Amazon CEO Jassy Talks Alexa+, AI Demand and Trump

Bloomberg

27-02-2025

Live on Bloomberg TV
CC-Transcript
00:00How is this going to really well excite your customer base? What are you most excited for? Well, you know, we've had Alexa's been around for ten years. We have 600 million devices in customers, homes and offices. And Alexa Plus is our next generation, Alexa personal assistant, and she's meaningfully smarter and more capable and useful than her prior self. You can do all the things that you used to do, but every single one of those functions is better. I can give you just an example. There are so many examples, but you know, if if you have smart home controls with Alexa now you can say Alexa. I have guests coming over at 7 p.m.. Could you could you raise the drapes? Could you raise the temperature by five degrees, Turn on the porch, lights in the driveway, Lights and put on mellow dinner music and a mellow music. You're not a guy. And this is so easy. Me, I choose my fighters, but most people will choose male dining music. But you can do that all verbally and simply, using conversational language like that. You don't need an app, it just happens. And so Alexa plus with what we've just announced and what we're launching, it really is there have been a lot of chat bots around that are good at answering questions, but they don't take actions. Alexa Plus is really Alexa is going to be the first one that not only is highly intelligent to answer various questions, but she can do so many things for you. She can play music and play video and control your smart home and make reservations for you and hire people to fix your oven. I mean, it really is the first big, large scale practical use of gender by the consumers. I'm going to be able to see and use. Naturally, I can see how you're going to obsess over it, whether it's ordering the latest Buffalo Wings or New York and seeing where the best space is to do that. But how are your stakeholders of shareholder base going to obsess about this, about generative AI in Alexa Plus? How is that adding value for them? Well, you know, I think if you think about what Alexa allows customers to do, it makes shopping more easy because it's so much more intuitive now to buy products. It makes enjoying music easier. It makes enjoying video and streaming media better. It allows you to control your smart home in a different way. So every single one of our consumer customer experiences gets better with Alexa Plus. And then, you know, of course, Alexa has its own business model. We have a brand new lineup of devices that are coming in the fall, I think, that are beautiful. I think people going to really like we have opportunities for to service new products and advertising in various interfaces like our mobile and our desktop interface that's coming in. Alexa And then we have subscriptions and you know, so I think there's a sustainable business model there as well. Talk to me about subscriptions because you're getting it free if you've got Prime. I get a lot with Prime now. Are you able to increase the prices that you think of Prime. Well, from Prime is an incredible value. If you think about getting free shipping on 300 million plus items, we launch Prime with free shipping on about a million items. Today, it's 300 million items. And most of the time you're getting your products now inside of a day. And so, you know, between that and what you get with prime video and prime music and the grocery subscription and you know, our unique selling events, it primes an incredible value so you could increase to add on top of it Alexa Plus is just great value So we don't have any plan right now but it's I do think Prime's unusual value and it's why people use it so expansively and to increase that use has been this invention, this innovation of generative AI and that costs money. And you've actually just took to the stage yesterday to say out of all companies you are spending the most on AI. How much are you spending on it? Well, you know, I think we don't disclose the exact amount, but, you know, we you know, we're spending a pretty significant amount of CapEx and the lion's share of it is on generative AI. We've said in our business, even though general AI for us is a many billion dollar a year business and growing triple digit percentages year over year, that if we had even more capacity, we could use it to monetize it. And we have this really interesting and very fortunate flywheel and AI inside of Amazon, which is if your mission is to make customers lives easier and better every day, which it is for us. And if you believe that all the customer experiences we know of today are going to be reinvented through generative AI, which we also believe you believe those two things, you're going to be building a lot of generative AI apps if you build a lot of gender, of if, by the way, other companies are too on top of it, if you ask which is the leading technology infrastructure platform. So if there are a lot of gender of AI apps being built on top of you, you can't help but get a lot of feedback from people on how they want those building blocks that create generate to be better. And if you're willing to invest in those building blocks, which we are, as you know, with our own chip. Subterranean and with our own frontier model with Amazon, Nova and model building services and stage maker I bedrock. If you're willing to invest in those building blocks and you're getting a lot of feedback, they get better much more quickly, which can't help but make it easier and quicker for people to build gender VR applications, which means you get more running on the platform. So that flywheel is very unique for Amazon. The lion's share, though, you did say basically $100 billion run rate for CapEx expenditure. Can you give us even like a percentage breakdown of how much that goes to distribution, logistics and how much goes to Air Alliance shares? You know, so most of it you know, most of it, you know, the lion's share is more than 50%. Yes. Is it more than 80%? We're playing the warmer and colder game. Yes, exactly. I love a game. But talk to us about that capacity that you talked about. Yeah. Areas could grow even faster if you had all the chips that you needed, all the power you needed, the motherboards you needed, How much faster could I best grow? I think it's hard to put an exact percentage, but but I do think it could be growing faster. I