AWS re:Invent 2025 | Rajiv Chopra, AWS

Clips
More from AWS re:Invent 2025

Rajiv Chopra

VP, AWS Healthcare AI, Life Sciences and Just Walk Out

AWS

play_circle_outline AWS re:Invent coverage highlights advancements in cloud computing and smarter AI agents.

play_circle_outline Just Walk Out technology sees momentum with customers expanding deployments to multiple stores.

play_circle_outline Revolutionizing Retail: How Just Walk Out and AI Factories are Transforming the Shopping Experience

play_circle_outline Integration of AI in retail enhances customer experiences and operational efficiency.

Info
Transcript

Rajiv Chopra, AWS

Rajiv Chopra

VP, AWS Healthcare AI, Life Sciences and Just Walk Out AWS

In this interview from AWS re:Invent, Rajiv Chopra, vice president of AWS Healthcare, AI, Life Sciences and Just Walk Out at Amazon Web Services, joins theCUBE’s John Furrier to discuss the rapid expansion and technical evolution of Amazon’s frictionless shopping technology. Chopra details how Just Walk Out is shifting from pilot programs to large-scale deployments across diverse verticals, including airports, hospitals and universities like UC San Diego. The conversation examines the concept of the "AI factory" moving to the edge, highlighting how deep learn... Read more

explore Keep Exploring

What event is being covered, who is the host, and who is the guest being interviewed? add

What are the recent trends in customer adoption and market expansion for Just Walk Out technology? add

What are your thoughts on the development and functioning of systems like Just Walk Out, particularly in relation to the integration of IoT devices and advancements in AI? add

What are some examples of AI applications in shopping experiences and how do they benefit customers? add

bolt Powered by CUBE AI

Rajiv Chopra, AWS

search

John Furrier

>> Welcome back everyone, theCUBE's live coverage of AWS re:Invent. I'm John Furrier, host of theCUBE. It's our 13th year covering re:Invent. We've seen the change of AWS over the years, cloud computing, abstracting away the infrastructure, and now with agents and the open training models. You're starting to see smarter agents hit the scene. And with that, infrastructure comes more scale, robotics, AI is really popping. Rajiv Chopra is here, VP of AWS Healthcare, AI, Life Sciences, and Just Walk Out. One of the most compelling, fun things to talk about. Rajiv, thanks for coming on, spending time out of your busy schedule to speak with us at theCUBE.

Rajiv Chopra

>> Thanks for having me, John.

John Furrier

>> So healthcare and life science is two of the hottest areas in AI right now. A lot of advancements. Physical AI too, we're seeing that. And I think Just Walk Out as one of those examples of a human experience that's been automated. There's all kinds of IoT devices. It's just a marvel of engineering. Give us the update. What is the momentum of Just Walk Out from a technology perspective? And then how's that translating into the market?

Rajiv Chopra

>> Sure, John. Over this last year, we've seen our customers really scale up. It's moving from pilots to customers signing up for 10 plus stores that they would want to deploy because they've seen either in other places or from their own experience, the business value that merchants are seeing. Obviously, shoppers continue to be delighted and more and more as they experience it. They're coming back, repeat purchases increasing. And so overall, we are seeing expansion, not simply in the United States. We've expanded to additional countries in different verticals. And so the number of stores we've launched this year far exceeded last year, and we already have commitments for next year that is like more than double what we'll be launching this year. So it's super exciting to see that this has now gotten the momentum and we are scaling Just Walk Out in many, many different verticals.

John Furrier

>> What's interesting is AWS and Amazon, they invent these big products like Connect was a great example. We talk about that a lot. We're going to speak to Colleen Aubrey later today where it's kind of built on the scale of Amazon. So you get the benefits of scale. Just Walk Out was one of those those inventions. But if you look at now, like say the retail market, the big show coming up that we'll be covering NRF in January, it's basically an AI show now. It's a tech show.

Rajiv Chopra

>> That's right.

John Furrier

>> So a lot of things are happening in retail and where the AI factory concept is interesting. So I want to get your thoughts on this.

Rajiv Chopra

>> Sure.

