AI in 2026: Understand the value spectrum to get ahead

Asutosh Mohapatra, Head of AI Strategy: Global AI Office at NTT DATA, and Ankur Dasgupta, Executive, Head of Marketing: India at NTT DATA

It almost seems hard to believe, but AI turns 76 this year, tracing its roots back to Alan Turing’s seminal paper titled Computing Machinery and Intelligence, published in 1950.

Since then, AI has progressed from theory to large-scale implementation — a progression that seems to be getting exponentially faster, raising new possibilities and concerns almost daily.

Now, 2026 feels like a turning point: a good time to assess where you are by evaluating where your various AI initiatives lie on the AI value spectrum.

How many have reached a point of established value, delivering measurable productivity and clear ROI? Where are you stalling at the point of correction, chasing use cases that don’t scale beyond experimentation? How are regulatory, strategic and market forces driving a new set of focused, high-priority use cases that are now gaining traction at the point of emerging momentum?

Understanding where your organization sits on this spectrum, and how these various placements interact, is a useful way to determine what’s really working and what isn’t so you can make a clear case for ongoing investment in AI.

3 prominent positions on the AI adoption spectrum

Let’s take a broad look at where the biggest AI technologies and trends fall on this spectrum.

1. At the point of established value: AI infrastructure and operations

The current state: Hyperscale data centers, direct-to-chip cooling, FinOps for GPUs and managed large language model (LLM) operations have moved out of the lab and into the ledger. They have reached a point where productivity expectations are steady and performance benchmarks are holding up over time.

The NTT DATA view: This is where infrastructure maturity becomes a competitive advantage, and we’re investing significantly to support growing demand in this area.

We operate over 160 data centers globally, including 22 in India, making us the largest data center provider in the country. Each of these facilities is either AI-ready or already supporting AI workloads.

We’re investing $1.5 billion to expand our data center capacity in India to beyond 800MW over time, with large-scale GPU clusters designed for AI workloads.

As we continue to scale, long‑term efficiency and sustainability become critical. Across the Americas and India, our high-density deployments are reaching multi-megawatt scale, using direct-to-chip cooling to manage heat at the source and support racks approaching 100kW.

• ALSO READ → PhonePe drives efficiencies with sustainability in focus

2. At the point of correction: Horizontal AI without support

The current state: After an initial burst of experimentation — DIY copilots, loosely defined “innovation labs” and a scramble to try the latest models — many organizations are now reviewing their early expectations against the pointed question: Where’s the return?

This is particularly noticeable with horizontal AI, the general-purpose AI capabilities that are reusable across many functions, teams or industries. Operating models and data foundations haven’t caught up with technological progress, and organizations have hit a slump, forcing them to assess what’s practical to deploy at scale right now.

The NTT DATA view: “Enterprise-scale AI adoption is still in its early days,” says Avinash Joshi, Executive Managing Director: India at NTT DATA. “While global capability centers (GCCs) in India are becoming global innovation engines, fewer than 20% of GCCs use AI as a core capability today.”

The demand exists, but the structure to support it isn’t in place. This mismatch is common in services-led ecosystems, where adoption tends to follow proven breakthroughs rather than create them.

3. At the point of emerging momentum: Sovereign and regulated AI

The current state: Data sovereignty, industry-specific regulations and national AI agendas are creating new demand for localized AI, particularly for inference and fine-tuning closer to where data is generated.

The NTT DATA view: In this position, focused, policy-backed use cases are quickly advancing. Momentum is being determined as much by regulation as by innovation, with governments and industries influencing where AI investment flows next.

According to NTT DATA research, 95% of organizations say private and sovereign AI are important. However, far fewer are ready to act on it. Some are redesigning how their AI works: where it runs, how it’s governed and how private and sovereign requirements are built into the architecture from the start. Others are still trying to layer AI into systems that were not designed for that level of control, locality or data-flow constraint — and the difference between the two is starting to show.

• ALSO READ → Sovereign AI, cloud and data resilience: The next frontier

Where you create value depends on your position on the spectrum

To move your organization to a different point on the AI value spectrum, you need to decide where to compete and which strengths to lean on, based on where you already are.

As NTT DATA, Inc., CEO and Chief AI Officer Abhijit Dubey puts it: “AI is a massive secular trend … It’ll be a short-lived bubble, and AI will come out of it stronger.”

He adds: “The biggest issue with AI in terms of getting value is not the technology. The technology is there. It’s about whether a company’s workforce is ready and if they have the data architecture set up.”

That’s the real divide. For full-stack AI players, the challenge is to navigate all three points on the spectrum at the same time. For IT services and global systems integrators, it’s about helping clients move from the point of correction to established value at speed. For industry-focused AI players, it comes down to proving value before the next round of investment.

• ALSO READ → Beyond the hype: How AI is delivering real business results

There is no single path to AI value

At NTT DATA, we work across the AI value spectrum because the market itself isn’t in one place. Different parts of the enterprise ecosystem mature at different speeds, and real value comes from meeting them where they are, rather than forcing a single path:

• Our infrastructure business sits at the point of established value, with steady demand, well-defined SLAs, and uptime that’s measured and delivered.
• Our services business helps clients move past the point of correction by managing risk early in AI projects, proving what works and scaling it with confidence.
• Our global innovation network supports what’s emerging, from GCC-led initiatives to sovereign cloud and regulated AI environments.

Much of the market is still operating at the point of peak hype, promising larger models and more agents. But hype rarely delivers value.

Our view is this: Use your position on the spectrum to your advantage. Invest in what’s proven to create stability. Partner with experts to move past the point of correction faster. Build early where it counts to stay ahead.

Find the gains you can build on now

For enterprise leaders, the conversation is now about where AI will deliver lasting value. Spending is following that shift, away from broad experimentation to where gains can build over time. Horizontal AI, in particular, is being reassessed through a more practical lens: governance, data readiness and operating maturity. Investment continues to flow to the foundations that make AI dependable at scale: computing, power, security, cost control and model operations.

Sovereignty and regulatory requirements are also influencing how priorities are set. Instead of defaulting to global, one-size-fits-all platforms, organizations are leaning toward more localized and controlled deployments.

If previous technology cycles tell us anything, what endures isn’t the initial wave of excitement but the systems that make the technology dependable. Infrastructure, operating rigor and trust frameworks rarely draw attention, but they transform AI from a series of pilot projects into something your organization can rely on.

To move forward with intent, you need to be clear about which part of the spectrum you’re investing in — and what that position demands.

• ALSO READ → From AI ambition to real impact: Why infrastructure makes the difference