mean, I'm confident could be growing faster. And, you know, there's there is, you know, for a long time, there's still aren't as many chips as we all what we have. We're fortunate in that we're very big partners with video. But then we also have our own custom silicon and trading, too, which we just released and reinvent, which is 30 to 40% more price performance than the CPU power instances, which is a big deal at scale if you're doing Gender VII and the inference side. So we have maybe more chips and some others might have access to, but we still don't have enough. And then there's just is not enough power in the world right now and we're all working really hard on that. I expect that to relieve some the second half of this year. But right now and, you know, the world can change, but right now we have just insatiable demand. We had amazing demand coming from Jensen Huang, who had his numbers out yesterday and Nvidia. Was there a limitation on the in video chips in particular that pulled back capacity? And at what point do you think can depend even more on your own in-house built chips to offset any of that? Well, you know, I would say I mean, there's a lot of demand for general AI right now. People are very excited about it. I think that all the different providers of chips have been constrained to some extent. I think some of the some of the new generations maybe have gone through different evolutions. And when they're going to be released maybe a little later than people thought, there are some components like motherboards and things like that that we all use, that particular ones are in shorter supply than others. So, you know, I do think there's some there are some supply chain issues which I expect to get better. I do you know, people are very excited about training and to to your question and whether we could see more demand there. And we have gone back at least a couple of occasions to make more training too, than we had intended because we have so much demand. So I expect we'll have customers for for as long as I can foresee wanting to run compute on instances in NVIDIA chips. But I think a lot of the demand will also be served by training the customers that you couldn't serve because of the limited capacity. Is it that you just had to pull it back from everyone a little bit more generally, or who lost out here, do you think? It's it's always a combination. I mean, there's four people that just have a very small amount of accelerators that they need. They don't usually have a problem with something called capacity blocks with. It's kind of like an on demand way to use accelerators in and generate chips, and that continues to grow. It's really the folks who have built, you know, have an idea for a new application, but they need a lot of of chips where if we don't have the capacity, we have to you know, we have to give them what we can give them and give it to them as fast as we can and push our partners to get it in sooner. And they can't get their initiatives done as quickly as they want to have. The capacity, isn't there? Anthropic said. All capacity it needs you to place relationships. Yeah, we have a very close partnership with Anthropic. We have this project called Project Rainier with them. They're building their next model, their their next version of their frontier model on top of training too. And our customized silicon, they're going to use over 400,000 training to chips. And so, yes, they have they have capacity. They're ramping up. And we're excited about that partnership and what they're building. You mentioned the power side. Now. How much is that something you're talking to the administration about? How much is the administration supportive of the build out that you need to do? We have been talking to administration, you know, multiple administrations in this country over the years as well as in other countries as well. And I think the power shortage really snuck up on people, you know, really right after the pandemic. And I would say that the current administration is very receptive to it. They understand the constraints it's having on the economy right now and or are convicted about solving it. One of. Out the restrictions around chips. And what's interesting is Microsoft just called on the administration today to say this limitation on chip access for some of our close allies around the world is going to limit all global business. You are a global business. It's not something you're worried about. And we are you know, I think I think that you really talking about that air diffusion act. I think and I'm going to be curious to see where that goes. I mean, it was it was enacted pretty quickly at the very end of the last administration. I don't know how this administration feels about it, but I would say that we share the concern that it has limitations on certain countries who are natural allies of the US who just to be able to do their business and those companies to be able to get done what they want to get done on top of these technology infrastructure platforms like ours, they're going to need more chips. And so I think if we don't do it, we're going to basically give up that business and those relationships to to other countries who can provide those chips. And I think we're better off being partners with them. Is it a risk to us? It's not so much risk. I mean, it has it you know, in the scheme of things, it's not a big swinger, but. But I also think that there are so many countries who are in the early stages of their economic development who both really need access to the most cutting edge, sophisticated technology to build the right customer experiences. And that could be big geographic markets for companies like ours and lots of other technology companies, where I think it would be a shame to limit them and to limit the companies that the air diffusion rules that was brought in very swiftly by the Biden administration. All of this is in the context of U.S. versus China and not wanting to get the most sophisticated equipment and chip and technology into China. That's what great for us for a moment. The administration and whether it's been net positive or negative for your business when it comes to China, de minimis, for example, the fact that Chinese competitors, as we call them, that Sheehan and the like, are going to have to pay more to get goods into the country. Does that help or hinder you? Well, you know, I would