John Furrier

>> If AI factories goes to the edge, you're going to have a hyper converged interface. All that IoT and wireless with back-hauling to ethernet is interesting because right now it's kind of almost there. You got inference at the edge, but the training is not yet there. I'm imagining there'll be an AI factory in a retail outlet with wireless mesh networks everywhere, kind of connecting. So computer vision kicks in, all kinds of cool things. How do you see that? Because I think Just Walk Out to Me as a template for kind of where that goes, because the seamless experience that could be generated. What's your reaction to that? What's your thoughts?

Rajiv Chopra

>> Well, I mean, so everything you described is sort of like how Just Walk Out today works. And we pay attention to the details in getting all of those IoT devices connected, operating the cameras, the hardware, gates, the networking equipment working. And over the decade long of fine-tuning and improving the models, the hardware and the systems, you've gotten to a great product. But how I think AI factories is going to change it is like there'll be so many more instances where building solutions like that on AWS will be a lot easier, a lot faster. Obviously getting to ... The devil is always in the details and getting to the level that we've taken Just Walk Out in terms of the accuracy, in terms of the reliability. I mean, Just Walk Out as an offering is a fully hosted, managed offering where if there's any kind of hardware failure or an upgrade that needs to be done, it's fully covered in our costs, right? But that level still requires awesome ownership of a solution.

John Furrier

>> Talk about the current momentum with stores, for example. Where are you guys at on the progress relative to the nirvana? Which is I walk in, it recognizes me. There's John, he's walking in, he's got behavior, maybe it's a model kind of in the future there. I can see that happening, but like right now, I just want to go in and pick up my stuff and walk out. How far along are we now on that journey? How would you scope that progress?

Rajiv Chopra

>> Well, John, just to be very clear, in our Just Walk Out stores, while we track human beings, we do not identify and recognize, we like obfuscate the faces from every part of our systems. So there is no recognition that, "Hey, John walked in." It's when you identify yourself with your credit card or your wallet that we know, "Okay, this is the person who's going to be paying," and so we can track.

John Furrier

>> I'm a number.

Rajiv Chopra

>> You're a number. You're a number.

John Furrier

>> I'm a person, identity is preserved.

Rajiv Chopra

>> Exactly. With that technology, we continue to invest in it with the latest in the same transformer based machine deep learning systems that are powering GenAI or the kind of machine learning systems that Just Walk Out now is built on. And you can imagine in a stadium, you're in line to get a beer and you like just walk in and there are like 100 people there all grabbing and coming in and out. And our technology is able to handle scale of that nature. And you're going into a place where you might have a fixture in one location and then there is a gap, there's maybe a bathroom door and another fixture somewhere else. And we are able to track, even though there are no cameras, we're able to track that number across multiple locations. So we continue to invest in the Just Walk Out computer vision technology to the level of accuracy, which our shoppers care about and our customers care about. Shoppers don't want to be charged for products they didn't buy and customers obviously want to make sure that they're getting the-

John Furrier

>> You also have history with Prime, which is a great service. Everyone loves Amazon Prime. Faster, less expensive, has been the flywheel for Amazon. On Just Walk Out, share some of the use cases. You mentioned getting a beer. Share some of the deployments for people who might not be tracking the progress.

Rajiv Chopra

>> Absolutely.

John Furrier

>> What kind of use cases? Share some real world deployments.

Rajiv Chopra

>> Absolutely. I can share three great use cases. One, apart from stadiums, if you can imagine in hospitals where either patients, loved ones or the hospital staff needs access to more than a vending machine 24/7. And they can just badge in with their corporate badge and get fresh selection, get a wide variety of selection anytime, day and night. And so that kind of thing in a hospital saves the medical staff time, as well as gives them broader selection in a 24/7 operation. You can imagine at a university such as University of California, San Diego, they operate very large student and resident staff that lives on campus and they need access not simply to cafeteria food, but fresh food. And our installation there works on groceries and fresh foods and packaged foods, managing large crowds. So in a campus environment for students, again, 24/7, it's open, lots of selection, handles queues, kids coming out of class, not wanting to stand in line.

John Furrier

>> How about retail?

Rajiv Chopra

>> Yeah, totally. I think the third would be in airports, right? So you can imagine in airports, we now have many airports. You can just browse through and see what you want to get before a flight. You want to pick up something, you forgot to pick up some toothbrush. You can just walk into a Hudson store.

