10 Lakh AI Jobs, But Only 16% Of IT Workers Have AI Skills!

India does not have an AI jobs problem. It has an AI talent problem. That distinction matters. Companies are already hiring for artificial intelligence and related skills, but finding people who can actually build, deploy and run these systems is proving harder. TeamLease Digital’s Digital Skills & Salary Primer FY2026-27, based on an analysis of 37,000 technology and digital roles, estimates a roughly 53% talent supply gap in GenAI and a 55-60% gap in cloud. Cybersecurity has a 25-30% gap. Only around 16% of India’s IT professionals, according to the report, currently have AI skills.

The report also puts the scale of the training problem into perspective. Around 20 lakh professionals have been upskilled, TeamLease says, but only about 3 lakh have advanced AI skills. The demand for GenAI talent is expected to cross 10 lakh roles by 2026.

The 53% figure should not be read as 53% of AI jobs being vacant. TeamLease is measuring the gap between the skills employers need and the available talent supply, and the report notes that some of its percentages are directional estimates. But the broader trend is clear: demand for AI capability is rising faster than the supply of people with the skills to deliver it. Recent reporting on the TeamLease study has highlighted the same 53–60% gap.

This isn’t just a shortage of AI engineers

One reason the talent shortage is easy to misunderstand is that the AI workforce is much bigger than people writing models.

TeamLease divides AI-related jobs into three broad layers. AI-Core covers roles such as GenAI engineers, ML engineers, LLMOps engineers and AI product owners. AI-Adjacent includes data engineers, cloud engineers, cybersecurity engineers, analytics engineers and platform SREs. AI-Support covers operational roles such as data-quality and application-support professionals.

This matters because an AI model cannot operate in isolation. It needs data, computing infrastructure, deployment systems and security. That makes shortages in cloud, data engineering and cybersecurity part of the same problem.

The Economic Survey 2025–26 makes a similar point from an education and skills perspective. It says building and scaling AI applications requires both algorithmic expertise and software-engineering skills. It also argues that much of the necessary knowledge comes from hands-on experience with actually building models, and suggests stronger industry-academia links, apprenticeships and pathways for practitioners to contribute to training.

Hyderabad is already paying for the shortage

The shortage becomes more visible when you look at where companies are willing to pay a premium.

TeamLease’s salary data for professionals with three to 10 years of experience puts Hyderabad’s median fixed pay at ₹34.5 lakh for GenAI and LLM engineering, ₹30.5 lakh for ML engineering and MLOps, and ₹28 lakh for applied data science, NLP and computer vision. Data engineering comes in at ₹18.8 lakh, cloud and DevOps platform roles at ₹15 lakh, and cybersecurity and cloud security at ₹13.8 lakh.

That puts Hyderabad close to Bengaluru in the premium end of the AI talent market. TeamLease places the corresponding GenAI and LLM median at ₹37.5 lakh in Bengaluru and ₹32 lakh in Mumbai, and identifies Hyderabad as one of India’s premium metro hubs for AI, data, cloud, cybersecurity and GCC roles.

The city’s GCC expansion adds another layer to this. Hyderabad accounted for 16.8% of GCC office leasing among the major Indian markets in the first quarter of 2026, according to the TeamLease report.

India is training more people. The harder part comes next.

The government has also been expanding the formal skilling pipeline. FutureSkills PRIME covers areas including AI, big data, cybersecurity and cloud, while the government’s AI-focused skilling initiatives have expanded into additional qualifications and training programmes. By August 2026, FutureSkills PRIME had recorded more than 34 lakh registrations and over 13 lakh certifications, according to the government.

But training volume does not automatically translate into production-ready talent. Someone who has completed an AI course may still lack the experience required to take a model into production, build the surrounding infrastructure, secure it or operate it at scale.

That distinction is becoming increasingly important as companies move from experimenting with AI to putting it into their everyday operations.

The awkward part: even the entry-level ladder is changing

There is another consequence buried in the transition.

The report identifies repetitive reporting, manual testing, basic operational support and other rule-based work as areas where technology roles are being compressed. At the other end, roles around AI-native products, data foundations, cloud infrastructure and secure AI systems are expanding.

That creates a strange situation for India’s technology workforce. Companies can simultaneously have a shortage of experienced AI talent and fewer conventional entry-level tasks through which workers traditionally gained experience.

For Hyderabad, therefore, the AI opportunity is not simply about producing more people who know how to use an AI tool. The bigger requirement is a workforce that can build, deploy, secure and operate AI systems, and combine those skills with knowledge of an actual business or industry.

India already has one of the world’s largest technology workforces. The challenge now is turning that scale into the deeper, production-level expertise the next phase of AI adoption will require.