Reliance's eight perfect minutes and a sixteen year old SEBI case

ELI5 / TLDR

In 2007, Reliance Industries quietly bet that its own subsidiary’s shares would fall, then sold a huge pile of those shares in the final eight minutes of trading, making about 513 crore on the bet. SEBI called it fraud and clawed back the money; sixteen years later the Supreme Court said it broke a rule but didn’t commit fraud, because the evidence didn’t add up. The second story: UPI killed the everyday ATM withdrawal, but India is holding more physical cash than ever, so cash isn’t dead — the business of running ATMs is, and the survivors are turning machines into mini bank branches.

The Full Story

The eight-minute trade

Picture the last day a particular futures contract is allowed to trade. A futures contract here is just a side bet on where a share price will go. These particular contracts had an unusual settlement rule: when the market closed, every bet would be cashed out at the average price of the final half hour, not the final price. So whatever happened in those last thirty minutes set everyone’s payout.

Reliance Petroleum was a subsidiary; Reliance Industries owned about three-quarters of it. The stock had run from 60 rupees at its 2006 listing to nearly 250 by October 2007, and analysts agreed it was overpriced. Reliance Industries decided to sell down 5% of its stake — roughly 22.5 crore shares. That much stock can’t be dumped at once without crushing the price, so the plan was to sell slice by slice over weeks.

The risk: if the price fell halfway through the selling, every remaining sale would fetch less. The fix was a hedge. A hedge is an insurance bet — you take a position that pays out if the thing you fear actually happens.

They could also take a side bet which would pay out if share prices fell. That would effectively let them have their cake and eat it too.

So Reliance shorted 9.92 crore shares’ worth of futures — “short” meaning a bet that the price goes down — less than half of what it was selling. The catch: the rules cap how big a single trader’s derivatives position can be. To get around the cap, Reliance routed the bet through 12 separate front entities, each secretly agreeing to trade only on Reliance’s orders and hand back every rupee of profit or loss. To the market these looked like 12 independent traders. They were one trader wearing 12 masks.

Then came the final eight minutes. Reliance sold 1.95 crore shares into the close, the average price for the settlement window crashed, and the short bets paid out 513 crore.

Rule-break or fraud?

This is the heart of the case, and the video frames it cleanly: breaking a rule is not the same as committing fraud.

Imagine a parent that tells their teenage child to make sure they’re home by 9:00 p.m. That’s a rule. If they break it, they might be punished… If the child wasn’t home because they were drinking alcohol, that would be a much worse problem. And the fact that they were home late would be secondary.

Being home late is the rule-break (the position cap). Drinking is the fraud (deliberately rigging the market). SEBI argued the late arrival proved the drinking. To win the fraud case, it had to prove three things: Reliance built the position in order to manipulate, it cornered enough of the market to do so, and the final eight minutes were the kill shot.

The Supreme Court took apart all three:

Was it a hedge or a weapon? The short position was less than half the shares being sold — too small to be a manipulation engine, exactly the size of a sensible hedge. Reliance was under no obligation to unwind the hedge every time it sold a share.
Did it corner the market? SEBI’s headline number was that 93.63% of the November futures shorts traced back to Reliance’s 12 entities. Damning — until you notice the share only peaked there late, when others had exited that particular contract, and across all futures months Reliance held about 40% of the short side, not the whole market.
The smoking gun. If Reliance wanted to crater the stock, it would have sold into weakness. Instead it had a price floor: it sold only above 210 rupees, held off when the price dropped to 190, and sold only when it recovered to 224. That’s someone taking advantage of a good price, not manufacturing a bad one. And in that final half hour, other traders were dumping shares too — more than half again of what Reliance sold — which SEBI never investigated.

The killer point: Reliance owned two-thirds of the company. A temporary 513-crore short profit would be dwarfed by the loss on its own enormous shareholding if it genuinely destroyed the stock. Nobody burns down their own house to win a bet on the house burning.

Most profitable trades can be made to look sinister in hindsight. Seby’s job wasn’t to poke at oddities, but to prove that there was manipulation.

The verdict split the difference: the penalty for the 12-entity rule-break stands, but the 500-crore-plus disgorgement is overturned and SEBI must return the 250 crore Reliance had deposited. Rule broken, fraud not proven.

India’s cash paradox and the dying ATM business

UPI made ATM withdrawals feel obsolete in cities. Yet India holds more banknotes than ever — 41 lakh crore, more than double the ~18 lakh crore just before the 2016 demonetisation. The currency demonetisation was meant to remove has more than come back.

The resolution: UPI replaced cash for spending, not for holding. Of that 41 lakh crore, about 35.2 lakh crore sits in 500-rupee notes — the note you stash, not the one you spend at a tea stall. The RBI’s own 2023 research found digital payments are substituting cash for transactions while the “store of value” motive stays intact. Rural India, real estate, and the unorganised economy still run on cash, often to stay outside the formal banking system.

