human history have we had the need to build at this type of scale for energy infrastructure. And it's getting to a point now, like the bottleneck is just we don't have enough humans to even build this type of infrastructure, let alone it being in very remote areas. They're lifting very heavy equipment and typically very dangerous. So, what we've developed over the past couple of years is we've been leveraging physical AI capabilities, but also building our own heterogeneous fleet of robots that go and essentially enable this technology, enable solar construction sites to be built faster and more efficiently.

Gemma Allen

>> Wow. It's a fascinating story and time. Alla, for you and your team and your relationship with Jay and his team, talk to us a little bit about how you guys work together, how this relationship even was formed, what exactly it is that your group does. And I'd love to hear too about how you at AWS and at Amazon broadly think about the world of physical AI, why you guys are so invested in this.

Alla Simoneau

>> Yeah, thank you. I mean, we are lucky to be working with Luminous Robotics as part of the physical AI fellowship. So, at the AWS organization, we have a physical AI organization that's looking to advance development of solutions in this space. And so, we've had the great pleasure of launching the fellowship with AWS teams here with NVIDIA and MassRobotics that have encompassed 17 different startups, who are doing breakthrough innovations in the space of robotics and physical AI. And so, in the GenAI Innovation Center, we have an esteemed team of capable scientists, strategists who are technically advancing, call it the tech venture of many of our customers, working alongside them to advance really critical solutions. So, in the case of Luminous, really helping to build the brain or reimagine the brain of the system that's powering these devices that are doing the installation in these solar power plants. Tremendously, what's actually happening and why we're investing so much of our time capacity at AWS, from infrastructure to model development and optimization for these customers, is because the industry has come to an interesting sort of standpoint. You have cheaper hardware, sensors, LIDAR. It's just easier to acquire. You have new advances with generative AI, but really general model development. And the kind of models that we're talking about here are visual language action models, world models, and those capabilities are advancing. And all of this is coming really to a standpoint because of the human side of things, as you've alluded to. You have massive labor shortages across various industries, whether it's manufacturing to solar panel installation and agriculture, to healthcare, that are catapulting just to the front fold. So, all of those factors together are the reason why brilliant startups like Luminous Robotics are reimagining the possibilities for industries to advance and drive impact. So, to give you a little sense, like in the space of solar installation, just here in the US, there's currently a shortage of 53,000 labor workers. That's today. If we talk about manufacturing and supply chain, there's a shortage in the United States of a half a million jobs. 40% of that labor in that industry will retire in the next 10 years. If you look at healthcare, there is, according to the World Health Organization, a shortage of 11 million healthcare workers by 2030. If you start to add all of these elements up, what physical AI is doing is it's transforming the capabilities previously that were just like digital or manufacturing optimization, workflow optimization that has been an innovation for the last 50 or so years from many organizations. Taking this now into the digital space and real environments that are learning, interacting, and unstructured environment. But fundamentally, transforming how people are living, working, and engaging with these types of devices, so to speak, because I don't want to call them robots, but really devices or systems of operation. So, it is really just a fascinating time to be in this. And AWS teams are not just building the infrastructure for these companies to scale, to learn, to train their data, to be able to operate at such scale, but also to optimize the brain or the models that are being deployed at the edge. And all of this knowledge and know how is also inspired by some of the work that we're doing at Amazon Robotics. We have the largest fleet of robots deployed anywhere in the world, with over a million robots across our warehouse operations as of September of last year. Many of those robots are powered by AWS. And so, we ourselves as an organization recognize the massive opportunity ahead of us to rethink how we're working every day. So, this is a very exciting time to be in this industry.

Gemma Allen

>> And to Alla's point, I mean AWS and Amazon, it's a monolithic force in tech. In terms of partnering with AWS and with Alla's team, talk to me a little bit how that helps you from the perspective of go to market. I know there's a lot of product and technology advantages. How do you think about that compound ecosystem effects when you think about your own GTM strategy here and globally?