John Furrier

>> Which airport? Can you share the airport?

Rajiv Chopra

>> I have a long list of airports.

John Furrier

>> So you have a lot of airports?

Rajiv Chopra

>> Yes. Yes. We have multiple airports that we are in. And most recently, which we'll be talking about tomorrow as well, in Amazon's own fulfillment centers, we now in 40 plus break rooms in fulfillment centers where our staff in their break, rather than standing in line to get some food, they can just-

John Furrier

>> Walk in and walk out....

Rajiv Chopra

>> walk in and walk out.

John Furrier

>> This is the frictionless shopping experience.

Rajiv Chopra

>> This is the trend. Correct.

John Furrier

>> That's what you're going for?

Rajiv Chopra

>> That is correct.

John Furrier

>> Save time, get what you want, no hassles at the register.

Rajiv Chopra

>> No hassles, no lines. Absolutely.

John Furrier

>> How is AI impacting this? Obviously, the announcements today, AI factory, we mentioned a little bit on the retail. We see that kind of a big deal. It's going to impact a lot of retail. We see the frontier agents. How are you infusing AI? What's the mandate on your side? Because you have a lot of tech. You mentioned deep learning.

Rajiv Chopra

>> That's right.

John Furrier

>> What are some of the AI powered shopping experience? How does that translate to the customer?

Rajiv Chopra

>> Sure. So clearly, Just Walk Out is an example of AI at scape, at work, at high level of accuracy because all of the computer vision models are form of AI. When we think about the experiences for our merchants in order to manage the store or the associate in the store that's stocking the shelves or counting what needs to be replenished, their tools and apps, we're simplifying and automating those with simpler interfaces that are again going to be agentic in nature. If you think about what we are doing for deformable products like clothing, in the merchant store, which we now have many of those, we are using RFID technology. And in order to be able to read RFID tags, there is ways by which you can read them by putting it in a box and there's a reader there, but it's slow. How do we have a Just Walk Out lane where people are just scanning and you can experiment with that down here, you can see it right here at the re:Invent. The re:Invent AWS merch store is an example of Just Walk Out RFID gates. The machine learning there can determine between you walking ahead with five T-shirts and a hat, me walking behind you with a hoodie will associate you with your products, me with mine automatically. So I think the machine learning algorithms to work on RFID signals is, again, a unique use case.

John Furrier

>> You mentioned the box or the bag. I've seen some cool shopping apps at let's say grocery stores where the actual shopping cart is its own IoT device. It's got like an iPad on it. It's essentially a reader. Talk about that role, and is that a key part of this where the container or reader? Or do I need to have a mechanism or-

Rajiv Chopra

>> Well, in Just Walk Out, we are focused again on the core value proposition where you really want to have avoid long lines. And so we're focused on our RFID fast lane, which allows you to not have to take bags and stuff. You can just pick up your stuff, tap a card and Just Walk Out with the products in your hand, or even wearing a jacket and a hat and you can Just Walk Out with it and we'll charge you appropriately.

John Furrier

>> Talk about the customer perspective, working backwards from the customer. Who's the customer? What's an engagement like? What are some of the operational mechanics, installing cameras? Take me through like, okay, I'm in, I want to do, Just Walk Out, what do I do? I just call Amazon, you drop in, is there-

Rajiv Chopra

>> Absolutely....

John Furrier

>> any truck roll, is there gear installed? It's a managed service, I get that. Take me through the architecture.

Rajiv Chopra

>> Sure, sure. If you can imagine, obviously, the end shopper is a customer shared with a merchant who is our customer that we are partnering with or selling the Just Walk Out technology to. Several of them have existing locations as they say, "I have this retail location and I want to install Just Walk Out here." Whereas others are like, "I'm rebuilding this stadium and I'd love to install Just Walk Out in these five mechanisms."

John Furrier

>> So new construction, new build out, and preexisting.

