Every few decades a new technology arrives, which promises to change everything. And often, it does. We saw computers, the internet, e-commerce, smartphones, and each transformed how we live and work. Artificial Intelligence (AI) will no doubt join that list.
But every revolution comes wrapped in its own hype. The technology is real, yet the timelines are always exaggerated. Right now, the hype around AI is racing years ahead of what it can realistically deliver.
That gap is visible in the extraordinary flow of money chasing AI dreams. Over the past year, ten unprofitable AI startups have seen their combined valuations surge to nearly a trillion dollars, according to Financial TimesUS venture capitalists have invested $161 billion in AI this year, almost two-thirds of their total spending, And in the public markets, Nvidia has become the first company in history to breach $5 trillion in market capitalization, its chips powering the global supercomputing build-up, That tells the story of an industry priced on what investors hope it might earn one day,
This surge is propelled by extraordinary bets. Big Tech is building at a massive scale: Data centers, chip foundries, and cloud infrastructure to handle the next wave of model training and inference. Citi expects AI-related capex at half a trillion dollars by 2026. Yet so far, only a fraction of it has brought in any measurable business impact. A recent MIT study found that barely 5% of enterprise AI pilots delivered rapid revenue growth; the rest have stalled or are anaemic.
While the West chases giant general-purpose models, India is at a different inflection point, with its focus on digital infrastructure, data registries, and applied use cases. We have the advantage of scale, data, and a massive base of skilled engineers. We’ve shown how scale and affordability can coexist: In telecom, fintech and vaccines. The same can happen with AI. If channeled well, AI can power a new industry — built in India, for the world — leveraging our IT services, startup ecosystem and Global Capability Centres.
To make that leap, we must invest far more in applied research, computing infrastructure, and home-grown innovation. Our frugal realities can actually work in our favor if we produce cost-efficient AI products and platforms designed for emerging markets. With the right public-private push, India could create not just an AI-ready workforce, but globally competitive products and brands.
The valuation mania has detached from economic logic. Startups with limited revenues and half-built products are being priced unrealistically high. Big Tech can afford those bets; their profitable cloud and advertising businesses act as a cushion. But smaller players and their investors cannot.
The public may feel insulated as these are seen as balance-sheet gambles by trillion-dollar firms. Yet, bubbles have a way of spreading beyond their origins. The dot-com boom at the turn of the millennium began with the same conviction that the Internet would change everything instantly.
When it didn’t, the crash was severe. Nearly $5 trillion in market value vanished, and the Nasdaq fell almost 80%. The internet survived, and even thrived. But only after a decade of recalibration.
Technological revolutions come in waves. The internet’s early years were defined by a gold-rush mentality; the true transformation came later. AI will follow the same arc. The first movers blaze the trail and burn through the cash. The second wave — the Oracles, Amazons, or perhaps tomorrow’s equivalents from India, China or Europe — will build sustainable value.
So far, most industries, including manufacturing, remain largely untouched. Areas such as education, law, design, marketing and consulting are seeing early AI-driven changes. However, the shifts are incremental. The real transformation will come when enterprises re-engineer products, processes, and profit engines around AI. That shift will take years. What we’re seeing now is the front end of the revolution: Pilot projects, prototypes, and promises, not the full system rewrite.
And amid the rush for digital dominance, it’s also worth remembering that the physical world is holding its own. Take bricks and mortar retail, which has defied most predictions. Online shopping boomed during the pandemic but has since leveled off at about 20% of total US retail sales, and physical stores have been opening faster than they’ve been closing. The digital and physical economies are finding an equilibrium rather than outright displacement, a reminder that new technologies often coexist with, rather than erase, the old.
In the late 1990s, Cisco, Lucent and Intel were the Nvidia and OpenAI of their time: Critical to the next generation of technology, yet caught in speculative exuberance. The problem then wasn’t that the internet failed, but that capital arrived too early and expectations rose too fast. We are seeing the same pattern now. Nvidia is a great company in a great position, but the valuation also shows how the narrative can start to run ahead of economic fundamentals.
Bubbles rarely burst because of one big event; they deflate when confidence does. Some of today’s growth rests on circular financing, or AI firms investing in one another while buying each other’s chips or services. For now, that keeps valuations afloat, but eventually investors will want to see real profits. Microsoft, Amazon, Google and Meta are together spending sums equivalent to the GDP of Portugal on AI infrastructure this year. If enterprise demand doesn’t rise as quickly as expected, we could be left with overcapacity, squeezed margins, and a painful reset.
None of this means AI is not real. It is — as revolutionary as electricity or the internet once were. But potential does not equal immediate payoff. It is crucial to build patiently, integrate deeply, and adapt continuously. Policymakers should encourage innovation while staying off speculative euphoria, and investors would do well to separate possibility from probability.
The promise of AI is immense, but the timing and discipline will decide who benefits. The technology will mature, the noise will fade and, as seen earlier, the true value will emerge.
Pramod Bhasin is chairman, Icrier, and founder, Genpact. The views expressed are personal