Jay M. Wong

>> So, actually, we have been deploying these robots globally already. We're across the United States, as well as Australia. And something that has been super powerful working with Alla's team revolves around the data that's being acquired by our robots. We have a fleet of robots that go and do useful work on these construction sites. But at the same time, they're accumulating massive amounts of data, which is being processed on the edge for reasoning. But at the same time, we're periodically pushing and uploading this data back up to AWS, fine-tuning our AI models, and then essentially redeploying those new models that are refined onto the robots. So, we essentially get this very positive feedback loop where the more robots we deploy into the space, the more data we acquire. And as we acquire more data, it essentially gets to make the entire fleet more efficient, more capable, allowing us to deploy more robots to collect more data. So, we get this amazing data flywheel effect by working with Alla's team. And they have some of the most awesome top-notch engineers really helping us.

Gemma Allen

>> Including Alla herself.

Jay M. Wong

>> Including Alla herself, that have been really wonderful to work with. Our team has been super fortunate to be able to leverage these resources and be able to deploy at scale.

Gemma Allen

>> Because you're in such an interesting point, in terms of solar and energy, you mentioned energy is the biggest bottleneck. I think there's a debate between energy and compute, but they're highly interconnected. What source of challenges and constraints do you guys see in the industry broadly, especially when you think about physical AI and AI on the edge where there is lower compute, more flexibility needed? What sorts of constraints are you seeing globally even as you are looking to scale this out?

Jay M. Wong

>> From my lens, it's interesting that the data center compute isn't measured in flops anymore. It's measuring gigawatts. That's because it's one of the biggest bottlenecks, in terms of being able to have sufficient energy to go empower these data centers. From the actual construction perspective, solar has gotten to a point where it's the lowest cost of energy, but what's more interesting is it's actually the fastest to build. So, you can build a solar power plant in order of 12 to 18 months. Nuclear is awesome, but it could be eight years to never.

Gemma Allen

>> If it ever can happen, right?

Jay M. Wong

>> Yeah. And typical other sources of generation are plus minus five years. So, in the context of having to build generation in a very short amount of time, I think there's an interesting case to be driven to build these large scale solar power plants, just because of speed to energize. Now, within that construction process, as Alla was saying, we literally don't have enough human labor to do this task. Most of the products that are being installed, like the solar panels, the racking infrastructure, they're all hundreds of pounds. It's not like one guy can go and pick things up easily. So, typically, now it's like two to five humans, working crews, hundreds of people on these sites. They're in remote areas. They're not in the middle of the city. So, now you have to think about logistics, lodging. And it's gotten to a point where the big unlock is to be able to automate and not remove the people from the process. We still want to have these people on the site, doing the dextrous manipulation, the electrical work, QA, QC, but more so enabling them through the context of automating the mechanical heavy lifting. Robots are very good at that. Humans are very good at other things. And that's the kind of synergy that we're going into market with, really trying to increase efficiency to increase the speed energized.

Gemma Allen

>> I mean, it's all about moving up that value chain, ensuring that there is less bottlenecks and more opportunity for innovation by great minds like yourselves. And on the topic of innovation, last question to both of you. There's a lot happening here today and it's going to be some exciting announcements. What are you both most looking forward to from the perspective of Luminous Robotics? What sorts of hopes do you have, I guess, for this next 12 months out? And I'll put that to both of you and this relationship continues to grow.

Jay M. Wong

>> You want to go first?

Alla Simoneau

>> I have a lot of hopes. I am most hopeful that the public understands, that was what was once written sort of as sci-fi possibilities of technology are now very much the reality. And that the advances over the last 15 years in traditional robotics are now transforming out of a data virtual environment into a physical space. And that means a massive transformation. And I think public understanding of what the possibilities are and rearchitecting the way we live, the way we work across different industries. And I'll just say as I think about Luminous, even as a space and thinking about solar panel installation. Just in the work that they're doing, you have to understand the constraints where two people cannot lift according to OSHA requirements more than 50 pounds each. So, a panel couldn't have been more than 100 pounds based on human constraint. Now, take that constraint away just in this one industry and now you have 500 pound panels or massive panels, because there is no limitation. Think about the size, the scale, the energy production. Think about those factors that had formally blocked innovation in this industry and completely rewrite them. This is what physical AI is doing and this is why AWS so committed to companies like Luminous, but really all organizations who are building at this intersection of physical AI and robotics, because we are at this transformational pivot point in history to completely change the dynamics of all industries all across the globe. So, I am hopeful that this will become a mainstream, truly mainstream topic about conversation, and every event here at AWS and everywhere around the world, amongst leaders who are reimagining the future of their business.