Rajiv Chopra

>> New construction or preexisting, we support both. And so we partner with them in giving them very quickly, "These are your plans. Under these circumstances, wonderful." Or like, "Well, you really have some limitations and we need to make sure that we can install this." So we work with them to figure out given the actual footprint of the store, if it's an existing store, what would a Just Walk Out installation look like? And once we have that, we will send either our people or a third-party to go in as the store is being constructed or renovated. It's down for a very short period of time if it's an existing store. We've got it down now to under five days where we can come in and install the tech either through a third-party or ourselves, install the cameras behind the scenes, light them up, and that's it. You're good to go. The gates are mostly now in a way that they're light and they're movable. You can pick your number of gates in the number of locations, cameras amounted once. Over this year, what we have realized is the sheer variability in the type of stores and the flexibility that customers need, we've gone and innovated even in the fixtures of the cameras and the networking gear. We've shrunk the size of our edge computer appliances, which we install at the edge. We are innovating on making our cameras be more resilient such that because we own them and we take care, we don't want customers to have any downtime. We're investing in all of the technology on a continuous basis to make it be easier, faster and cheaper. We've cut down this year, the cost of a JWO installation by pretty much 50%. Not for us, customer's costs have come down by 50%.

John Furrier

>> Yeah. I just love the innovation because we're predicting and we're doing a big survey on this and research report around the hyper converged edge. The AI factory will come to the edge with the kind of capabilities we see with agents.

Rajiv Chopra

>> That's right.

John Furrier

>> Where I could opt in maybe for my identity, because I may want to be seen. Let me know when I'm in the store. I've used my app now to get my coffee at Starbucks. I'm out in Seattle, you guys have Just Walk Out in Seattle, which would be great, but the frictionless experiences gets more customized.

Rajiv Chopra

>> That's right.

John Furrier

>> I mean, the theme of re:Invent this year, the word custom I think was in multiple announcements. Custom transformation efforts, custom models. So in a way, the personalization becomes a huge deal. What's your vision on that?

Rajiv Chopra

>> Well, I think we are hearing from several customers that they would love to imagine a future where those frictionless stores are personalized. You can imagine as we are doing more corporate cafeteria and build out there, that they would be like, "Look, I recognize these are my employees. They're badging in with their employee badge. I'd love to offer them a personalized experience. They always get this coffee or they always want the salad this way. That's prepared and coming out." So I think thinking about such future vision, we don't have that yet, but we are absolutely-

John Furrier

>> It's a headroom for you. You can grow it. Yeah.

Rajiv Chopra

>> Totally. Exactly. And clearly, Amazon and personalization.

John Furrier

>> Yeah, it's clear. Yeah. Again, that's just given. On the customer side, the merchant is your customer, but they have a customer.

Rajiv Chopra

>> That's right.

John Furrier

>> Your customer's customer is also in mind. Take me through the authentication. Is it my Amazon account? Is it biometrics? Or does the merchant decide?

Rajiv Chopra

>> I think in our case, what we've done with Just Walk Out with our merchant partners, we've given them full control over deciding on whether they want to delegate the payment processing and the reading of the card or your wallet, your digital wallet to us, or they want to work with their existing solutions, existing providers. Any mix and match of doing so, we give them full opportunities, including this year supporting loyalty, multi-tiered loyalty. If they have already a loyalty program and they have multiple tiers of loyalty, they can give additional discounts on it. We've made the integration for our customers into these systems of-

John Furrier

>> You give them the choice.

Rajiv Chopra

>> We give them a choice. They choose.

John Furrier

>> So let's take the airport one, because I have to keep a lot of travel on. We all do the whole team. We know all the airports, so why I ask. I love to get that list by the way too. I want to maybe do a little video on that. I can see the airports being kind of the more general purpose. So what would happen if I go into the airport? Do I check in with my credit card first or do I need ... Take me through that use case, because they have to keep me more generic.

Rajiv Chopra

>> Yeah. What we have seen is there are some places where to try and manage the throughput, the Just Walk Out, by tapping your card first before you enter the store, is the most natural and works very well. It's easy to do, you get in. You're intentional, you got that and you can leave. In other places, like in an airport, for example, the Hudson News is a great partner of ours that we've installed Just Walk Out. What we saw was by moving away from this traditional model to a place where it's open entry and you pay when you're leaving the store, it invited customers into browse more because it's not that the customer ...

John Furrier

>> They do the browse.

