India’s white-collar workforce is staring at a potential bloodbath as AI rapidly moves from experiment to execution, according to Saurabh Mukherjea, who warned that millions of jobs could disappear as companies race to automate work once done by humans.
Speaking on a podcast, the founder of Marcellus Investment Managers pointed to projections tied to a report from US software giant ServiceNow suggesting that as many as 18 million Indian jobs could be affected by AI automation by 2030, while only about 3 million new tech jobs may emerge in their place.
Mukherjea framed the numbers as a potential net loss of 15 million jobs. ServiceNow’s own language, however, describes many of those roles as being “redefined” or transformed by AI rather than entirely eliminated.
Still, Mukherjea argued the disruption could be severe and far more immediate than most workers expect.
“Abhi toh yeh trailer hai. Yeh shuruaat hai,” Mukherjea warned, saying the layoffs currently rippling through the technology sector are only an early glimpse of what happens once AI systems become capable of performing digital work at near-human levels.
According to Mukherjea, AI has already reached roughly 50% human reliability across a range of coding related tasks. Within the next two to three years, he believes those systems could hit 90% or even 100% reliability across large sections of tech work.
“That’s when the big threat to AI jobs will come,” he said.
The shift is being fueled by an unprecedented wave of AI spending from hyperscalers including Microsoft, Amazon, Alphabet, Apple and Anthropic.
Mukherjea claims those firms have collectively doubled AI capital expenditure in just the last 12 months, pouring roughly $800 billion into AI infrastructure, chips, data centers and large language model training.
“That’s 25% of Indian GDP,” he noted.
What has changed dramatically in recent months, he argued, is that hyperscalers are finally beginning to see explosive revenue growth from those investments.
“I think tipping point aa gaya,” Mukherjea said, pointing to exponential AI revenue growth over the last three to four months.
He cited Anthropic as a key example, claiming enterprise revenue tied to its Claude AI platform surged from roughly $1 billion in early 2025 to nearly $30 billion in annualized run rate revenue within just 15 months.
“That’s economies of scale. That’s takeoff. That’s tipping point,” Mukherjea said.
The result, he warned, is that hyperscalers now have even greater incentive to automate aggressively and cut labor costs faster.
An EY survey cited by Mukherjea found that 58% of Global Capability Centers in India are already investing in agentic AI while over 80% are investing in generative AI. The same survey showed that 24% of tasks inside GCCs are already fully automatable and another 42% can be significantly augmented by AI systems.
“If I was running a GCC, those 24% tasks which are fully automatable, I would immediately say there should be no human doing this work anymore,” he said.