Gemma Allen

>> I love it.

Jay M. Wong

>> I think what I'm very hopeful about is the scale of impact that we're going to see at this inflection point. If you think about how much impact these LLMs have made over the past couple of years, this is the digital space. And now we think about bringing some of that technology into the physical space, enabling work to be done at large scales. Alla had a really good point about offloading some of this mechanical work that is just beyond the scope of what a single human can achieve. Even today, modern solar power plants are being built with 100 pound panels. No one person can do that. It's already two to three people lifting these things. Being able to go and really rethink just work in general, through the context of being able to remove these human constraints, allows us to build some really interesting things in the future. I think that's what I'm really hopeful about and that's only being enabled now through these new physical AI capabilities. I've been building robotics my whole life. And I think it's never been a more exciting timeframe than today, where we can use these tools, really have robots reason at a level that has never been previously attainable, at like accuracies and rates and truly delivering value to these customers. Luminous wouldn't exist 10 years ago, just because these tools weren't around. It's a great time to be.

Gemma Allen

>> And I guess that's what it's all about, continued value to customers and to society. So, Alla and Jay, thank you so much for joining us on TheCUBE.

Alla Simoneau

>> Thank you so much for having us.

Jay M. Wong

>> Thank you very much.

Gemma Allen

>> I'm Gemma Allen here at the AWS Summit 2026 in New York City. We are bringing you all the action on the ground, talking to some of the folks pioneering what's next for AWS and this phenomenal ecosystem. Don't go anywhere.

Jay M. Wong, Luminous Robotics & Alla Simoneau, AWS

search

Gemma Allen

>> Welcome back to theCUBE coverage here on the ground at the Java Center. It's AWS Summit 2026 and we are bringing you all the action on the ground, talking to some of the folks shaping this next phase of AWS's story. Joining me now are two folks who are going to talk about the world of physical AI, Jay Wong, CEO of Luminous Robotics and Alla Simoneau, Emerging Technology Physical AI Lead at AWS. Welcome, guys.

Alla Simoneau

>> Thank you.

Jay M. Wong

>> Thank you.

Gemma Allen

>> We're going to talk about solar panels, which I think are something, Jay, that most of us can at least visualize and understand. You have made a bet though that it's not really about the cost of the panel itself. It's about the human element of having to install this. Talk to me a little bit about that. Maybe start there and break down what it is that Luminous Robotics does and what the before and after looks like for your customers.

Jay M. Wong

>> Sure. So, what we do is we build fleets of industrial robots and we deploy those robots onto active solar construction sites. And essentially, we go and try to slot into the existing workforce, and help increase the speed and safety of building out these large scale solar power plants. We're in a very interesting timeframe right now within decarbonization, electrification efforts, as well as the AI data center demand for energy. No other time in human history have we had the need to build at this type of scale for energy infrastructure. And it's getting to a point now, like the bottleneck is just we don't have enough humans to even build this type of infrastructure, let alone it being in very remote areas. They're lifting very heavy equipment and typically very dangerous. So, what we've developed over the past couple of years is we've been leveraging physical AI capabilities, but also building our own heterogeneous fleet of robots that go and essentially enable this technology, enable solar construction sites to be built faster and more efficiently.

Gemma Allen

>> Wow. It's a fascinating story and time. Alla, for you and your team and your relationship with Jay and his team, talk to us a little bit about how you guys work together, how this relationship even was formed, what exactly it is that your group does. And I'd love to hear too about how you at AWS and at Amazon broadly think about the world of physical AI, why you guys are so invested in this.