So cash is alive, but the economics of dispensing it are broken. ATM operators earn an “interchange fee” — when you use another bank’s machine, your bank pays the machine’s owner a small fee per withdrawal. That fee was 15 rupees for nine years, nudged to 17 in 2021, then 19 in 2025. Meanwhile transactions per machine collapsed (an ATM that did 200 a day now does a fraction), and costs — rent, diesel for cash vans, and especially wages under new labour codes — kept climbing. The industry body called it a “force majeure” situation: the business can no longer run as designed.

The casualties are real. AGS Transact, the second-largest ATM services provider with ~72,000 machines at its 2021 peak, went from a 175-rupee IPO to insolvency, the stock at 5 rupees, with 13,171 crore in creditor claims. White-label operators — firms whose only income is interchange — are squeezed hardest; one pioneer exited the business entirely in 2026.

The survivors are changing the machine and the contract. The new “cash recycler” both dispenses and accepts deposits, does video KYC, issues debit cards, even cross-sells loans — a branch with no teller. ICICI Bank flipped from 80% dispensers to 80% recyclers in two years. And the pay model flipped from per-transaction (operator eats the volume risk) to a fixed monthly fee per machine (bank shares the risk). CMS Infosystems’ CEO declared the transaction-fee model “dead,” and CMS now holds 46% of the cash-management market.

The catch: recyclers cost ~6 lakh versus ~3.5 lakh for a dispenser, so they go where customers are dense — cities and tier-1-to-3 towns. Off-site ATMs (petrol pumps, small-town streets) fell by ~8,000 even as on-site machines grew ~5,000. Rural India gets pushed to weaker substitutes — micro-ATMs capped at ~10,000 rupees, business correspondents tied to one shopkeeper’s hours, RRB branches with queues. None is a true 24/7 ATM, and it’s unclear if they fill the gap.

Tidbits

JSW will build a 30 GWh lithium-ion cell plant via a JV, investing over $1.3 billion, eyeing Southeast Asia to exploit free-trade agreements, targeting its first new-energy vehicles by early 2027. Maruti Suzuki launches India’s first E100 (near-pure ethanol) car on June 5 — potentially 25–35% cheaper per km, but hampered by thin refuelling infrastructure and low energy density. Realty firm Anantraj signed a 20,000-crore MOU with Haryana for data-centre infrastructure expected to create ~6,000 jobs.

Key Takeaways

These futures settled at the average price of the final half hour, which is why the last eight minutes of selling mattered so much.
A “short” position is a bet that a price falls; a “hedge” is an insurance bet that pays out if the thing you fear happens.
Reliance hedged only 9.92 crore shares against a 22.5 crore-share sale — under half, the hallmark of a hedge rather than a manipulation weapon.
It used 12 front entities with secret agreements to evade the per-client position cap. This was the proven rule-break.
SEBI’s three-part fraud test — intent, market cornering, smoking gun — failed because the position was too small, the 93.63% share was a late peak in one contract (40% across all months), and Reliance sold into strength with a price floor, not into weakness.
The decisive logic: owning two-thirds of the company, Reliance would lose far more on its shareholding than it could gain on a short by genuinely crashing the stock.
Outcome: rule-break penalty upheld, ~500-crore disgorgement overturned, 250-crore deposit returned.
India holds 41 lakh crore in banknotes, double pre-demonetisation; ~35.2 lakh crore is in 500-rupee notes held as a store of value, not spent.
UPI replaced cash for transactions, not for holding — so cash demand persists but per-machine ATM usage collapsed.
The interchange fee (one bank pays another per cross-bank withdrawal) rose only from 15 to 19 rupees in 13 years while costs climbed — breaking ATM unit economics.
AGS Transact, the #2 operator, went from a 175-rupee IPO to insolvency at 5 rupees; the per-transaction pay model is being replaced by fixed monthly fees.
New “cash recyclers” double as teller-less branches (deposits, KYC, cards); they cluster in cities, leaving rural areas with weaker substitutes (micro-ATMs, business correspondents, RRBs).

Claude’s Take

The Reliance segment is the strong half — a genuinely clean explanation of a subtle legal distinction that most coverage botches: the gap between breaking a rule and committing fraud. The teenager-and-curfew analogy does real work, and the closing insight (you don’t torch a company you 70% own for a one-off short profit) is the kind of thing that reframes the whole case. The “spoton,” “alliance industries,” and “disorgged” transcription garbles aside, the reasoning is faithful to how the actual judgment ran.

The ATM half is solid business journalism — the cash-paradox framing (more notes than ever, but as a mattress, not a wallet) is the memorable hook, and the interchange-fee mechanics are explained well. It leans a little on quoted figures and one-off corporate anecdotes (CMS, AGS) without much skepticism about whether those firms’ framing is self-serving; the “transaction fee model is dead” line is, after all, coming from the company that just won 46% market share by saying so.

Scoring a 7: tight, well-structured, two non-trivial concepts made genuinely clear, no padding. Not an 8 because it’s a daily-news brief rather than something with lasting reference value, and the second story is more reportage than analysis.

Reliance's eight perfect minutes and a sixteen year old SEBI case | The Daily Brief #478