The Looming Crash: Why India Should Pay Attention

I’ve been closely watching the rise of LLMs like ChatGPT, coding assistants like Cursor and Claude, and the early emergence of AI agents. The disruption they’re creating is already visible — even in small companies like Vajro, that I co-founded. And the ripple effects across India’s IT ecosystem could be catastrophic if we’re not prepared.

This isn’t alarmism. It’s a warning. And it’s time we paid attention.

India’s IT Backbone Is More Fragile Than We Think

The Indian IT industry employs about 5.4 million people — that’s just 0.4% of our population. But it contributes nearly 7.5% of India’s GDP. That’s a massive multiplier effect.

The two-decade boom we’ve enjoyed — in real estate, automobiles, travel, and retail — was largely fueled by this one industry. If that engine slows down or breaks, the consequences will echo across the economy.

And here’s the problem: that engine is under threat.

Consulting Is the First Domino

Let’s start with consulting firms. Most of them still run on a pyramid model — billing clients based on people and hours. But now, AI agents can do much of what junior consultants used to. LLMs can write, summarize, strategize, and analyze faster than ever before. And they only get better with time.

The traditional consulting model is crumbling — not in the future, but right now.

Even legacy firms are not safe. Market cap drops for consulting majors (as of June 2025):

• Accenture: –18%

• Cognizant: –12%

• Infosys: –14%

• Wipro: –20%

(Source: Bloomberg, Financial Times — June 2025)

The old “people-rental” model is failing.

The Coding Layer Is Next

Coding assistants like GitHub Copilot, Cursor, and Claude are already accelerating developer workflows. At Vajro, we’ve used these tools to build features in hours — features that used to take days or weeks.

The impact is clear: high-performing developers become 10x more efficient, and average developers? They risk becoming redundant.

And remember — India never really built an IP-led economy. We built a “labor-for-hire” economy.

We’ve Been Here Before

Take the textile industry. India had the workforce, the history, the infrastructure. But we didn’t control the designs or the machines.

Eventually, Bangladesh and Vietnam overtook us — because labor was cheaper there. And we had no IP to defend our advantage.

It’s happening again. Just that clothes have been replaced by code.

What Services Companies Must Do Next

If you’re running a services company in India, here’s the harsh truth:

You need to reinvent. Now.

The current model — billing by the hour — is obsolete. Clients no longer care how many engineers you deploy. They care about speed, outcomes, and cost.

To survive, you must:

• Shift from people-led to outcome-led models

• Leverage AI as a core capability — not just a tool

• Automate low-value work and focus on strategic value delivery

Yes, this means letting go of old assumptions. Yes, it may involve reducing headcount.

But denial is not a strategy.

Crash or Opportunity?

There’s Still Time. But Not Much.

Here’s the silver lining: large enterprises are the slowest to adopt AI. Their bureaucracy and legacy systems give us a small window of time — maybe 12 to 24 months.

That’s our chance.

A chance to rethink, retool, and rebuild. A chance to pivot from labor-driven services to outcome-driven innovation.

As someone building with AI daily, I can tell you this:

The crash is real.

The opportunity is bigger.

But only if we act now.