AI Is Making Half of TCS's Deals Cheaper: What It Means for Indian IT and Your Career

When India's largest IT company says AI is pushing prices down on half its contracts, it is worth stopping to listen. That is roughly what TCS chief executive K. Krithivasan said after the company's second-quarter results.

Today is October 10, 2026. Indian IT employs millions of people and earns a large share of the country's export income, so a change in how this industry gets paid is not just business news. It affects graduates looking for their first job, mid-career engineers, and every company that sells services by the hour.

What the TCS CEO actually said

In an interview with CNBC-TV18 after the Q2 results, Krithivasan made four main points.

About half of TCS's IT contracts now see "AI-led deflation". Deflation here simply means the price of the work is going down. If a foreign client used to pay 100 rupees for a piece of software maintenance, testing or coding work, AI tools now let the same work be done for roughly 10 to 20 percent less, around 80 to 90 rupees.

This was expected. He said the industry has been discussing this effect for about a year and a half to two years, so a figure of around 50 percent is not a surprise.

New projects should balance it out. Many people only think about AI in terms of productivity and lower costs. He pointed out that businesses are also starting brand-new projects because of AI, which he called net new work, and that this new demand should offset the deflation fairly soon.

Not every deal is shrinking. In large, complex AI engagements, the contract value has not fallen. It has held steady.

The interview also noted that TCS's margins, the share of revenue it keeps as profit, have remained healthy.

Why this is a challenge for India

IT services are one of the pillars of the Indian economy, so the downsides deserve a clear look.

The old billing model is under pressure. For decades, Indian IT firms have charged on a "time and material" basis, which means billing for how many people worked how many hours. If AI finishes a task in less time, the old way of billing earns less for the same result.

Fewer easy entry-level jobs. Routine work such as basic data entry, manual testing and simple bug fixes is exactly what AI now does fast. That makes the old pattern of mass hiring fresh graduates into such roles much harder to sustain.

Risk for people who stopped learning. Engineers who have done the same routine task for years without updating their skills will find it harder to hold on to their roles.

Why it is also a big opportunity

The same shift opens doors.

Moving from low cost to high value. India does not have to be known only as a source of low-cost IT services. Indian engineers can lead the world in building complex AI systems, cloud architecture and cybersecurity solutions.

Better margins for companies. If smaller teams with AI help finish more work faster, profit margins can rise, and TCS's own margins are already holding up well.

A boost for startups. One developer with good AI tools can now do work that used to need a whole team. That makes it far easier to build new tech products and startups in India.

In one line: this is a warning and an opportunity at the same time. It is bad news for anyone who stays in the old way of working, and a huge opening for engineers and companies that adopt AI and keep raising their skills.

The skills that will keep you in demand

Demand is falling for people who only write code or do basic tasks. Demand is rising fast for engineers who understand complex system design, deployment, security and real business problems. Here is a practical path, step by step.

Step 1: Become an AI builder, not just an AI user

• Learn to design applications on top of large language models (LLMs, the AI models behind chat assistants) with frameworks such as LangChain, LlamaIndex or Semantic Kernel.
• Understand RAG (retrieval-augmented generation): connecting a company's private documents to an LLM, using vector databases such as Pinecone, Qdrant or Chroma, and embeddings, which turn text into numbers a computer can compare.
• Explore agentic workflows, where several AI agents split a task among themselves, using tools such as CrewAI or AutoGen.
• Learn to evaluate model output, reduce hallucinations (confident but wrong answers), and fine-tune small models for a specific domain.

Step 2: Learn modern DevOps, platform engineering and MLOps

• Build CI/CD pipelines (automated build, test and release) that use AI tools and deploy to the cloud automatically.
• Practise platform engineering: internal developer portals and self-service infrastructure with tools such as Terraform or OpenTofu, and Pulumi.
• Learn MLOps / LLMOps: running AI models in production at scale, controlling latency, managing GPUs and monitoring models.
• Get comfortable with AWS, Azure or GCP, including cloud cost control (often called FinOps), serverless designs and Kubernetes scaling.

Step 3: Add AI security and governance

• Study the OWASP Top 10 for LLMs, which covers risks such as prompt injection (tricking an AI with crafted input), data leakage and unsafe handling of AI output.
• Learn how privacy and compliance rules apply to AI systems, including GDPR, India's DPDP Act and ISO 27001.
• Secure the software supply chain by scanning third-party packages, libraries and container images for known vulnerabilities.

Step 4: Grow into system architecture and domain knowledge

• AI can write small functions, but designing a large enterprise system that is scalable, fault-tolerant and fast still needs an experienced architect. Practise high-level and low-level design.
• Pick a domain, such as banking, fintech, healthcare or e-commerce, and learn its real problems well enough to design technology that solves them.

Step 5: Master AI-assisted development

• Use AI coding tools in your editor or terminal, such as Cursor, Claude Code or GitHub Copilot, to work several times faster.
• Sharpen your code review and debugging skills, because someone still has to spot what the AI got wrong and where performance suffers.

The shift in one table

The old approach (at risk)	The future approach (in demand)
Only writing code	Solving problems as an architect
Manual testing and deployment	Automation and platform engineering
Staying inside one technology	Understanding end-to-end systems and security

Conclusion

TCS's comment that about half its contracts now face AI-led deflation is a clear signal of where Indian IT is heading. Prices for routine work are falling, and the old way of billing by the hour is under strain. At the same time, new AI projects are arriving, complex engagements are holding their value, and engineers who can design, secure and run AI systems are becoming more valuable. The engineers who keep learning will come out ahead.

Merits

• Lower costs make Indian IT services more attractive to global clients.
• New AI projects create fresh demand that can balance falling prices.
• Engineers can move from routine work to higher-value design and architecture roles.
• Better margins give companies room to invest in skills and products.
• Small teams and solo developers can now build real products and startups.

Demerits

• Hour-based billing earns less when AI finishes work faster.
• Entry-level hiring for routine tasks is likely to shrink.
• Professionals who do not reskill face a real risk to their jobs.
• The benefits may reach large firms and skilled engineers first, widening gaps.
• New skills take time and effort to learn while the market changes quickly.

Caution

This article is for educational purposes and summarises a TV interview and an independent analysis in plain language. Figures such as the 50 percent share of contracts and the 10 to 20 percent price change come from the CEO's remarks as reported, and may be rounded or change in later results. Tool names are examples, not endorsements. Please check the company's official statements and do your own research before making career, investment or business decisions.

Frequently asked questions

• What did TCS's CEO say about AI and pricing? — K. Krithivasan said about half of TCS's IT contracts now see AI-led deflation, meaning the same work costs clients roughly 10 to 20 percent less.
• What does AI-led deflation mean? — It means prices for a piece of work fall because AI tools let it be done faster or with fewer people.
• Will AI reduce jobs in Indian IT? — Routine entry-level roles are likely to shrink, while demand grows for engineers with AI, cloud, security and architecture skills.
• Are all TCS contracts losing value? — No. The CEO said large, complex AI engagements have held their contract value.
• How will IT companies make up for lower prices? — Through net new AI projects, which the CEO expects to offset the deflation, and through better margins from higher productivity.
• Which skills should Indian IT professionals learn now? — AI engineering with LLMs and RAG, platform engineering and MLOps, AI security and compliance, system design, and AI-assisted development.
• Is this good or bad news for India? — Both: it pressures old billing models and routine jobs, but opens a path to higher-value work and more startups.

Tags

#TCS #IndianIT #ArtificialIntelligence #AIDeflation #ITJobs #CareerGrowth #FutureOfWork #AIEngineering #DevOps #Upskilling