It's been one month since the changing of the guard at the top of Amazon Web Services, with Matt Garman becoming CEO, and he joins us now for an update on how things are going and to lay out his vision for AWS, the leader in the cloud market. Missus Garman, good morning to you. Let's start with the basics. What have you spent the last thirty days or so doing and how quickly have you tried to implement changes that you think were helpful important AWS?
Oh, thank you and thanks for having me this morning. It's been a great first month. I've been diving into various teams and understanding where all the teams are working on. In particular, I'm excited about the opportunity in front of us. I've spent some time sitting down with customers and understanding what's important to them. Spent a bunch of time last week in Silicon Valley talking to some small startups to understand what's going on in the general of AI space.
And there's a lot of excitement about new technologies that are coming and the potential for where business as go. So I'm incredibly excited and looking forward to the business.
Now we will talk about generative AI. But your background is so interesting because you've held pretty long periods of both engineering and sales responsibilities in the cloud market at AWS. You know, before we were talking about AI, the story with cloud was really clear. There was this massive total addressable market where companies all around the world had not yet transition workloads and storage to the cloud. And increasingly we're actually talking about on prem being back a little bit.
Just talks to me about your strategy for going after that addressable market the basics of cloud computing.
Yeah, the opportunity ahead is really enormous for the business. If you look out there and you actually talk to most customers, the vast majority of their workloads aren't yet in the cloud. Some estimates are only ten to fifteen percent of workloads have actually transitioned to the cloud. And actually we hear from customers that, if anything, they're looking
to accelerate that transition to the cloud. If you think about genitive at AI or other technologies, most customers are finding that it's hard to actually take real advantage of some of these new technologies that are out there if
their data is not out in the cloud. And so when we talk to a lot of customers, they see this new era of generative AI as a accelerator for reasons, to get their data in the cloud, to put it in a data lake, and to be organized in a secure and safe environment where they can actually go and innovate and deliver value for their business. And so we're
actually seeing this as an accelerator. Actually, when I talk to most customers, they're trying to find ways that they can migrate more of their workloads and more of their data into the cloud faster than they ever have before. And so we think that that's a huge opportunity for
AWS business. But we also think that's a big opportunity for our customers as they can gain agility and can innovate faster for their own customers once all those workloads are in the cloud and they can focus on delivering value to their own customers.
Alphabet and Microsoft and IBM and all that is also see this as their opportunity too, Matt, And I'm interested as someone who basically helped build cloud as a business model back from two thousand and five, when you're intenning how you think Amazon and AWS is different at this moment.
Yeah, you're right.
So, and as you mentioned, I've been working in AWS since two thousand and five, two thousand and six, since we started AWS, and so if you look about where we're different, there's a couple of things that I would focus on. Number one is we have a lot of experience going and helping customers in the cloud world.
And from the very beginning, when.
We very first started AWS, we said Priority zero for us is security. Customers are trusting us with their business, they're trusting us with these critical workloads, and so security is first and foremost.
Second is operational excellence.
And so we know that for our customers, if we have to be secure and we have to perform an outstanding way, so performance security, those are the two things that we focus on. And then we think about how do we help customers innovate, And we really like to partner with our custom We want to lean in. We want to understand where their struggles are, where they're having problems with their on prem environments, or where they're slowing down because they're doing too much of the muck we
call it, instead of actually innovating for their customers. And then we deliver a really rich set of services. And if you go look at AWS, we have by far the widest set of services and capabilities that allow customers of all sizes, whether they're startups or enterprises or governments,
to go and innovate out there for customers. And so when you have that baseline of security and operational excellence and then you have the broadest set of capabilities, it's for many customers and most customers out there, it's a no brainer as to what's the optimal platform to go from?
And so, yeah, so is that no brainer? Meaning you're still accelerating for AWS growth as we see that pivot point in the previous quarter. Are you still accelerating profits too?
Absolutely?
AWS is still growing faster than any other cloud provider out there on an absolute basis, and we see customers leaning in and so we see again this is such early days for the cloud. It's really is early days, and so we see massive growth and acceleration ahead of us and huge opportunity for the business.
Matt, since you've been in the job for thirty days, what's your understanding of what a run rate in the billions of dollars tied to generative AI actually meaning what have you come to see across your desk? Is that being a business of Yeah?
And as you mentioned, general AI is already a multi billion dollar business for AWS, and so it is a big business for us today. And when you sit down with customers, I think you look at kind of where this transition has gone. It started out called a year year and a half ago with chat GPT, and everybody thought this was a fantastic technology, and it.
Is and it was.
And then so the first thing that everybody did is they basically built chatbots for their website, right. And so as customers, though, are transitioning, they're trying to say, where's the real enterprise value for me? How am I delivering value to my customers? And so when again when they think about AWS, they say, I have a platform that I can build on. It's secure at the baseline, it has a bunch of capabilities that I can actually go
and build value for my customers. And so we increasingly are seeing customers say, how do they get real value for their business? How are they delivering revenue growth? How are they delivering real cost savings? How are they completely transforming their business? That's just different than how anyone has done their industry before. And I think that's the real potential. Even though today it is a multi billion dollar business. I think it really is the nascent part of this
technology because it's what it's capable of doing. Is progressing so quickly that I think a year two years from now you're going to see many industries out there have completely changed how they think about work, how they think about delivering value to their customers. And it's a super exciting time for innovation and for technology as a whole.
