The infinite Loop #193

The sovereign flywheel: AI brains, lending muscle

Mayank Jain

Head - Marketing and Content

·

Feb 20, 2026

Hello everyone,  

Welcome to the 193rd edition of The Pattern — a weekly newsletter on the latest in finance, technology, and the economy. 

UPI just delivered another record month. In January, Indians moved ₹28.33 lakh crore through 21.7 billion transactions — up from December's already massive numbers. Digital lenders now disburse 11 crore loans totalling ₹2.9 lakh crore, with one in three happening inside shopping apps or payment flows.  

Forget payments infrastructure. India built credit infrastructure at population scale. 

This week proves it. NPCI partners with NVIDIA for sovereign AI that makes payments more fraud-proof resistant. The RBI restores Default Loss Guarantees for NBFCs, unlocking billions in lending capital. 

First, the brains: NPCI × NVIDIA strengthen India's payment intelligence 
 
When millions of people rely on you every single minute, staying rock-solid is non-negotiable. Even marginal fraud or latency issues can materially impact user confidence. 

That’s why NPCI is teaming up with NVIDIA to weave homegrown AI right into the heart of payments. The goal? Slash risk as transaction volumes keep skyrocketing. 

Essentially, NPCI is leveraging the tech giant’s ‘Nemotron’ (a suite of open models) to create a payments-focused AI foundation that fully respects India's data sovereignty and regulatory needs. It's a smart move toward true sovereign AI, keeping sensitive financial data local and compliant. 

They've also rolled out the UPI Help Assistant, powered by FiMI (Financial Model for India). It's a custom fine-tuned small language model built just for payments, helping resolve user grievances with quicker, more consistent answers. This could be a meaningful upgrade for customer experience, turning frustrated queries into seamless resolutions and boosting adoption even further. 

Taken together, efforts like these could shift India’s role from payments pioneer to a benchmark for how AI-driven financial systems are designed and scaled. 

Capital efficiency returns 

The renewed clarity around DLG structures is equally significant. 

When earlier tightening increased Expected Credit Loss provisioning requirements, co-lending slowed and return ratios came under pressure. Some lenders absorbed provisioning impacts running into hundreds of crores, visibly affecting quarterly profitability and making boards more cautious about fintech-linked growth. 
With clearer risk-sharing norms now in place, including capped guarantees embedded within loan structures, predictability has returned. 

Capital that was defensively parked can now be redeployed into fresh originations. 

What changes structurally? 

  • Lower capitaldrag per loan originated 

  • Improved visibility on loss distribution between fintechs and NBFCs

  • Faster co-lending approvals as board-level uncertainty reduces

  • More stable quarterly earnings outlook 

For NBFCs, this improves RoA, or Return on Assets, which measures how efficiently profit is generated from the loan book. Even marginal improvements in RoA can materially expand lending capacity when balance sheets run into tens of thousands of crores. 

For fintechs, it restores confidence to scale sourcing engines without overwhelming partner capital buffers. 

In practical terms, this means: 

  • Retail and MSME pipelines can accelerate again

  • Embedded credit models regain momentum

  • Balance sheets can grow without disproportionate provisioning spikes 

This shift is happening in a market already disbursing roughly ₹2.9 lakh crore digitally each month across about 11 crore loans.  

At that scale, even marginal gains in capital efficiency can unlock significant additional credit supply over a year. 

This is how systems mature 

First came access, then scale, and now comes optimization. 

NPCI's AI capabilities make UPI more resilient to fraud, while the RBI’s DLG clarity unlocks billions in lending capital. This creates a flywheel: smarter payments lead to fewer defaults, which enable leaner capital structures, which support more digital loans and deeper adoption. 

At 11 crore loans and ₹2.9 lakh crore per month, we're past gradual gains. You've got a self-reinforcing credit engine powering India's financial leap. 

India is now focused on the long game — building systems robust enough to evolve with constant change, because a digital economy must be designed to endure. 
That’s the pattern.  

Reading list 

That’s all for this week. See you next time!   

Thank you for reading.If you liked this edition, please forward it to friends, colleagues, and your network. Do encourage them to subscribe as well. You can also follow FinBox on LinkedIn and myself on X to keep up with all the updates.        
    
Cheers, 
Mayank 

Press release

FinBox raises $40M Series B to power faster, fairer, and more inclusive credit

Solutions

Products

Resources

FinBox raises $40M Series B

FinBox raises $40M Series B

FinBox raises $40M Series B