Clarence Reynolds, TelecomTV (00:02):
The race to implement AI Factory Solutions is accelerating across industries with telecom companies uniquely positioned to leverage artificial intelligence. So how can Telco's best capitalize on this technology? Mark Armstrong from Hewlett Packard Enterprise joins us to explore the opportunities. Mark, thank you for being with us today.

Mark Armstrong, HPE (00:21):
Pleasure.

Clarence Reynolds, TelecomTV (00:22):
So Mark, what is HPE doing to help customers with AI factory deployments?

Mark Armstrong, HPE (00:27):
So HPE building on solutions from partners such as Nvidia provide secure multi-tenanted AI platforms that help businesses to create services for internal use and also for external business to business use and business to consumer usage. So in essence, HPEI would describe as creating the engine really inside the factory to enable customers to take advantage of AI in all senses.

Clarence Reynolds, TelecomTV (00:58):
So which core components does HPE uniquely deliver in its AI factory solution?

Mark Armstrong, HPE (01:04):
Well, I mean, it starts with the expertise around helping customers to plan and design these things. Then it moves through to infrastructure, the servers, the storage, the system requirements that you need that are very, very specialized. We have acquired a number of organizations that give us a lot of intellectual property that differentiates our technologies. We have a lot of intellectual property that we develop inside HPE that really help customers to extract more value from the capital investments that they make. And then we work with partners. Nvidia is a great one in this telco. In this telco space. We take what Nvidia do and we build on top of it. So the technology, the GPUs, the Cuda software tooling, HPE has a number of tools above that that, again, extract more value. Ops RAMP is a really good example of that. Morpheus is another where we have an operational layer that sits above the infrastructure and creates operational services for our customers to distribute the service of AI at scale to its customers, either internally or externally in a multi-tenanted way, in a very secure way using sovereign data, which clearly businesses want to use, want to extract from their own data sources.

Clarence Reynolds, TelecomTV (02:34):
In this increasingly virtualized world, we're actually standing in front of a piece of hardware.

Mark Armstrong, HPE (02:41):
What is this that we're standing in front of? Well, I mean, this is one of our largest systems actually. I mean, if you are thinking about an AI factory, this is at the very top end. It uses significant direct liquid cooling to cool it. I mean, these systems are immensely, immensely large. So AI factories in a telco environment, let's just talk about a telco specifically. As I've said, they may require to deliver AI internally, externally, to other businesses as a service or to consumers. And they will build systems. They may start really small, they may start with a few GPUs, but we see them building significant at scale platforms, a hundred and twenty six, two hundred fifty six GPUs, 5,000 GPUs. And when you're getting up to that sort of scale, you're getting to a scale that needs this sort of innovation and this sort of technology to help run and cool systems like that. So HPE really provides the end to end, and we wrap that with a whole host of services that enable telcos to get on with doing what telcos do best, finding customers, delivering specific services. And we will provide the technology that underpins that. In an AI context.

Clarence Reynolds, TelecomTV (04:02):
How can a telco build a business case on AI services?

Mark Armstrong, HPE (04:06):
Well, I think if you just go back to the tenant, that information is the new oil. Telcos hold masses of information, their own information and customer's information, a lot of which is sovereign. It's very, very secure. And our systems enable telcos to mine that data and to create AI outcomes with that data. So the business case is really around, first of all, what data sources the telcos have that they can create value from in AI context, how do they develop and pipeline that data through to train models and to create inferencing applications that deliver the value ultimately of what that data has within it. So the business cases are huge and varied, but I believe we're at relatively early stage of the journey. And actually organizations are finding new use cases for AI applications on a daily, weekly basis. So the use cases just grow and grow.

(05:14):
The key thing right now I think is to act and to create these platforms that enable the extraction of value because these sorts of things take a significant amount of time to develop and build. And I think the leaders in the telco industry will be the ones that are developing these kinds of services first. And we're seeing that already a lot of first mover advantage in the industry with some organizations already building these kinds of systems. So I think the business cases will only increase in number and value, but we are definitely in a place where this is very serious and telcos are thinking now about what to invest when and how

Clarence Reynolds, TelecomTV (06:00):
It is an exciting time in our industry. Mark, thank you for being with us today.

Mark Armstrong, HPE (06:03):
It's an absolute pleasure. Thank you.