Sikka, an early supporter of OpenAI during its non-profit phase, cautioned that India must not remain a passive consumer of AI tools developed elsewhere. “We are too significant. The stakes are too high for us to be excluded from any of these layers,” he remarked, noting that AI is already transforming software development, business operations, and productivity metrics.
He believes India should not merely harness AI but also innovate, scale, and export it. This involves developing core models, creating tools, and training millions to utilize AI effectively. “We have to engage, and we must excel,” he asserted.
He illustrated this with an example where a project that previously required 15 people and nine months to complete was accomplished by a single individual in just two weeks with AI assistance. This, he pointed out, highlights how rapidly the landscape is evolving. “It’s not a question of when it arrives. It’s already here.”
Thus, the key challenge lies in how swiftly we can shift from being consumers disrupted by AI to providers of it.
Sikka identified India’s readiness challenge as twofold. “It consists of two aspects. One is the skill issue… The other part relates to its disruptive impact,” he explained.
Skills represent a significant hurdle. Today, only a minuscule fraction of the global population possesses the ability to build advanced AI systems. He noted that the number of individuals capable of creating foundational models is “less than 25,000 worldwide.” However, these are not unattainable skills; they can be taught, learned, and scaled. India must act swiftly to facilitate this development.
The second challenge revolves around the potential disruption AI may cause to existing jobs and business models, particularly in software services. “Large language models are exceptionally adept at writing software,” Sikka warned, indicating that it will have direct implications for professionals who develop and maintain software. India’s expansive services sector will need to “adapt rapidly” to this transformation.
Sikka has recognized these trends for years. Back in 2015, he was one of the initial supporters of OpenAI when it was still a nonprofit organization. “We contributed to OpenAI… and we were among the first supporters,” he recalled, mentioning his advisory role. At that time, it was already evident where AI was heading. “The impact of AI on our work processes was becoming increasingly clear,” he stated.
While enthusiasm for AI is considerable today, Sikka expresses skepticism regarding current valuations. He remarked that the financial figures don’t seem to align. Referring to Nvidia, he noted that the company generates around $250 billion annually, and when considering CPUs, memory, and networking, total AI infrastructure spending reaches about $750 billion each year. “This is tangible,” he asserted.
However, he questioned the sources of revenue to justify such expenditures. “That $800 billion indicates there should be approximately $200 billion in revenue to support it all. Where is that $200 billion?” he challenged, asserting, “The existing economic realities don’t correspond to the spending levels.”
A recent Bloomberg analysis by Jonathan Levin indicates that excluding the hottest tech names like Nvidia, Palantir, and Tesla would reduce the S&P 500 index’s valuation from 22.7 to 21.9, making it appear significantly less expensive.
For reference, the S&P 500’s average valuation over the last 20 years has hovered around 16%. This has led some, including AI researcher Gary Marcus, to suggest that we are amidst an AI bubble, although this viewpoint is not universally accepted.
Despite potential risks, Sikka remains hopeful. He believes India still has time—but not much. Using a cricket metaphor, he stated, “If it were a T20 match, we are in the first over.” This implies that while there is still opportunity, action must commence immediately.
For the full interview, watch the accompanying video
