CALIFORNIA: India is racing ahead on artificial intelligence. Three out of four enterprises say their AI bets are paying off. Investment is accelerating faster than the global average. Yet beneath the success story, a quieter risk is building: data infrastructure is straining under the weight of scale, complexity and security demands.
That is the central finding of Hitachi Vantara’s 2025 State of Data Infrastructure Report, which shows Indian organisations deploying AI more aggressively than their global peers, but struggling to keep their data foundations fit for purpose.
Based on a survey of 1,244 business and IT leaders across 15 countries, including 104 in India, the report paints a picture of momentum colliding with operational reality.
India outpaces the world on AI returns
Indian enterprises have largely moved beyond pilots. A striking 89 per cent have either widely adopted AI or made it critical to operations, far ahead of the global average of 69 per cent. Nearly two-thirds, 63 per cent, rate themselves as strong or established in achieving return on investment from AI, again beating global benchmarks.
“Indian enterprises have moved well beyond experimenting with AI and are now focused on delivering measurable business outcomes at scale,” says Hemant Tiwari, managing director, Hitachi Vantara, India and saarc. “What’s becoming clear is that as adoption accelerates, infrastructure complexity is rising just as fast.”
That complexity is rising faster in India than elsewhere. As many as 87 per cent of Indian organisations say their data infrastructure is becoming rapidly more complex, compared with 80 per cent globally.
Data growth turns punishing
The numbers explain why. AI investment in India is expected to jump 75.6 per cent over the next two years, outpacing the global projection of 70.3 per cent. Data storage needs are forecast to rise 73.9 per cent, again above the global average of 68.9 per cent.
Four in ten Indian organisations now manage between 50 and 200 petabytes of data, a scale at which inefficiencies quickly become painful. Almost half store operational data in public cloud environments, with even higher usage for general business data, creating fragmented, hybrid setups that are harder to govern, secure and optimise.
Security takes centre stage
As AI spreads deeper into core operations, security has become the dominant concern. Two-thirds of Indian enterprises, 67 per cent, cite data security as a top challenge in AI implementation, placing India among the most security-conscious markets globally.
Yet Indian firms are responding with stronger governance than many peers. Some 81 per cent report executive teams with clearly defined AI visions, compared with 71 per cent globally. Nearly four in five have dedicated AI or machine learning teams, defined KPIs and expected business outcomes.
Operational discipline is also higher. Around 71 per cent rate their mlops capabilities as strong and established, versus 63 per cent globally. Advanced governance models are in place at 63 per cent of firms, compared with 49 per cent worldwide. Sustainability considerations are embedded in infrastructure decisions at more than half of Indian organisations.
Success, but not for everyone
Overall, 75 per cent of Indian organisations report successful AI outcomes, with virtually none reporting outright failure. The most common use cases are workflow automation and insight generation, each cited by 29 per cent, followed by productivity and decision-making improvements at 26 per cent. High-quality data, monitoring, employee adoption and skilled teams emerge as the biggest success drivers.
Still, the report flags a growing readiness divide. While 55 per cent of Indian enterprises have reached managed or optimised data infrastructure maturity, the remaining 45 per cent operate on weaker foundations, making AI harder to scale and more expensive to run.
Only 32 per cent have predictive, automated and cost-efficient infrastructure scaling in place, raising questions about how sustainable current returns will be as data volumes explode.
“The real bottleneck for AI isn’t the models, it’s when the data backbone lags behind,” says Sanjay Agrawal, cto and head of pre-sales, India and saarc, Hitachi Vantara. “As Indian organisations operate at higher data volumes across hybrid environments, complexity and security risks increase rapidly.”
Talent gaps push firms to partners
People, not just platforms, are part of the problem. More than half of Indian organisations cite hiring skilled workers as a major AI challenge. To cope, 76 per cent are turning to partners or outsourcing key parts of their AI and data initiatives, well above the global average.
That reliance underlines both the opportunity and the risk. India’s enterprises are clearly ahead in ambition and execution, but sustaining the lead will require faster modernisation of data infrastructure, sharper governance and deeper talent pools.
India may be winning the AI race for now. Whether it stays ahead will depend less on algorithms and more on fixing the pipes beneath them — before success itself becomes the constraint.
