0:09
Everyone, this is Jordan Novitz from Semi Analysis.
Search this video
0:10
I'm here with Patrick O'Shaughnessy Legal from Radiant.
0:14
The uh the newly launched Neo Cloud. Uh yeah, welcome. >> Well, thank you. Very happy to be here. >> Awesome.
0:22
So, why don't we start with a little intro?
0:23
Just tell everybody uh where you're from and yeah, what's new with Radiant? It's brand new, right?
0:29
>> Yeah, it's uh it is brand new.
0:29
It's been announced only a couple of weeks ago.
0:32
So, yeah, I'm I'm Patrick.
0:32
Uh I'm an engineer at heart.
0:34
I've been doing high performance computing for for the past past 12 15 years.
0:40
And uh yeah, I joined the uh the Radiant adventure about, you know, 2 years ago before before it got announced as uh looking after the infrastructure.
0:46
So, basically deploying um supercomputers, you know, bringing them to production and you know, operating them.
0:54
And uh today I'm at Radiant.
0:54
I'm I'm the head of infrastructure.
0:56
Uh so, I'm looking after the infrastructure roadmap and procurement and things like that. So, yeah. >> Awesome.
1:03
So, talk me through the infrastructure roadmap a little bit.
1:04
I know you guys have big plans.
1:07
Um there is, you know, existing sites, but you're expanding globally. >> Mhm. Yeah, no, absolutely.
1:11
I mean, yes, we obviously we we basically Radiant's is is is a new venture, right?
1:16
It's been uh it's built around, you know, four key, you know, pillars.
1:20
You know, the the the software platform is one, the compute is is very important because, you know, this is the the the underlying infrastructure that we're uh we're bringing to market.
1:28
Uh we've got access to land.
1:30
We've got access to uh a lot of capital in order to execute.
1:35
And our infrastructure global roadmap is really global, right?
1:37
So, we've got multiple sites available to us, you know, that represent, you know, several hundreds of megawatts.
1:42
Uh and we're looking to to fill these sites with the latest Nvidia architecture.
1:46
So, being here at GTC is really it's really good because uh we get to see the the the technology very very closely, right?
1:53
We're going to see what what's It's up, to Rubin architecture and these kind of things.
1:59
Uh and also, you know, looking at all these uh providers doing all the ancillary equipment and CDUs.
2:03
So, combining all these technologies um is really nice and very much looking forward to deploying these new clusters globally. >> Awesome, yeah.
2:11
So, what do you think is uh I guess the key constraint?
2:13
You know, there's uh you said you got four key pillars.
2:17
Is there things that you think you'll have to be working on over the course of this year and next year as you guys try to expand? >> Yeah, I know.
2:24
I mean, we're working on all all all angles, right?
2:26
Uh our software platform is, you know, we we've got an engineering team that keeps developing and bringing new capabilities in the in in that.
2:34
Uh the infrastructure, you know, building the new designs, getting ready for for what's coming is also a very important angle.
2:39
And uh I think that one of the key constraints until until the region region announcement, right, was access to to land, access to power.
2:45
Uh and the capital to to deploy this.
2:48
So, as uh the ability for us to to merge with Brookfield uh is incredibly exciting because that resolves that problem, right?
2:58
Um and we now can basically take greenfield sites uh with planning permission.
3:02
Uh we've got the power plants in order to to supply the energy, which is, you know, one of the key bottlenecks in in the industry.
3:09
Uh and we can now, you build these data centers and and bring this uh this infrastructure to market, you know, in record times in, you know, about 12 months we can go from uh you know, a field to a fully productive AI environment delivering services.
3:22
So, it's uh it's incredibly it's incredibly exciting and uh yeah.
3:29
>> Can you talk me through maybe the Brookfield portfolio across how that that helps?
3:33
I know, you know, going from greenfield or dirt to data center requires many many different organizations to contribute something.
3:44
Um equipment, like you said, permitting.
3:48
Are there, you know, special things that you're doing with the Brookfield portfolio there? >> Yeah, no, absolutely.
3:52
I mean, this is this is one of the key aspect, you know, that Brookfield is bringing with the in this merger, right?
3:57
Um so, we are part of the Brookfield AI fund, uh which has quite a lot of companies, and I couldn't name them them all, right?
4:06
Um but uh we're obviously looking to partner with uh with these companies.
4:11
We're also looking to partner with a number of data centers that also own and operating with uh with Brookfield.
4:15
Uh which means that, you know, we can we can start talking to them now about the future capacity that they're going to bring online the next, you know, 24, 36 months, uh and basically design these these environments for the upcoming infrastructure that is that is being, you know, rolled up by Nvidia.
4:32
I mean, we we've had discussions about, you know, GPU racks of, you know, 400 kW as soon as, you know, January next year, right?
4:41
And all the data centers need to be ready to to be able to accept these these new infrastructure.
4:45
So, being able to collaborate with these players that are part of the the the Brookfield portfolio uh puts us in a really unique position in order to to execute on that. >> Yeah.