Alla Simoneau

>> Yeah, thank you. I mean, we are lucky to be working with Luminous Robotics as part of the physical AI fellowship. So, at the AWS organization, we have a physical AI organization that's looking to advance development of solutions in this space. And so, we've had the great pleasure of launching the fellowship with AWS teams here with NVIDIA and MassRobotics that have encompassed 17 different startups, who are doing breakthrough innovations in the space of robotics and physical AI. And so, in the GenAI Innovation Center, we have an esteemed team of capable scientists, strategists who are technically advancing, call it the tech venture of many of our customers, working alongside them to advance really critical solutions. So, in the case of Luminous, really helping to build the brain or reimagine the brain of the system that's powering these devices that are doing the installation in these solar power plants. Tremendously, what's actually happening and why we're investing so much of our time capacity at AWS, from infrastructure to model development and optimization for these customers, is because the industry has come to an interesting sort of standpoint. You have cheaper hardware, sensors, LIDAR. It's just easier to acquire. You have new advances with generative AI, but really general model development. And the kind of models that we're talking about here are visual language action models, world models, and those capabilities are advancing. And all of this is coming really to a standpoint because of the human side of things, as you've alluded to. You have massive labor shortages across various industries, whether it's manufacturing to solar panel installation and agriculture, to healthcare, that are catapulting just to the front fold. So, all of those factors together are the reason why brilliant startups like Luminous Robotics are reimagining the possibilities for industries to advance and drive impact. So, to give you a little sense, like in the space of solar installation, just here in the US, there's currently a shortage of 53,000 labor workers. That's today. If we talk about manufacturing and supply chain, there's a shortage in the United States of a half a million jobs. 40% of that labor in that industry will retire in the next 10 years. If you look at healthcare, there is, according to the World Health Organization, a shortage of 11 million healthcare workers by 2030. If you start to add all of these elements up, what physical AI is doing is it's transforming the capabilities previously that were just like digital or manufacturing optimization, workflow optimization that has been an innovation for the last 50 or so years from many organizations. Taking this now into the digital space and real environments that are learning, interacting, and unstructured environment. But fundamentally, transforming how people are living, working, and engaging with these types of devices, so to speak, because I don't want to call them robots, but really devices or systems of operation. So, it is really just a fascinating time to be in this. And AWS teams are not just building the infrastructure for these companies to scale, to learn, to train their data, to be able to operate at such scale, but also to optimize the brain or the models that are being deployed at the edge. And all of this knowledge and know how is also inspired by some of the work that we're doing at Amazon Robotics. We have the largest fleet of robots deployed anywhere in the world, with over a million robots across our warehouse operations as of September of last year. Many of those robots are powered by AWS. And so, we ourselves as an organization recognize the massive opportunity ahead of us to rethink how we're working every day. So, this is a very exciting time to be in this industry.

Gemma Allen

>> And to Alla's point, I mean AWS and Amazon, it's a monolithic force in tech. In terms of partnering with AWS and with Alla's team, talk to me a little bit how that helps you from the perspective of go to market. I know there's a lot of product and technology advantages. How do you think about that compound ecosystem effects when you think about your own GTM strategy here and globally?

Jay M. Wong

>> So, actually, we have been deploying these robots globally already. We're across the United States, as well as Australia. And something that has been super powerful working with Alla's team revolves around the data that's being acquired by our robots. We have a fleet of robots that go and do useful work on these construction sites. But at the same time, they're accumulating massive amounts of data, which is being processed on the edge for reasoning. But at the same time, we're periodically pushing and uploading this data back up to AWS, fine-tuning our AI models, and then essentially redeploying those new models that are refined onto the robots. So, we essentially get this very positive feedback loop where the more robots we deploy into the space, the more data we acquire. And as we acquire more data, it essentially gets to make the entire fleet more efficient, more capable, allowing us to deploy more robots to collect more data. So, we get this amazing data flywheel effect by working with Alla's team. And they have some of the most awesome top-notch engineers really helping us.

Gemma Allen

>> Including Alla herself.

Jay M. Wong

>> Including Alla herself, that have been really wonderful to work with. Our team has been super fortunate to be able to leverage these resources and be able to deploy at scale.

Gemma Allen

>> Because you're in such an interesting point, in terms of solar and energy, you mentioned energy is the biggest bottleneck. I think there's a debate between energy and compute, but they're highly interconnected. What source of challenges and constraints do you guys see in the industry broadly, especially when you think about physical AI and AI on the edge where there is lower compute, more flexibility needed? What sorts of constraints are you seeing globally even as you are looking to scale this out?

Jay M. Wong

>> From my lens, it's interesting that the data center compute isn't measured in flops anymore. It's measuring gigawatts. That's because it's one of the biggest bottlenecks, in terms of being able to have sufficient energy to go empower these data centers. From the actual construction perspective, solar has gotten to a point where it's the lowest cost of energy, but what's more interesting is it's actually the fastest to build. So, you can build a solar power plant in order of 12 to 18 months. Nuclear is awesome, but it could be eight years to never.