say, you know, de minimis specifically. You know, we have a certain number of items that are shipped in that way as well for things like Haul, which is our new low price offering. We maybe have lots of it, and some other companies like the ones that you mentioned, but I think it's early in this administration. But what I would say is that it is encouraging to us that we have an administration that wants to hear from business. I would say that, you know, we've we've been a business through six administrations, every single one of them. Our primary focus is to take care of customers, but we try to build a productive relationship with the administration because we want to help the country. And I would say that some administrations are more receptive to it than others. But this administration cares about what business thinks. And I've always been surprised that it isn't obvious that the best economic results for a country are going to be when the public and the private sector collaborate. And, you know, I don't expect the the government to kowtow to what companies want, but they should get their feedback and their input because they're going to make policies better together if they collaborate. And I'm encouraged early on that this administration wants to talk to businesses. You're taking calls personally. You know, I. I take calls. We talked, you know, same thing with all the administrations we talk to. People in the administration we share was working for us. What's not working for us? Concerns that we have. As I said, some administrations care more about our feedback than others. Has Trump cared about yours? Have you spoken to President? I have spoken to the president. You look, as I said, this administration has been pretty busy the first month. And but I am encouraged that they are having conversations with businesses and they do care about our feedback. And we'll see what happens. But I you know, it starts with a dialogue. You have to have a dialogue to have any kind of relationship. And they care about infrastructure. Just take Stargate. Are we going to hear more from you on how much you're investing here in the United States on the air infrastructure build out as well? Well, you know, as as we talked about earlier, where we said it was, you know, directionally right. In terms of the run rate on a CapEx, you know, we're spending a lot and I infrastructure and the lion's share of it is not just on air, but also in this country. In the US we spend elsewhere because we have a global business. We have customers everywhere. We've customers in know a couple hundred countries, but we have we have a pretty substantial investment there. I don't expect it to attenuate soon. We could get a number. I'm interested in something that perhaps is going to feel a more sensitive topic and and it comes around perhaps some words that were missing in your annual report. Not this year, which for diversity and inclusion. I put this in the context of the administration as it stands, because I know that Amazon's tries to be the Earth's best employer, and I'm just wondering how your employees react to perhaps the lack of certain words now involved and whether or not programs might be forced to change or not. Yeah, well, what I would tell you, just at a high level, if you serve as many customers as we do in as many diverse groups as we do, we intend to moving forward. You have to have a diverse team to be able to build products that, you know, work for the for for everybody. And that has always been our intention. It continues to be our intention. I think that, you know, there were so many programs that we launched and other companies launched in the pandemic. And as you probably have seen over the last three years, we've gone through very thoroughly every single one of our business areas and the programs that we had conviction about, we doubled down on and the programs that we don't have conviction about. We streamlined and we stopped doing. And so we did the same thing in looking at all of our programs on diversity. And, you know, we have some programs that we have double down on. A good example is our career choice inside our fulfillment network. Our fulfillment center teammates are able to get an advanced education for free on us to advance their career, and they're in their own development. And that has been very, very successful and very meaningful. So we double down our program like that. There are other programs that really just haven't been that successful and haven't moved the needle much in those we just moved away from. But we we have a diverse group. We're trying to continue to build out a diverse group and that won't change. And you can't you can still use that word ultimately. You're not having to reframe it. That is I mean, you can call it lots of people call it lots of different things, but we have a giant customer base of every imaginable group of people where we want builders who can build for them. Culture is key. Yeah. And you just talked about how you're doing more with less is always a focus on frugality That's in your very principles. That's something you've been doing at the employee base as well. Wrote out, made it very clear that we're going to be reducing layers. How is that going? Yeah, it's going well. You know, look, if if you have a company where the culture is an important ingredient in your success, which is absolutely been true for Amazon, you it's not your birthright to keep having a strong culture, you know, especially as you grow the number of people and the number of businesses you're in, the geographies that you're in. And so you have to work at it all the time. And for us here, there were two areas when we looked at it as a leadership team last year that we wanted to strengthen. One was we have always hired really strong owners, smart, ambitious, strong owners who get to own the lion's share of it. In this case, you know, I would say, you know, 90 plus percent of the decisions and and they you know, as you add a lot of people, you end up with a lot of middle managers. And those middle managers all well-intended, want to put their fingerprint on everything. So you end up with these these people being in the pre meeting for the pre meeting for the pre meeting, for the decision meeting and not always making recommendations and owning things the way we want that type of ownership. So we took a goal collectively to increase the ratio of