For our Bloomberg television and radio audiences around the world. We're speaking with the AWSCO Matt Garman thirty days into the job. An important development your predecessor was the anthropic investment, most recently the Aqua hier so to speak of a debt labs, which is more focused on non AWS at Amazon. How much do those deals indicate that Amazons had to bring in engineering talent to work on large language models.
How we think about it is the potential is massive, and we think that if you look at the AI space, there's not going to be one solution for everyone. There's not going to be one model that's perfect for all use cases. Our view is that customers are going to need a variety of capabilities, They're going to need large
models to do reasoning. And the anthropic models, by the way, are fantastic, and the new claud three point five sonnet model is by far the best in the world right now, and that's fantastic and we love working with that team.
But a lot of customers, we find are combining those large models with small models, and then they also need capabilities, and one of the areas that we're really excited about is agentic workloads, where models are actually going to be able to call out and do things out there in the world, real world, not just summarize data or give you information, but actually go and perform actions. And that's part of where we're really excited about the technology from ADEPT.
And so we think that there's a lot of these different components.
Many of them are going to be built in house and have been built in house with AWS. Many of them come from external capabilities, and we think that there's going to be a really rich ecosystem of capabilities that customers are going to want to build from. And many of them are partners with us, and many of them will provide US first party offerings.
A rich ecosystem needs rich infrastructure and energy. At that map, how are you seeing the need for energy growing an aws, how are you satisfying it?
Then?
Much talk about nuclear for example.
Yeah, it's a great point, and it does need a lot of energy, and that's why if you rewind even several years ago, we're the single largest purchaser of renewable energy in the world for many years in a row now, and part of that is we saw this coming and we've been thinking about how do we build up enough power in a sustainable way that's carbon neutral that we can support the energy needs for technology.
And so.
We think that that growth is going to be significant, and there's a lot of work that's going to go forward in continuing to make sure that we have enough power to power the workloads that our customers need.
But that's part of.
What I think we bring to the table. We've spent a lot of time getting good in this space. We've spent a lot of time thinking about how do we make sure that we have enough sustainable power, how do we make sure that we have enough components in our supply chain, and how do we manage that spend. As you mentioned, it's a pretty significant infrastructure investment. How do we make sure that we manage that in a reasonable way, so that it's good for the AWS business and that
we're being responsible with how much capital we spend. But we can also grow the power footprint that the world needs in a sustainable way. I think that's an important thing of what ABS brings for people.
Can I ask about therefore that profitability because there has been marked improvement in profitability, But the question is at the time of investment, are you managing this by cutting costs by alsomny having let go of people, You're managing to equate it more by just more higher pricing for these AI related services.
Now we think about this business like this business will have ebbs and flows over time, particularly as we make large financial investments in the future, whether it be large investments in power or data centers or components and things like that. So the profitability of the business will go up and down over time. But long term, we think that AWS will continue to be a very profitable business
for Amazon. But for us, it's all about investing in the future and making sure that we have the right team, that we have the right setup in place, and we have a long term vision of what this business is going to be. So we're not thinking about particularly short term, but we're thinking about how do we really go after that long term potential so that AABS continues to grow
at a rapid rate for many, many years. And we want to make sure that we have that right team in place, we have the right infrastructure in place, and we have the right investments in place so that we can be there for customers when they need us.
Now, we dedicate a lot of time on this program, covering your proprietary silicon efforts Trainium for example, So we've done that. Could we focus on Nvidia and in discuss with me your relationship with that company, how supply of their AI accelerators is, and whether you actually will give us a forecast of how many accelerators you think you'll take from Nvidia in the next twelve months or calendar twenty four.
It's a great question. And Nvidia is an incredibly important partner of ours. They make a fantastic processor, They have executed incredibly well and have by far the world's best technology in this space and have for the last couple of years. And we believe that AI technologies are going to rely on Nvidia and GPUs for many years to come, and so we think that they're a super important long term partner of ours. In fact, we're getting deeper and
deeper with them. Jensen has said on stage that AWS is their technology partner that they're building their own models on top of. They believe that, and they've seen that AWS has by far the best performing technolog whether it's on the network side or the availability of the GPUs, which actually is a super important part of building models.
And so Nvidia trusts AWS to go and build their long term models, and we lean in on Nvidia for that too and learn about what's working well, what in the next generation technology will they need from our data centers and networking, And we really have a great collaboration there to ensure that our AI customers can jointly take advantage of the best technology that Nvidia has to offer and AWS has offer in a combination so that they can really deliver the best technology for the world.
Partnerships being built. Matt Common, thank you so much for joining us today, awsk you for having mea