All of India's micro-enterprise credit growth came from ticket size. None of it came from reach.

Between 2023 and mid-2025, credit outstanding to India's micro-enterprises grew 37.6%. Over the same 21 months, the number of micro-enterprise borrowers fell 6.3%.

Those two facts are usually reported separately. Put together, they decide who gets funded in this country for the next decade.

Credit decomposes exactly. Total credit = borrowers × average ticket. Borrowers moved by a factor of 0.937. Average ticket moved from ₹5.74 lakh to ₹8.42 lakh, a factor of 1.467. Multiply: 1.375, against a reported 1.376. The identity closes to within a tenth of a point.

All of the growth came from ticket size. None of it came from reach.

Inflation explains about nine of those forty-seven points. The other thirty-eight are a decision.

I've spent eighteen years building technology for NBFCs and the last several running one, so let me be fair about that decision. It wasn't greed. It was arithmetic. A ₹2 lakh loan to a tea shop in Villivakkam costs roughly what a ₹20 lakh loan costs to source, underwrite and collect, and earns a tenth as much. Every credit committee in India has run that sum. Most of them ran it quietly and moved up the curve.

What sits below the line is roughly ₹18 lakh crore of real, creditworthy demand that nobody can serve profitably by hand.

"By hand" is the operative phrase, and it is what I want to argue about for the next hundred days.

Because the constraint everyone names is capital, and it almost never is. I was talking to the COO of a ₹10,000 crore NBFC recently. He has treasury headroom. What he does not have is the ten-thousand-and-first employee. Five percent of his field force leaves every month, which compounds to 46% of the workforce gone in twelve months and an implied median tenure of about thirteen and a half.

He is not running a lending business. He is running a recruitment business that also lends.

The plan for a hundred days, roughly in order. The line lenders stop at, and the equation that puts it there. The ₹1 crore per employee ceiling. Why sourcing, not credit and not capital, is the binding constraint. What underwriting a salon actually requires. Why repayment is a design problem with a variance term inside it. What AI changed in the last two years and what it did not. And what we have built, including what broke.

Real numbers throughout, including the ones that don't flatter us.