individual contributors to managers by over 15% of the company by the end of this quarter. We've made very good progress in that. We beat that already and it's going to allow us for the people that are doing the work, they're going to have more ownership and they're going to be able to move more quickly. And then I think the other thing we saw was that if you're a culture that invents a lot and collaborates a lot like we do, if you don't have people in the office together doing that invention, it's just meaningfully worse. And, you know, you you know that the same way you don't collaborate, the same way you don't connect with each other, the same way you don't learn the culture the same way. And so having people back in the office together more frequently, we felt very strongly that it's going to be better for customers in the business. Let's talk about that invention and that collaboration. We've got a new quantum chip ocelot. Love the name. What does that mean for you to have this more efficient chip at this exact moment? Will we start seeing it be practically useful soon? Well, quantum compute quantum computing is very high potential. It has the chance to solve some very computationally intense problems, and I still think it's realistically a few years away from from having a real shot at solving this problem. But you have to solve a bunch of these challenges that relate to quantum computing along the way. And one of them really is around error correction on the qubits. And that's what also it does. It's a very unique, inventive way to do error correction with qubits that makes a meaningful difference. And we're excited about that milestone. You can't get to something that has real impact unless you get those milestones done along the way and you invent along the way. And that's just No. Other example invention. I feel the learning of how generative I suddenly became, not just in the business parlay, but suddenly everyone was discussing it. I think people are aware Quantum could do the same thing. So when you hear people debating between two decades or five years, where do you sit on the ground, the scope of that? Gosh, you know, I don't know for sure. You know, I would say I'm hopeful that it's more in the five year range. That is the 20 year range. All these things that are successful are seven, ten, 20 year overnight successes. You know, it wasn't like, you know, if you look at generative AI, it's just kind of another evolution of AI. But we've been working on AI for 50 some odd years. I mean, like and then it just, boom, it happened and it really shocked us when it actually was more accessible and worked. And I think the same thing could happen with quantum computing, which is, you know, takes a long time, takes a long time, takes a while, and then all of a sudden it's functional and solves problems that you couldn't solve easily or cost effectively before. And it just feels like it happened overnight. But quantum computing we've been working on now for ten plus years, and then the euphoria comes and then people try to make head or tail of how long that euphoria lasts. Going back to the investment that you make, and particularly in Generative, I do you think it's going to be peak year for generative AI in terms of that investment that Amazon makes? I don't know. You know, it's funny, I have this feeling that you're going to end up with some people that feel a little disillusioned about gender of AI because the investment. Are you getting the commensurate return? It's still very early of so many companies that are really doing pilots right now, their gender of the AI applications. And here you see a little bit of it. But I think that the smart companies are going to figure out which are the initiatives that can really change their customer, experience their businesses and keep investing in general AI. And the the slower companies are going to wait to see if it's safe to go outside and they'll they'll be behind by two or three years, maybe more, because the reality is even more so than software development generally. AI is very iterative. It's not in software development. You can get on a whiteboard with a team of architects and design something and maybe it doesn't work exactly as you design it, but largely, you know, whereas in general they are the models, they, they get better at kind of disproportionate rate, sometimes failing or we got a new scaling law. I mean, I think that a lot of times when you're building models in the model get so much better. You talk to the scientists and the team, they just can't believe how much better because the model is learning itself. And so I think that if you actually aren't investing in general, you're going to be behind by even the amount of time that you waited, because there are so many lessons you get from iterating and building applications today and you want to be the supermarket of AI. It feels like agnostic to the models as it is with Amazon Alexa Plus. Well, you know, look, the truth is what we care most about in all our businesses. But as it relates to eight of us and I, we want our customers to be able to change their customer experiences and improve their businesses so they can last over a long period of time successfully. And if we do right by our customers and take the long term approach, we do and they're able to run their applications successfully on top of our technology infrastructure services, then they're successful and we ride along with them. And so, you know, we in all these areas, we have services, we build ourselves in the models area. We have Amazon Nova, which people are really excited about because it's it's got comparable intelligence, the leading models in the world, but it's meaningfully less expensive and lower latency. But if customers prefer to run other models, we have huge partnership with Ropeik and with Llama and Deep Sea Mistral Deep Sea. We have the largest collection of leading foundation models in the world, and if customers are having success with those, then we're happy. And the truth is, if you build a lot of genera of AI applications, people don't realize this. You use multiple model types often in the same application. Even Alexa plus, as we talked about, uses multiple foundation models. And so we want people to use the right model for their applications and then we make it easy for them to switch between them and run it easily and successfully. Idea because.