NTT DATA Will Reskill 40,000+ India Employees Into AI-Native Developers
NTT DATA is making a clear move: turn its India workforce into AI-native developers and scale generative AI delivery across global projects. Leaders from the company say India's engineering depth, cost expectations, and data confidentiality needs will anchor how they build and deploy AI at scale.
"India is one of the most important countries for the future… our plan is to reskill all of them to be AI-native developers," said Carlos Galve, Head of the Global Generative AI Office, NTT DATA Corporation.
What's Changing Inside NTT DATA
The reskilling program targets productivity, not just training hours. Kenji Motohashi, Head of the Global Generative AI Office, NTT DATA Group Corporation, framed the goal simply: the same headcount should be able to support many more projects-an order-of-magnitude increase over time.
NTT DATA already concentrates R&D in Japan and Silicon Valley. There's no immediate plan for a core R&D center in India, though leaders left the door open for future software development expansions.
India-First Delivery: Cost and Confidentiality
Pricing matters in India. Motohashi noted that OpenAI is offering services, including ChatGPT Enterprise, at significantly lower cost for Indian customers-around 70% less than other markets-so solutions can meet local budgets.
Data control matters just as much. Some Indian clients prefer private or proprietary infrastructure rather than shared environments to protect sensitive data. Expect on-prem, VPC-isolated, or sovereign deployments to be part of standard solution design.
Support for GCCs
NTT DATA plans to double down on Global Capability Centres (GCCs) in India. The approach: embed trained Indian AI engineers to support client GCCs on site and remotely, turning them into a high-capacity delivery engine.
What This Means for Engineering Leaders
- Measure productivity by business outcomes and projects per team, not just velocity or story points. Token cost, model latency, and prompt throughput become core metrics.
- Build an AI stack that fits India-specific constraints: cost controls, private data paths, and compliance guardrails.
- Expect hybrid delivery: public APIs where appropriate, plus private inference and isolated data stores where required.
Skill Map: From Developer to AI-Native
- LLM integration: function calling, tool use, agents, retrieval-augmented generation (RAG).
- Data layer: embeddings, vector databases, document loaders, chunking strategies, PII redaction.
- Quality and safety: prompt testing, evaluation harnesses, guardrails, red-teaming.
- Ops: token and cost budgeting, observability, latency tuning, caching, model/version management.
- Platforms: secure SDKs, model gateways, policy enforcement, audit logging.
Infra Choices You'll Need to Make
- Where inference runs: public API, private cloud, or on-prem GPU clusters.
- How data moves: zero-copy patterns, encrypted stores, masked analytics, and strict role-based access.
- How to control spend: prompt optimization, response truncation, batching, and caching at scale.
Regulatory View: India vs Europe
Company leaders noted India's regulatory posture as more enabling than Europe for AI deployment. That's shaping go-to-market decisions and the speed of build-out in Indian GCCs.
What's Next: Photonics, Quantum, and Bigger Models
NTT leaders flagged photonics and quantum as the next wave that will hit data centers and talent models. They expect major leaps in capability over the next 5-10 years, with broader discussions around artificial general intelligence and advanced robotics.
If you're planning roadmaps beyond 2026, keep an eye on the NTT IOWN vision and other photonics-first architectures. This will influence how you plan networks, latency budgets, and compute placement.
Practical Next Steps for Teams
- Stand up an internal AI enablement program with a clear skills ladder and weekly delivery targets.
- Standardize a model gateway, evaluation harness, and cost observability early to avoid tech debt.
- Offer both public-API and private-inference options in every client solution, with data-handling patterns documented.
- Pilot with one GCC, prove the throughput gains, then replicate the playbook.
If you're upskilling engineers now, start with focused, job-ready training paths rather than generic theory. Curated tracks for developers and data teams can speed this up: AI courses by job role.
Bottom line: NTT DATA is turning India into a scaled AI delivery hub. For IT and engineering leaders, the advantage goes to teams that ship usable AI fast, keep costs predictable, and protect data by design.
Your membership also unlocks:
