Beyond the GPU: What NTT DATA and Blackfuel are building for the AI boom

AI has an appetite problem. As more organizations move AI into production, demand for inference — the computing work that’s required every time an AI model generates an answer or takes an action — is soaring. And providing that capacity in a reliable way takes much more than filling a data center with GPUs.

The infrastructure needs to combine compute, storage and high-performance networking. It also has to keep sensitive data secure and stay constantly available, all while dealing with enormous power and cooling requirements.

Then somebody has to integrate all those moving parts and keep them running. That’s where NTT DATA comes in.

We’re working with Blackfuel, an AI-native “neocloud” provider, on an ambitious new approach to AI infrastructure, with multibillion-dollar expansion plans and a major first deployment this year. Our role spans integration and operational readiness, as well as the ongoing operation and maintenance of the compute infrastructure.

There’s also a much bigger opportunity for us: helping a new generation of AI infrastructure providers turn extraordinarily complex technology into dependable, production-ready platforms.

• ALSO READ - Your AI strategy is only as strong as the infrastructure it runs on

The rise of the neocloud

Blackfuel is part of an emerging category of providers often called neoclouds: companies building cloud infrastructure specifically for demanding AI workloads. For these providers, acquiring AI accelerators is just the beginning.

Building an AI factory is very different from buying a server. Compute, storage, networking, cybersecurity, cooling and software all have to work together. The platform needs to be secure and resilient from day one, and when something fails, there needs to be clear responsibility for finding and fixing it as fast as possible.

As a full-stack AI infrastructure integrator, NTT DATA brings those pieces together across compute, storage, networking, security, operations and lifecycle management. We also work with technology partners, including Dell Technologies and AMD, to help providers move from infrastructure on paper to infrastructure that performs reliably at industrial scale.

Blackfuel shows what that can make possible.

What if you could buy AI output instead of infrastructure?

Blackfuel’s first inference-only deployment brings together serious computing power: liquid-cooled Dell Technologies PowerEdge systems powered by AMD Instinct™ MI355X accelerators, integrated by NTT DATA.

But the infrastructure itself is only half the story. What Blackfuel does with it is where things get interesting. Rather than requiring every customer to build and operate their own GPU environment, they let organizations reserve the AI output they actually need. Customers can match a model and capacity to their specific throughput, latency and data-location requirements. They can even reserve a fraction of a model’s serving capacity for a fixed monthly fee instead of renting an entire multi-GPU deployment.

Building AI capacity closer to home

There’s another reason this model is attracting attention: sovereignty. Organizations increasingly care not only about how much AI capacity they can access but also where their workloads run, where their data resides and who operates the underlying infrastructure.

Blackfuel’s model combines European-operated infrastructure with customer control over model selection, deployment location and capacity commitments. This presents a valuable opportunity for NTT DATA to turn those requirements into infrastructure and managed operations that deliver AI-enabled business outcomes reliably in the real world.

We turn AI investment into business results

As AI investment accelerates, organizations need to turn that spending into business results. That starts with secure, scalable infrastructure that can take AI beyond experimentation and support it reliably in production.

Blackfuel has an ambitious idea for turning the industrial machinery behind AI into a service that customers can consume more predictably. NTT DATA’s role is to bring the pieces together and make sure that machinery works when customers depend on it.

• ALSO READ - Infrastructure is critical again: Here’s what it means for every CIO

WHAT TO DO NEXT

Read more about NTT DATA’s partnership with Blackfuel and how it will help organizations deploy AI more efficiently.