Risky Business: Amazon's big bets | AWS Events

I have a portfolio that extends like from Zoox, if you've seen it here in Vegas, self-driving cars or Robotaxi. And on the other end of that would be what you would call Project Leo or Kuiper. So, those are the big kind of bets.

I'd say Alexa's a pretty big bet too, though. I also, I'm in charge of Alexa, Alexa+, and then the product lines are Hero, Fire TV, Ring, Blink, Kindle. And there's a long list. And then now that I didn't say them all, I'll get in trouble next week, but something like that.

Okay.

It's pretty tricky. Between Ring, Alexa, Fire TV, you're talking about hundreds of millions of customers using their products.

Every day. And I would say somewhat religiously, they use it for with purpose.

When we talk about customer obsession, it's not a tagline. It is literally like in our bones. And so when you have hundreds of millions of customers using your product every day, you need to bring them along on any journey you're on or where you think you're going.

I think this is probably the most critical thing because as AI comes down the pipe and every one of us is reinventing with AI right now, and then using AI to reinvent the experience, you can't just move people's cheese. You can't just take away something that they've been relying on for five or 10 years, critical. So, how do you do it? First you respect the customer and what they're telling you today, what they need, what they do every day. Understand the data, understand what they're using your product for, and make sure you respect that without any doubt. And don't get me wrong, there's also this clamoring of, 'Well, what's AI going to do to transform your product?' And when you get to that point, you sit back and say, 'What are the things you need to do that are useful for your customer? Where they respect not only what you're bringing forward, but they get to use the product for what it was meant for.' And I think this is kind of the tension we deal with every single day, because you're never done building these products. And if you can understand the usage patterns, that's great, but they also need to continuously be reinvented. I think an example, if I give you one is... So, there's a customer need, if you can meet it, maybe the customer doesn't know how to ask for it. But let's talk about Ring. You have Ring video descriptions right now. This is a pretty interesting. Previously if you were using Ring, your notification would just say, 'There's movement in front of your house,' or... Now you get a full video description of there are two people walking in the driveway, one of them has a shovel. And you're like, 'Oh crap, what the hell's going on in my driveway?' And you immediately go to your phone, you get the information you need, but that's AI in action. It's a need that the customer had. They didn't necessarily ask for it, and then when you present it, it has this incredible solution for information that AI brought forward. And it's a useful piece of AI, it's an incremental change to the product, but it's transformative.

Yeah, I think across, in my team's world, we're always thinking about the device and the service that goes with the device. So, it's the service and the device coming together, whether it's Alexa+, or what will eventually be the Leo Service. The Fire TV, the streaming service. Product making, if I answer your question, I talk to the team like this. There's three sections to this pie, when you talk about customer obsession. Sometimes people get lost and like the customer said they needed this and I gave it to them. I think that's true when you're making products about 33% of the time. This is the way I kind of think about it. This is my own construct. 33% of the time the customer, whether it's through a survey, through usage, through what they do every day for what they said they needed. If you're listening and you have to listen, you must listen, you must listen. They're going to be right. And that's about 33% of what being customer obsessed is for a device. The other 33 or 60, the next two sections are, one of them is in your labs, in your teams, there's technology that only you can see. And whether that's one, three or five years out, it's technology you know is coming that you know can solve new customer problems and find new needs to solve for that will make not only the product exceptional, but the experience for the customer off the charts. So, 33% of it is, you know what the tech is, you know what your customer is saying. Do you really solve the problem now or do you lead them to what's next? The last part of that pie is vision. It's never as simple as here's the tech, how do I fill it in? Or here's what the customer said, how do I answer it? It's also, what's your vision? Where are you taking the customer? I think this is really important. If you're developing now, you're developing to get the customer where they want to go. And these are slightly different things. Like here's what they asked for, here's what they're trying to get to. And in five years from now, are you going to have a product that's ready for them? If you only listened to what the customer said in feedback, you would never develop what was coming in five years. You just wouldn't be able to see it.

A lot of times, like there's a one list, or if you have a like from a commercial customer or CIO comes in and says, 'Here are the 10 things I need.' You got to serve those 10 things. But if you thought that was the only part of serving your customer, they're just going to have a new, they're going to have a new product before you know it.

Yeah, that's a perfect way to put it. Well framed.

We have a leadership principle at Amazon that we talk about that I just love, actually adore. And it lives every day with us, and it's called 'Think Big'. Kuiper/Leo, and I'm going to say both because I've taken feedback at this show. Like, 'Why did you change the name? It should just be Kuiper.' And then I've changed the, and like, 'Yeah, I love the name Leo.' So I'm just going to say both right now.