4:55
How about the sites themselves?
4:57
So, let's talk about maybe power for a second.
4:59
Um the big discussions this week at GTC is the difference between grid-connected versus behind-the-meter power generation.
5:05
Do you have preferences there?
5:07
Are there ways in which you can move faster versus, you know, have more durable sites over time that you're thinking about?
5:15
>> Yeah, that's a very good question.
5:15
I think that um connection to the grid is definitely, you know, one one aspect.
5:20
Um but I think that it can be slower, right?
5:25
Uh and Brookfield brings that ability to get behind the meter uh power generation.
5:30
And when we're looking at these new sites, we're definitely looking about, you know, how we're going to bring that infrastructure, how we're going to to to get that energy.
5:37
So, yeah, I think that we've we've got all the cards, right, uh at the moment. So, yeah, it's exciting. >> Okay, cool.
5:46
Uh maybe can you talk a little bit of about talent?
5:49
Uh, you know, you guys are obviously bringing a lot of talent at the software layer.
5:53
Um, you know, what what does that look like maybe across the stack when it comes to actually providing a service to an end user that's going to come in via tenant at one of these sites?
6:00
You know, that off-taker that people keep talking about. >> Ooh, good question.
6:07
So, yeah, I mean, our software stack is very comprehensive, right?
6:09
I mean, we've got a team dedicated on the lower-level platform, which is basically the ability to to take this bare-metal infrastructure and turn it into compute being, you know, bare-metal services or virtual virtual machine or Kubernetes environment.
6:26
And we've got so a team that is doing this, which is what we call the the Radiant Cloud OS, which is the lower-level platform.
6:32
And we are looking to bolster that team manifold.
6:37
And then there is, you know, all the applications that come at a higher level around data science and model registry and inference services, etc.
6:43
, which really augment that infrastructure for for our end users.
6:48
So, we also have already a quite quite a significant team doing these things.
6:53
And yeah, we'll we'll bring certainly more more of them over the the coming months, that's for sure. >> Makes sense, yeah.
7:00
What do you think that breakdown is between like a a large-scale bare-metal service that you might provide to, you know, a very large customer versus the paradigm where you've got many small customers who want to consume smaller services?
7:15
>> It's it's an interesting one.
7:15
Um, I don't think that there is so much of a difference between the two because even when you're talking to one very large customer, you may still have many teams operating within that organization who would have different needs and different expectations.
7:33
And basically the idea with with our software platform is the ability to to re-provision the infrastructure and reshape it based on the needs of the customer, right?
7:43
And based on you know, we've been engaged with a lot of enterprise customers and you know, one day they want to do lots of training.
7:49
So they will be provisioning the infrastructure as a big monolithic kind of Slurm bare metal environment and just run these batch batch nodes and this batch and this batch and jobs.
8:03
And you know, maybe a few few months later or a few weeks later they want to run inference, right?
8:07
So you want to decommission this and provision it in a in a different fashion and our our software stack is built to handle that level of flexibility.
8:15
So whether it's one big customer that you know, has multiple little teams or whether it's multiple customers that are coming from a public cloud angle for instance, from our perspective it's the the the same software stack that that that will do this. >> Awesome, yeah.
8:30
And earlier you talked a little bit about, you know, monitoring or reliability and how that helps.
8:33
I think that's a big discussion for many of these big systems.
8:37
They're very expensive, they're hard to run.
8:38
If they're down it can be even expensive because you're paying for something you can't use.
8:43
Um maybe can you talk me through anything interesting that you guys might be doing in terms of like reliability and uptime and how vertical integration to these sites that are all under one roof might be different than others that have to bring a bunch of different partners to the table and and have contracts instead of you guys are just you know, it's just all employees under the same umbrella if that makes sense. >> Yeah, no absolutely.
9:08
So in in the region Cloud OS we've got multiple plugins that I mentioned the lower level platform, the AI application services.
9:14
There's also as you as you rightly point out this whole other angles is kind of more more vertical about observability, monitoring, logging, triggering and alerts, etc.
9:23
So yes, in in our software we we have this capability, right?
9:27
we we have this capability, right? Which basically will feed on data, we will capture all the data from you know, the servers, the switches, the equipments, the racks, even the data center, and compile this into actionable um
9:41
you know, alerts that we can actually operate and uh and that allows us to effectively ensure that we we meet like very, very high SLAs in terms of uptime and availability of the systems, uh which is what enterprise customers, you know, widely ask and and deserve to get, right? Um so, this this this is
9:56
Um so, this this this is critical for us and because at the end of the day, right, you for for successful success, you want to make sure that your customers remain happy and the way that you make them happy is make sure that, you know, the the service remain available at all time. Um so. >> Awesome.
10:12
Well, this has been a great interview.
10:14
Thanks for taking the time.
10:14
Again, this is Jordan and Patrick.
10:15
Uh I'm signing off for Semi Analysis and Sale from GTC 2026. Thanks for watching. Thank you.