Gemma Allen

>> If it ever can happen, right?

Jay M. Wong

>> Yeah. And typical other sources of generation are plus minus five years. So, in the context of having to build generation in a very short amount of time, I think there's an interesting case to be driven to build these large scale solar power plants, just because of speed to energize. Now, within that construction process, as Alla was saying, we literally don't have enough human labor to do this task. Most of the products that are being installed, like the solar panels, the racking infrastructure, they're all hundreds of pounds. It's not like one guy can go and pick things up easily. So, typically, now it's like two to five humans, working crews, hundreds of people on these sites. They're in remote areas. They're not in the middle of the city. So, now you have to think about logistics, lodging. And it's gotten to a point where the big unlock is to be able to automate and not remove the people from the process. We still want to have these people on the site, doing the dextrous manipulation, the electrical work, QA, QC, but more so enabling them through the context of automating the mechanical heavy lifting. Robots are very good at that. Humans are very good at other things. And that's the kind of synergy that we're going into market with, really trying to increase efficiency to increase the speed energized.

Gemma Allen

>> I mean, it's all about moving up that value chain, ensuring that there is less bottlenecks and more opportunity for innovation by great minds like yourselves. And on the topic of innovation, last question to both of you. There's a lot happening here today and it's going to be some exciting announcements. What are you both most looking forward to from the perspective of Luminous Robotics? What sorts of hopes do you have, I guess, for this next 12 months out? And I'll put that to both of you and this relationship continues to grow.

Jay M. Wong

>> You want to go first?

Alla Simoneau

>> I have a lot of hopes. I am most hopeful that the public understands, that was what was once written sort of as sci-fi possibilities of technology are now very much the reality. And that the advances over the last 15 years in traditional robotics are now transforming out of a data virtual environment into a physical space. And that means a massive transformation. And I think public understanding of what the possibilities are and rearchitecting the way we live, the way we work across different industries. And I'll just say as I think about Luminous, even as a space and thinking about solar panel installation. Just in the work that they're doing, you have to understand the constraints where two people cannot lift according to OSHA requirements more than 50 pounds each. So, a panel couldn't have been more than 100 pounds based on human constraint. Now, take that constraint away just in this one industry and now you have 500 pound panels or massive panels, because there is no limitation. Think about the size, the scale, the energy production. Think about those factors that had formally blocked innovation in this industry and completely rewrite them. This is what physical AI is doing and this is why AWS so committed to companies like Luminous, but really all organizations who are building at this intersection of physical AI and robotics, because we are at this transformational pivot point in history to completely change the dynamics of all industries all across the globe. So, I am hopeful that this will become a mainstream, truly mainstream topic about conversation, and every event here at AWS and everywhere around the world, amongst leaders who are reimagining the future of their business.

Gemma Allen

>> I love it.

Jay M. Wong

>> I think what I'm very hopeful about is the scale of impact that we're going to see at this inflection point. If you think about how much impact these LLMs have made over the past couple of years, this is the digital space. And now we think about bringing some of that technology into the physical space, enabling work to be done at large scales. Alla had a really good point about offloading some of this mechanical work that is just beyond the scope of what a single human can achieve. Even today, modern solar power plants are being built with 100 pound panels. No one person can do that. It's already two to three people lifting these things. Being able to go and really rethink just work in general, through the context of being able to remove these human constraints, allows us to build some really interesting things in the future. I think that's what I'm really hopeful about and that's only being enabled now through these new physical AI capabilities. I've been building robotics my whole life. And I think it's never been a more exciting timeframe than today, where we can use these tools, really have robots reason at a level that has never been previously attainable, at like accuracies and rates and truly delivering value to these customers. Luminous wouldn't exist 10 years ago, just because these tools weren't around. It's a great time to be.

Gemma Allen

>> And I guess that's what it's all about, continued value to customers and to society. So, Alla and Jay, thank you so much for joining us on TheCUBE.

Alla Simoneau

>> Thank you so much for having us.

Jay M. Wong

>> Thank you very much.

Gemma Allen

>> I'm Gemma Allen here at the AWS Summit 2026 in New York City. We are bringing you all the action on the ground, talking to some of the folks pioneering what's next for AWS and this phenomenal ecosystem. Don't go anywhere.