The team really wants me to say Leo, but okay. Is one of those big bets. But to your question, the way I think frame it and how does a big bet really work? There's always going to be hard decisions. There's always going to be challenges and their time-based bets, generally when they're big, they don't happen overnight. But I will say this, that the most critical thing you can do when making a big bet is be unbelievably crystal clear on what your mission is. It is the number one thing by far, without any doubt. As a matter of fact, you should all ask yourselves, if you're making a big bet right now at your company, how clear are you on your mission to yourself and to the people that are working on it. This I found is to be, it doesn't solve every problem, but it's for sure the anchoring point on where you're trying to get to next. Mission must be clear. So, think about Leo, the mission is basically to solve the digital divide, provide internet for everyone on the planet. It's super clear. And to do that, you have to do it. You start understanding the principles. You have to do it at lowest cost possible. You have to do it in a way where connectivity never fails you. You have to do it in a way that's inventive in new ways. That then leads you, once your mission is clear, it leads you to decisions on speed of building satellites, how fast you need to get them up. How do you think about designing your silicon? How do you make sure you take as much cost out of the product as possible upfront so your customer terminals are reasonable. So you actually can solve that digital divide. And when you start breaking it down, you then end up in this bet that's massive. Here's the trick. We don't blink. At Amazon, we don't blink because the mission is so clear, specifically on Leo. And so, when the investment has to be made or when the bet needs to be laid down, you go back as, does it serve the mission or does it not? Now, there's a lot of nuance in there, but at the end of the day, where I would challenge yourselves and I always challenge myself is like this serve the mission or not? And what it also leads to, which is really important, and I'm a big, like, I believe in teams long before I believe in technology, because it's always the people that are going to create it and serve it.

If your mission is clear, you end up with missionaries on your project. Critical. It's the missionary versus mercenary. Have you ever heard that conversation?

And a project this big where the bet is so long, I mean, we're six years, billions of dollars into Kuiper. We're right on the cusp of that knee bend, right, where we're, that hockey stick, that curve up where all of you're going to be able to use it, and we can talk about that. But ultimately then you end up with missionaries who are very invested, they understand the mission and they run at it hard. And I think that's where we're at right now with Leo.

It's tricky that, I'll go back. You don't compromise the mission, but when you do have to make trade-offs, then you have to find the gray areas. There's a couple of ways to think about it. First go back to big decisions. Along this journey, we're six years in and it's a lifelong project to connect the world. You end up in so many decisions that are one-way doors, what we call one-way doors and two-way door decisions. And you try and get to as many two-way door decisions as you can. So, you can kind of create the journey as you go and go ahead and make mistakes. On big bets, you're going to make a lot of mistakes and we've made plenty. But if your mission's clear, you just stay on the path. Maybe veer every now and then, but if it was a decision that was wrong, it's okay, come back. When you get to those one-way door decisions, those are the ones that can't be compromised. So, now you're at a place of trade-offs, you're either going to get more investment. Which is, that's tricky. Or you are going to look at what is the customer promise you're making. And where is it that you do have to now start, not trade off what the customer needs, but look at the mission, hold it customer needs, hold it. And now you can always, and I promise you, always find the trade-offs within that. And if you're disciplined about it, you know when that trade-off is truly a one-way door decision. And I would bet it's very rare that that's the case. And so you take the risk on the two-way door, and if you're wrong, you just come back to it, and you just get vocally self-critical, and you say, 'I missed.' And you come back and you said, 'Trade off was wrong,' and you just rehash and you keep moving forward.

Yeah, for me it's huge. You know, I look at it as it's so critical because you then start choosing where you should put your time. This is a two-way door decision. Like empower your team is everything. I've just seen a couple team members out here today, and I look at them and I get to take so much pride because I go, 'These are leaders who just make the critical decisions every single day.' But of course when there's a big one way door, like that one decision that you just can't come back from, let's all get in a room and let's make it. And I think when that's the case, you just end up moving faster. And Andy talks about, being the world's largest startup, that's kind of how you do it. You empower your teams to run as fast as they can on as many decisions as possible. And you just keep moving.

It wasn't on my list.

Yeah, well, here's the status. It's pretty awesome. There's 150-plus satellites that are up right now. Ultimately, we have four different launch providers, five different rockets that we launch on. We have a series of launches coming up. We're building multiple, many satellites. We're in pretty good shape. In the early 2026, customers in this room, you can start testing it now, but in early 2026, you're going to be into being able to use it as commercial customers.

We're pretty excited. Like, we're right there. It's exciting. And then it's just this powerful, like, you have this opportunity, it's such a unique product for our commercial customers because you can go direct to AWS. It's about as secure as it possibly gets. And going through the internet, nothing, just direct connection. We're pretty pumped about it right now.