Rajiv Chopra

>> Right. The traveler there is not like, "I know I need X." They were like, "Maybe I'm going to see, maybe I don't." And so I think removing the gate from the entry and moving it to the exit. Hudson, you saw almost, I forget, 40% increase in same source sales. So it was a big boost and it does address it in different situations, the customer has different intentions. And so by accounting for the fact that here, customer does intend to browse more, but still wants the convenience. The merchant still wants the convenience of being able to operate the store with easy flow, low shrinkage. I think this was a-

John Furrier

>> So they choose the operating model based on what they want?

Rajiv Chopra

>> They choose the operating model.

John Furrier

>> All right. So final question. Next year, obviously, this re:Invent, you see a lot of change with the agents, and you're probably well ahead of everyone else, kind of AWS because you're kind of on the Just Walk Out team. But as you see this re:Invent, the mainstreaming of agents, a lot more user behavior is going to probably align with the vision that you guys have. What are you optimizing for, for next year? What's the goals? What goals do you have for next year for your business?

Rajiv Chopra

>> Well, I think how we started, Just Walk Out now is seeing that momentum. The reason it's seeing the momentum is because our customers are seeing the ROI. I don't think we are done. I think that by focusing on our customers and drastically reducing their costs to install, to operate and to grow their stores using Just Walk Out, we're barely getting started. So I see us continuing the momentum.

John Furrier

>> What are those KPIs for momentum? Number of customers? What's the internal? What's your benchmark look like? I guess what's your dashboard look like for momentum? Happy customers, of course, but like number of customers, size of deployments?

Rajiv Chopra

>> Well, I think I love to focus on what would be the key customer metrics. And so to me, the value realization that they're getting. How are their sales increasing? How are their costs coming down? That's what we are focused on. And then obviously, we are seeing the momentum is because now there are more customers.

John Furrier

>> Word gets out.

Rajiv Chopra

>> Having larger deployments.

John Furrier

>> Word gets out.

Rajiv Chopra

>> Word gets out, and that's how this grows.

John Furrier

>> What's the one thing that you wish you could throw the magic wand at and that would change? Would it be adoption, behavior, attitude towards the tech? Is there a blocker, I guess, or condition in the market, or is it full steam ahead?

Rajiv Chopra

>> I do think that we're still early on in Just Walk Out being a household name in shoppers being familiar with it. So we still have work to do to make the Just Walk Out shopping experience be as natural and shoppers aren't confused about how the store operates. And so that's still work we have to do along with our merchant partners, make it be seamless.

John Furrier

>> Who wouldn't want to go in? Just grab a pair of jeans, I know my size. Just go in, grab it, walk out.

Rajiv Chopra

>> Exactly.

John Furrier

>> Probably some branding, a little bit of operational kind of visual use, kind of sign-in.

Rajiv Chopra

>> That's right. Yep. A little bit more on making the entry, the payment, the product discovery be what you focus on rather than standing in line.

John Furrier

>> Well, Rajiv, thank you for coming on. I would be remiss if I didn't touch on healthcare and life science. I know you have that area too. Just general observations on those sectors. Another AI tailwind, a lot of unlock as they say the word of the year in AI is the unlock of traditionally retail, healthcare, life sciences really were lagging on the IT side mainly because they didn't have the supercomputing. Now with AI factories and edge, cameras, you're going to see a lot more just do it kind of thing. I mean, a lot more just get more AI into the system.

Rajiv Chopra

>> I think the opportunity to build AI enabled solutions in healthcare and life sciences, improving, reducing the time for drug discovery, reducing the cost in healthcare, right? It's just immense and we're just getting started.

John Furrier

>> Life sciences, the discovery science is pretty incredible right now.

Rajiv Chopra

>> That's right. That's right.

John Furrier

>> Retail, life sciences and healthcare are really the big areas we're focused on. So looking forward to following up on those areas. Thanks for coming on. Appreciate your time.

Rajiv Chopra

>> Absolutely. Absolutely.

John Furrier

>> We're just doing theCUBE interviews here and he's going to just walk out in a minute, but this is a great session. Just Walk Out was a great invention by AWS where they really leverage the technology to make great user experience. And again, as technology in the physical world and digital come together with AI and other advances, you're going to see a lot more frictionless experiences. And again, this is just one example of many more. This is theCUBE doing our part to share that with you. Thanks for watching.