This week, Oracle laid off approximately 30,000 employees globally, with nearly 12,000 of those layoffs occurring in India. This marks the largest tech layoff in the country this year.
The move comes after sustained pressure from Wall Street on the company’s stock, driven by its aggressive bets on AI infrastructure. Between June 2025 and February 2026 alone, Oracle spent $39.2 billion in capital expenditure, against $17.4 billion in operating cash flow, pushing it into negative cash flow — something markets don’t easily forgive.
This unfolding story, while firm-specific, deserves analysis through two distinct policy lenses. The first concerns what it reveals about the intensity of the global AI race and its implications for India’s industrial policy. The second examines what the accelerated substitution of labour with capital as a factor of production in the digital age means for social policy.
Consider first the AI race.
Oracle is investing in AI infrastructure with clear expectations of future demand. It is set to provide a significant portion of the cloud computing backbone to OpenAI, with the two firms recently entering into a landmark agreement worth approximately $300 billion for cloud and compute capacity over a period of roughly five years.
By contrast, in India, domestic players have invested in AI infrastructure without securing comparable long-term commitments. Policymakers have focused largely on the supply side, including subsidies for GPUs and welcome steps such as providing tax certainty for data centres servicing foreign markets. They have paid far less attention to cultivating demand, similar to the push for 5G telecom infrastructure.
Where, then, does demand for AI infrastructure originate? Application-layer companies drive it — firms like OpenAI in the US. Yet, in India, policy interventions aimed at the application layer often constrain rather than expand economic activity.
Just last week, we saw new rules put out for consultation that would expand government control of online content. Last year, policymakers abruptly shut down real-money gaming — an industry whose surpluses had fuelled a broader ecosystem of tech-intensive game development. India is also considering tighter regulatory oversight for online streaming businesses, heavy end-users of AI.
If India chooses not to actively cultivate domestic demand, its industrial policy must at least position the country as a hub for digital services exports. Policy efforts here are confined to procedural frameworks like tax administration in the case of the zero rating of software exports.
This raises more fundamental questions. What would meaningful reform look like for firms exporting digital services? Policymakers could consider hybrid economic zones that can service both foreign and domestic markets, facilitate inward remittances and investments, among other targeted measures. But such moves will require a shift in policy attention and state capacity from areas such as online content regulation.
The second lens is social policy.
Here, two practical issues stand out, setting aside the age-old debate on universal basic income in the age of AI. We don’t have the fiscal room to provide this as long as 800 million people are on the equivalent of food stamps, public debt markets remain shallow, and states battle with ballooning deficits.
The first is, how do we incentivise industry to invest in workforce upskilling? Firms treat existing apprenticeship and internship frameworks as compliance obligations, rather than as instruments for building a globally competitive workforce. Perhaps rightly so, given that the state has no real skin in the game. The state-subsidy incentive per apprentice is capped at less than $50 per month.
Policymakers often worry that deeper monetary incentives or regulatory concessions for firms that are investing in upskilling will lead to misuse. But good policy design requires distinguishing between opportunistic and genuine behaviour. To argue that this cannot be done is to concede the limits of governance.
The second issue relates to education. India’s National Education Policy contains many of the elements needed to foster interdisciplinary learning, a critical component of a resilient workforce. Yet, implementation has prioritised incremental changes to curricula over pedagogical reform. Just yesterday, the government announced a new “Curriculum on Computational Thinking and AI” for children in grades III – VIII.
These are useful steps, but they do not address the core question of whether teachers encourage out-of-the-box thinking, provide immersive learning, or are equipped to evaluate students using AI.
Simultaneously, the demand for specialised education continues to outpace supply, as is evident from student migration. In theory, private capital should help bridge this gap. However, political and judicial institutions prefer to treat education as a not-for-profit activity. This virtue signalling has limited the flow of institutional investment into the sector and encouraged the proliferation of elaborate workaround models to extract profit. India’s educational system, therefore, continues to produce large volumes of talent that are increasingly substitutable.
AI is not a passing phenomenon. It will evolve through cycles of expansion and contraction. But waiting for the “bubble” to burst is not a sound strategy.
Many business leaders who actively use AI express hesitation about expanding hiring in private conversations, particularly at the entry-level, where the demographic dividend is highest. This, in itself, is a litmus test for assessing the structural impact of AI. We must respond with the exceptional focus and clarity such moments require.
Vivan Sharan is a public policy expert and co-founder of Koan Advisory, a New Delhi-based technology policy consulting firm
