The Information's TITV
00:19:27 · Ken Gawrelski
[00:19:27] S2: Sure. So, Trainium is important to Amazon and to AWS. However, it's much earlier in its development. You have Trainium three out there now with Trainium four that is likely forthcoming. TPUs are on version seven on... And and eight is forthcoming, I believe. So we... TPUs... Google TPUs are much further along in their development. There's a broader ecosystem, and Google has become much more aggressive in selling its TPUs externally. Right? And this is, you know, has just started really, we think in the the first sales were really came in in a small amount in 2Q, more so in 3Q, and it will grow quite substantially from here. What Amazon is talking about is by the end of twenty seven, they would expect to start to see some delivery of Trainium chips. That's external sales, and that will be more of a factor in '28. But that doesn't... That's not the whole story on Trainium. The story on Trainium is that much of the inference that happens over AWS today is served over Trainium. It is... There's a inefficiency to that for Amazon. It's a lower cost for them. And so it is very important to the AWS growth. It is one of the pieces of capacity for AWS that's critical. Again, we're in a world where today where demand exceeds supply for compute capacity, and Trainium is a critical piece of of Amazon's compute capacity in addition to partners with NVIDIA GPUs and and CPUs. Right.
00:21:14 · Akash Pasricha
[00:21:14] S0: I wanna ask you lastly about CapEx and how they're building for the long term here. I wondered on your end, when you're doing your modeling, how are you doing the math on whether or not the CapEx is reasonable, whether it's too much, too little? Just give us a view into your own modeling exercises and what inputs you're putting into that.
00:21:38 · Ken Gawrelski
[00:21:38] S2: Yeah, I think the way we look at this is the payback periods and the return profile. Right? And this is not a static thing. Right? Because we can, in a moment in time, we'll say, okay, here's what we think the return profile is for the cohort of investments of CapEx in '26, in 2026. But then as market conditions change, both the revenue that they can attain for that capacity, but also the cost. And this has been a big issue. Right? You've seen supply chain inflation, memory, all along the supply chain. And so the key is to continue to monitor this. This is an ongoing thing. What we see here is this. We've done this work. We we see high teens IRRs on on... And on compute capacity investment. We're talking about a three to four year payback period. That is longer than you would might hear out of a SpaceX, for example, where they're talking about one to two two year paybacks, maybe even shorter than that. But I think there's a difference here, which is Amazon and AWS is doing longer dated contracts, longer term agreements that are not right at the, you know, spot market price. So we... What we do is we track to make sure that over time, these return profiles are steady to improving. And what we see so far is, if anything, would say a bit of an improvement in return profile. Some of that I think is because you have... We talked about the training chips at Amazon. You talk about the TPUs at Google. I think these alternative sources of supply, proprietary sources of supply, are allowing some of the return profiles to improve for these hyperscalers.
00:23:33 · Akash Pasricha, Joelle Pineau
[00:23:33] S0: Great. Well, I want to thank you for coming on. That is Ken Gawrelski from Wells Fargo here on TITV. Canadian AI startup Cohere is looking to raise around $7,000,000,000 Canadian, that is roughly $5,000,000,000 U. S. In a new funding round that would value the company at roughly 26,000,000,000 U. S. Dollars. The company announced an initial close. I want to bring on Joelle Pineau, Chief AI Officer at Cohere for a conversation. Joelle, welcome back to the show. It's great to have you here. [00:24:05] S3: Good morning. Great to see you. [00:24:07] S0: So how much of this target have you actually raised so far? [00:24:13] S3: Well, we've actually raised the majority of the target. I'm not going to go into the details of where we are, but definitely, you know, we're well on our way towards that. [00:24:21] S0: But like six of the seven? Six of the seven maybe? [00:24:25] S3: I will leave it to my CFO to decide when he wants to share the numbers, but definitely, you know, more than half of that and and well on our way to the target and and and a lot of interest, frankly, in terms of people who are seeing the value in Sovereign AI, seeing the value in giving customers autonomy over how their technology is built, where it's deployed, where the data is stored, really resonates really well with investors. So we just want to be a little bit more out there in terms of how we're doing and progressing and what's our what's our target. And also really gives us an opportunity to talk about what we want to build with those investments.
00:24:25 · Joelle Pineau, Akash Pasricha
[00:24:25] S3: I will leave it to my CFO to decide when he wants to share the numbers, but definitely, you know, more than half of that and and well on our way to the target and and and a lot of interest, frankly, in terms of people who are seeing the value in Sovereign AI, seeing the value in giving customers autonomy over how their technology is built, where it's deployed, where the data is stored, really resonates really well with investors. So we just want to be a little bit more out there in terms of how we're doing and progressing and what's our what's our target. And also really gives us an opportunity to talk about what we want to build with those investments. [00:25:02] S0: So talk about that to them. What do you want to build? Because you spotlighted... I mean, the product roadmap is one thing, but you said invest in long term compute infrastructure, which is kind of interesting because Sovereign AI, I mean, part of the value prop for you guys is that they're running it on their own systems. So what exactly is the compute infrastructure that you want to build?
00:25:02 · Akash Pasricha, Joelle Pineau
[00:25:02] S0: So talk about that to them. What do you want to build? Because you spotlighted... I mean, the product roadmap is one thing, but you said invest in long term compute infrastructure, which is kind of interesting because Sovereign AI, I mean, part of the value prop for you guys is that they're running it on their own systems. So what exactly is the compute infrastructure that you want to build? [00:25:23] S3: First and foremost, we're really committed to our modeling roadmap, the ability to build frontier models. And in some cases, it's very specific capabilities. In the past, we've focused on multilingual capabilities. Now we have a lot of efforts around cybersecurity space as well as enterprise use cases, which can be quite different from some of the consumer AI. And so a lot of that infrastructure, obviously, to continue to have a really ambitious roadmap in terms of training larger models, generic synthetic data to train those models. But there's also definitely an appetite as AI grows to adopt our solutions for small medium businesses who may want to lean on us for compute and also, you know, to provide a greater number of inference solutions. So those are some of the things that will certainly come into the picture as we go. Right now, we've shared actually is our commitment to build some of that compute in Canada, which is part of the government of Canada's involvement in this funding round.