Speaking to CNBC-TV18, Velamakanni said the industry is already witnessing the early impact of artificial intelligence-driven disruption, which is compressing deal sizes and slowing growth. “The worst is yet to come in terms of contraction. Growth in Indian tech is currently around 2–3%, and the next 12–18 months could be similar or worse,” he said.
The pressure is being driven by rapid advances in AI-led software development, which are significantly reducing the cost and time required to execute projects. Velamakanni noted that work which previously commanded large budgets is now being repriced sharply lower. “A $100 million project from three years ago may now be bid at $70–80 million, or even $50 million by AI-native players. So compression is real,” he said.
This near-term slowdown comes even as the long-term demand outlook remains strong. According to Velamakanni, the world still lacks sufficient software, creating a large opportunity ahead despite current headwinds. “We need 1,000 times more software than we have today,” he said, adding that faster development cycles — potentially 10 times quicker — could ultimately translate into a significantly larger market opportunity.
However, he cautioned that the expansion phase driven by AI adoption has not yet fully materialized. Enterprise adoption remains complex, requiring accuracy, reliability and deep contextual integration. “It’s not just about plugging in a model — you need accuracy, reliability, enterprise context, fine-tuning. That’s hard,” he said.
Francisco D’Souza, Co-Founder & Managing Partner at Recognize, who also spoke to CNBC-TV18, said the industry is undergoing a structural shift of a scale not seen in decades. “It’s a large transition, perhaps an operating model transition that we haven’t seen in some 35 years in the industry,” he said.
D’Souza highlighted what he described as a “paradox of value”, where the value of software code is rising sharply even as the cost of producing it declines due to AI. At the same time, a “paradox of complexity” is emerging, as abundant code does not necessarily translate into reliable or secure software systems.
He added that both compression in existing work and expansion from new AI-led demand are playing out simultaneously. “Existing work is experiencing deflation, but at the same time we’re seeing enterprise AI adoption at scale,” he said, describing the transition as a J-curve. He expects the industry to reach an inflection point within about a year.
Despite the near-term pressures, Velamakanni said the disruption opens up large opportunities across AI-led business transformation, building foundational data and knowledge layers, and reshaping the workforce. He also emphasized that AI models will act as enablers rather than replacements for services firms. “Models will become inputs, not competitors,” he said.
Also Read | AI ‘structural shock’ biggest since Y2K, Indian IT risks ‘sleepwalking into crisis’: Phil Fersht
Looking ahead, he expects only a section of the industry to successfully navigate the transition. “At least a third will be exceptionally successful,” he said, adding that companies will need to significantly increase investments in research and development — even at the cost of margins — to stay competitive.
D’Souza, for his part, said firms that adapt their business models toward outcome- and output-based pricing, invest in proprietary intellectual property, and shift their talent structures will be best placed to benefit from the next phase of growth.
