Menu
Sign In Search Podcasts Charts People & Topics Add Podcast API Blog Pricing
Podcast Image

Bloomberg Talks

AWS CEO Matt Garman Talks AI Race

02 Dec 2025

Transcription

Chapter 1: What is the main topic discussed in this episode?

0.031 - 18.756 Michelle Hussein

Hello and welcome. This is The Michelle Hussein Show. I'm Michelle Hussein. I speak with people like Elon Musk. I think I've done enough. And Shonda Rhimes. That's so cute. This will be a place where every weekend you can count on one essential conversation to help make sense of the world.

0

19.257 - 27.267 Michelle Hussein

So please join me, listen and subscribe to The Michelle Hussein Show from Bloomberg Weekend, wherever you get your podcasts.

0

27.367 - 29.43 Matt Garman

You certainly ask interesting questions.

0

32.887 - 37.868 Michelle Hussein

Bloomberg Audio Studios Podcasts Radio News

0

38.236 - 60.036 Ed Ludlow

Now let's narrow our focus from the broader markets to one single stock, Amazon. The tech giant hosting its annual Amazon Web Services Reinvent Conference down in Las Vegas this week. The cloud-focused confab draws developers, engineers, and other thought leaders in tech to explore the latest cloud and AI projects happening under Amazon's roof, including a new AI chip.

60.457 - 67.163 Ed Ludlow

Let's go live now where Bloomberg Tech co-host Ed Ludlow is joined by a special guest. Ed, take it away.

69.235 - 89.133 Unknown

Yeah, three pieces of news moved markets this morning. A new generation of frontier model from AWS, new agentic tools, and then a very quickly released, installed, and now ramping generation of in-house custom accelerator, which is Tranium 3. All points of discussion for Matt Garman, AWS CEO.

89.773 - 114.239 Unknown

The basic point with Tranium 3, and you've moved quickly to bring it to the real world, is cost, performance, efficiency over the prior generation. but also over NVIDIA GPUs, over Google TPUs. I think what people are trying to understand is that ramp part I was talking about, when real-world customers use it beyond this anchor customer of Anthropic, which relies on it currently.

115 - 127.435 Matt Garman

Yeah. Well, look, we're quite excited about Tranium and Tranium 3 in particular, as you mentioned. Excited to get it into customers' hands. And part of where we have a benefit that we can bring to bear is is, as you mentioned, getting it into market quickly.

Chapter 2: What is the significance of AWS's relationship with Anthropic?

127.455 - 145.041 Matt Garman

And it's because we control that full stack. We control the silicon development. We control the data centers that it land in. We know that full environment. And we can land that in very large clusters for people to take advantage of that. And the performance that we're seeing out of it is quite incredible. And so we're anxious and excited to get more and more people using it.

0

145.021 - 164.182 Unknown

I've been able to go inside Annapurna Labs and look at the engineering work between the first generation of Tranium and the second. It wasn't just the accelerator. It's at the server level as well. That's right. But part of the surprise of today is you appear to be committing to an annual cadence of new generation of Tranium. How do you keep that up?

0

164.462 - 186.267 Matt Garman

Well, the key thing that we're focused on is making sure that we can iterate on the technology. as fast as possible. The desire and the hunger out there for more power and more compute is almost insatiable. And so the more we can take an existing power footprint, an existing set of capabilities, and bring more and more compute into that for customers,

0

186.247 - 198.484 Matt Garman

to build cool applications and cool environments, and to get value from that, that's what we're focused on. And so we're going to be pushing that envelope as fast as we possibly can to get those new and new capabilities out to customers.

0

198.744 - 213.124 Unknown

The pitch for Tranium in both the training and inference use case is that it's a great deal. You know, cost-effective, performant. At the same time, you went on stage and said AWS is, quote, by far the best place to run NVIDIA GPUs.

213.625 - 213.705

Yes.

214.411 - 215.493 Unknown

How are both possible?

215.774 - 237.973 Matt Garman

Well, I mean, both are possible because that is a great environment to run accelerators and compute in. And so we've been working for 15 plus years with the NVIDIA team and Jensen and team to deliver outstanding capabilities for our customers. When you're running a large cluster of NVIDIA GPUs, people will tell you AWS is the best place.

237.993 - 257.235 Matt Garman

You get the best performance, the most stable cluster, the best capabilities out there, and broad scale. That's why folks like OpenAI and others are running in AWS. And we have that choice. And so for others that want to be able to take advantage of Tranium, and there's some use cases that are best for Tranium. There's other use cases where NVIDIA GPUs are going to be your best option.

Chapter 3: How does AWS's Trainium3 chip improve AI performance?

399.95 - 418.753 Matt Garman

And as I mentioned, through Project Rainier, it's a huge collaboration there to go build their current generation models. And all their models run today and launch on day one on top of Tranium and on top of AWS, which we're incredibly excited about. And we'll continue that partnership for a long time. I think from them, they have a huge demand for compute. And so...

0

418.733 - 432.367 Matt Garman

They'll go to other places where it makes sense to round out their compute needs because they just have such massive needs for compute. And they have customers in other clouds as well. But we're definitely their primary cloud provider and closest partner for sure.

0

433.448 - 447.943 Unknown

Supply constraints. So Amphropic is supply constrained. They can't get the compute they need. We've talked about the ramp on NVIDIA GPU and in-house silicon. Is there a supply constraint element with AWS? Are you able to get the chips that you need?

0

448.303 - 466.273 Matt Garman

Yeah. I think there's always, anytime you see an industry that's growing as fast as this is right now, when you think about AI and model development and chips, there are going to be constraints no matter what. There is more demand than there is supply. Sometimes it's in chips, sometimes it's in power and data centers, sometimes it's in

0

466.253 - 487.52 Matt Garman

You know, different parts of that, you know, at some points it's, you know, networking equipment. At some point it's it's transistors, resistors or whatever it is. And you look at the entire supply chain that is needed to ramp up at such a massive rate. Right. Never before has the technology industry ramped at the rate that we are right now. And so there are always constraints.

487.54 - 503.922 Matt Garman

And so it's not necessarily there is necessarily one constraint where it's like, wow, I can't get Nvidia chips. We can get Nvidia chips. And actually Jensen team have been incredibly supportive and great partners in helping us get capacity there. It's not that you can't get power. We're getting power all over the place, but it's just, we're ramping all of these places

503.902 - 510.111 Matt Garman

in such rapid rates that always there's a constraint in that system. And it'll change every month you ask me of what the current one is.

510.531 - 526.854 Unknown

Throughout the day, we were just speaking with your team about the idea we're moving from AI assistants to AI coworkers. You know, particular focus on the agentic offering that you've done. You're in the camp of people, if you don't mind me saying, that sees basically 90% of the value in enterprise coming from agentic technology.

527.315 - 531.681 Unknown

Do you have any data or evidence to support that all of your customers are ready for that?

Comments

There are no comments yet.

Please log in to write the first comment.