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. Let's go live now where Bloomberg Tech co-host Ed Ledllo is joined by a special guest. Ed, take it away.
Yeah, three pieces of news move 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. You know 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.
Yeah. Well, look, we're quite excited about Tranium and Tranium 3 in particular as you mentioned. Excited to get it in a customer's hands. Part of where we have a benefit that we can bring to bear is as you mentioned getting it into market quickly, and it's because we control that full stack. We control the silicon development, we control the data centers that it lands 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.
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. But part of the surprise of today is this. You appear to be committing to an annual cadence of new generation of Tranium. How do you keep that up?
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 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 capabilities out to customers.
The pitch for Tranium in both the training and inference use case is that it's a great deal, 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. How are both possible?
Well, both are possible because that is a great environment to run accelerators and compute in. We've been working for 15 plus years with the Nvidia team and Jensen and team to deliver outstanding capabilities for our customers. For when you're running a large cluster of Nvidia GPUs, people will tell you AWS is the best place. You get the best performance, the most stable cluster, the best capabilities out there at broad scale. It's why folks like OpenAI and others are running in AWS. And we have that choice. So for others that want to be able to take advantage of Tranium, 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. We want to have all of those available. And so we think that if we can continue to push the envelope on what Tranium can deliver for customers and make sure that we are supporting the latest and greatest from everything that the awesome team at Nvidia is delivering, that's going to be the best outcome for our customers.
The plan for AWS is to basically double capacity by end of 2027 to around 8 GW. Do you have a sense of how you apportion that capacity in-house silicon and server designs with Tranium versus Nvidia GPUs?
We're just going to keep pushing as fast as we can and we'll see where customer demand drives us as we go. As you said, we're massively adding capacity. In the last year alone, we've added 3.8 gigawatts of capacity and we'll continue to add more and more over the next couple years. And we'll let customer demand drive us a little bit on what they're looking for and what they want. That's what we always listen to and that's what we'll continue to listen to.
10 more exchanges in this transcript
Sign in free to read the rest of this interview. No card required.
Garman, M. (2025, December 2). AWS CEO Matt Garman Talks AI Race | Bloomberg Talks [Interview transcript]. Bloomberg Podcasts. CEOInterviews.AI. https://ceointerviews.ai/interview/498509/
MLA
Matthew Garman. "AWS CEO Matt Garman Talks AI Race | Bloomberg Talks." Bloomberg Podcasts, 2 Dec. 2025. Transcript, CEOInterviews.AI, https://ceointerviews.ai/interview/498509/.
BibTeX
@misc{garman2025_498509,
author = {Matthew Garman},
title = {AWS CEO Matt Garman Talks AI Race | Bloomberg Talks},
howpublished = {Interview transcript, Bloomberg Podcasts. CEOInterviews.AI},
year = {2025},
month = {dec},
url = {https://ceointerviews.ai/interview/498509/},
note = {Speaker-attributed transcript with timestamps}
}