Nabfid Sanctions ₹30,000 Cr+ Loans to Four Data Centers Amid AI Buildout

⚡ Key Financial Takeaways

Nabfid has sanctioned loans of over ₹30 billion (approx. $313 million) each to at least four data center projects.

Managing Director Rajkiran Rai G stated that these projects offer good cash flows despite long construction periods, with repayment terms spanning up to 15 years.

Nabfid estimates India’s data center sector will require approximately ₹1 trillion in funding by March 2031.

The lender is preparing its maiden dollar bond and aims to raise ₹1 trillion in the current fiscal year, with 40% sourced from overseas by end-December.

💡 Why It Matters

Nabfid’s aggressive lending signals institutional confidence in the long-term viability of India’s AI infrastructure. By structuring loans with extended moratoriums, the lender is aligning with the capital-intensive nature of data centers. The push for offshore funding, including the maiden dollar bond, indicates a strategic effort to reduce reliance on domestic rupee funding and access global capital markets, which is crucial for meeting the projected ₹1 trillion funding requirement by 2031.

Strategic Lending to AI Infrastructure

India’s National Bank for Financing Infrastructure and Development (Nabfid) has significantly increased its exposure to the data center sector, sanctioning loans upwards of 30 billion rupees ($313 million) to at least four facilities. This move underscores the state-run lender’s role in supporting India’s expanding artificial intelligence (AI) infrastructure.

According to Managing Director Rajkiran Rai G, the sanctioned loans come with a moratorium period of up to five years, followed by a repayment tenure of 10 years. Rai noted that despite the lengthy construction phases typical of such projects, they generate robust cash flows. He highlighted strong traction from hyperscalers and major cloud service providers seeking to establish compute capacity in India.

Funding Gap and Market Demand

The surge in loan applications reflects a broader trend of corporate spending on large-scale digital infrastructure. Global investment firms such as Blackstone, EQT AB, and Alphabet, along with domestic conglomerates like the Adani Group, have committed billions of dollars to AI compute projects in the country.

Nabfid’s internal estimates suggest that the Indian data center sector will require approximately 1 trillion rupees in funding by March 2031. This projection highlights the substantial capital gap that private and public financiers must address to support the nation’s digital transformation goals.

Diversifying Funding Sources

To meet this growing demand, Nabfid is actively diversifying its borrowing channels. The lender has prepared its maiden dollar bond issuance, a strategic move following recent interest rate hikes by the US Federal Reserve.

Rai stated that Nabfid aims to raise 1 trillion rupees in the current fiscal year (through March). Of this amount, approximately 40% is expected to be sourced from overseas markets by the end of December. Additionally, the bank has raised close to $1 billion in offshore loans under the Reserve Bank of India’s concessional window, which remains available until December 31. Foreign banks have been mandated to mobilize a 10-year dollar bond of a similar magnitude through this window by the end of the month.

🏛️ Background & Context

India is positioning itself as a major hub for global AI compute infrastructure. The involvement of both global hyperscalers and domestic conglomerates suggests a competitive landscape for data center development. Nabfid, as a key infrastructure financier, is playing a pivotal role in bridging the funding gap for these large-scale projects.

👁️ What To Watch Next

Readers should monitor the successful issuance of Nabfid’s maiden dollar bond and the extent to which the bank achieves its target of sourcing 40% of its annual funding from overseas by December. Additionally, the pace of loan disbursements to the sanctioned data centers will be a key indicator of the sector’s growth trajectory.