Who is buying Indian stocks, and who is selling?

Who bought and who sold, since December 2023

NSE and BSE have published net buying by every type of investor since 4 Dec 2023. From then to 30 Sept 2026, mutual funds bought a net ₹14.5 lakh crore and insurance companies ₹1.6 lakh crore, while foreign investors sold ₹8.0 lakh crore and others ₹7.8 lakh crore. Individuals trading on their own account were net sellers of ₹26,673 crore, with FY26 a net sale. 33 sessions with missing or broken exchange data are left out.

Domestic institutions against foreign investors

The two big blocs of institutional money, week by week. Domestic institutions are mutual funds, insurers, banks and pension funds, reported by the exchanges. Foreign portfolio investor figures come from NSDL, which dates trades a day after the exchanges, so compare the two over weeks, not days.

Every investor category

Since December 2023 NSE and BSE have published net buying by type of investor. Each gets its own chart because their scales differ widely: read the direction and persistence of each, and compare sizes using the axis labels.

Blue columns above the line are months of net buying; red below, net selling. Source: IndiaDataHub (NSE, BSE, NSDL, AMFI, PFRDA, CDSL, NSDL depositories)

Mutual funds and SIPs

SIPs are the monthly instalments that give domestic funds a steady supply of cash to invest. The inflow is designed to be stable, so the more telling numbers are net flows into equity funds and the stoppage ratio, which shows whether investors are quitting.

Pensions and accounts

Two slower signals of how much household money reaches the market: pension savings invested in shares through the National Pension System, and the number of demat accounts that let people hold shares directly.

Do flows lead returns, or follow them?

A common story is that foreign money moves the market. Week by week, the record points the other way: past returns predict flows, and flows do not predict later returns in any sample we tested. Foreign investors' net buying in a week correlates 0.47 with the previous week's return, so they tend to follow the market; domestic institutions' net buying correlates −0.36, so they tend to lean against it.

How to read it. Look down the correlation columns: the large numbers sit at "before the flow" and "same week", where the return comes first or alongside. Against any of the next eight weeks' returns, neither flow's correlation exceeds 0.13 in size. A Granger test measures predictability, not cause. Weekly sums also mix trade dates and reporting dates (NSDL dates foreign trades a day after the exchanges), and the foreign series covers only the years since 2021.

Source: NSDL (foreign portfolio investors), NSE and BSE (domestic institutions), NSE (Nifty 500 total return index). Pre-registered before it was computed: specification and results log (C3). JSON. How the flow series are sourced: method.

Bulk and block deals

Large trades in a single stock are disclosed by name. A bulk deal is one client's trades in a session adding up to more than 0.5% of a company's shares. A block deal is a single trade of at least ₹10 crore, made in a separate window. On 1 Oct 2026, the latest session recorded, NSE disclosed 157 bulk deals in 32 stocks and 6 block deals in 3. Our record begins on 1 Oct 2026: NSE makes only the latest session openly available, so there is no earlier history here, and it grows one session at a time.

How to read it. A buy and a sell in the same stock by different clients are often the two sides of one trade, and a block trade can also show up as a bulk deal. Adding up the values therefore double-counts, so the buy side alone is the cleaner measure of size. A large deal says who traded, not why.

A view of the most active stocks across recent sessions appears here once the record holds more than one session.

Source: NSE bulk and block deals (nsearchives bulk.csv, block.csv), captured daily; via Data bank nse_bulk_block_deals. Latest session (JSON), last 45 days (JSON), daily counts (JSON). Method.