NTT DATA unveils Agentic SDI

Data-center operations are turning into an AI problem. Not just because companies want GPUs, but because the infrastructure around those GPUs has become messy — multivendor hardware, hybrid cloud, security controls, power constraints, and too many manual workflows. NTT DATA’s move this week is basically an attempt to put a conversational control layer on top of that sprawl. On April 28, it launched Software Defined Infrastructure Services Agent, or SDI Services Agent, inside its broader SDI offering for enterprise infrastructure. (servic([services.global.ntt)at did NTT DATA actually launch? This is not a new server rack or a new cloud. It’s an agentic service layer embedded in NTT DATA’s Software Defined Infrastructure services — a system that lets IT teams interact with infrastructure through a conversational interface while background agents handle orchestration, monitoring, and actions across complex environments. NTT says the service is built for multivendor estates, which is the important part, because most large enterprises do not run a single stack from one supplier. (services.global.ntt([services.global.ntt)entic”? Because NTT is not describing a simple chatbot. The company says the SDI Services Agent works as a multi-agent system that can surface real-time visibility, explain decisions, and trigger the right specialized agents in the background. That means the pitch is less “ask a dashboard for data” and more “ask the platform to understand, route, and act.” The catch is that plenty of vendors now use “agentic” loosely, so the real test will be how much autonomy customers actually allow in production. (services.global.ntt)cture does it touch? NTT is aiming wide. The company says the agent covers networking, hybrid data center operations, cybersecurity, digital workplace environments, and sustainability insights. That breadth matters because infrastructure failures rarely stay in one box anymore — a performance issue can be a network bottleneck, a policy conflict, a capacity shortfall, or a power-and-cooling problem wearing a different label. (services.global.ntt)ittle and more expensive. NTT had already been repositioning SDI as an AI-powered service model, including a June 2025 launch tied to Cisco infrastructure and software products. The new agentic layer looks like the next step — moving from analytics and automation toward a front end that can coordinate actions across the stack. Basically, NTT is trying to make infrastructure management feel more like operating software. (us.nttdata.com)es with mixed environments. That means companies with on-prem systems, multiple hardware vendors, some cloud footprint, and internal teams that are tired of break-fix operations. NTT’s own framing is that SDI should move customers away from reactive support and toward business-outcome-driven operations — better utilization, lower operating cost, faster time to value. That is classic services language, but it lines up with a real pain point in enterprise IT. (nttdata.com)ch is also a land grab for the control plane. If infrastructure gets managed through an agentic interface, the company that owns that interface gets closer to daily operations, policy decisions, and future AI spending. In other words, this is not just about making admins faster. It is about making NTT harder to displace in the enterprise stack. That’s an inference, but it follows from how deeply the service is meant to sit across multivendor environments. (services.global.ntt)uld readers watch next? Watch for customer names, scope, and guardrails. Product language about explainable AI and orchestration is easy; real deployments are harder. The useful signals will be whether NTT publishes case studies, whether customers let the agents take action rather than just recommend, and whether the service can prove savings in labor, uptime, or capacity planning. (services.global.ntt)ess — and that may be the smarter move. Enterprises do not need more boxes. They need fewer humans stitching the boxes together.