Unusual options activity sounds intriguing at first, but it can be misleading.

Institutions, hedge funds, and big players around the world use options for a variety of reasons. They use them to hedge, to arbitrage, and (of course) to speculate.

Hedging is a major use, and it's the one that trips people up most.

You see, institutions often use complex options spreads to hedge a large position in something.

It’s not infrequent that I'll wake up, wait for the options data to publish at 6:30 a.m., and see that the biggest changes from the day before were cheap “put butterflies,” essentially insurance in case of a market correction.

Nothing to write home about. But if you saw it on a screener with a dollar figure attached, you might convince yourself a hedge fund was calling for a crash.

Then sometimes you see “short calls” on a stock. As a refresher, a call is a bet the stock goes higher. A short call is selling that bet to somebody else.

This is generally read as bearish, or more accurately, non-bullish. The assumption is that an institution is collecting income because it doesn't think the stock is going anywhere. Maybe.

But a fund that has quietly accumulated millions of shares in “dark pools” (a topic for another day) will often sell calls against that stock. That's a covered call. It's somebody who owns a pile of shares getting paid to wait.

Or we'll sometimes see a long call in a significant chunk on a stock and assume it's a bet the stock moves higher. (Were it so simple!)

Say they bought that call and sold another call at a higher strike. That's a bull call spread, and yes, they genuinely think the stock is heading toward that higher strike. Fine.

But if they sold a call at a strike below the one they bought, you're looking at a bear call spread — and the whole position is now net short.

That long call up top is the wing. It's there to cap the damage if they're wrong. The position is saying “this is the top,” not “this is going higher.”

And there's one more wrinkle that catches people…

Hedge funds will break a large order into much smaller chunks across their brokers to get a better average price.

So the thing you most want to find — the single enormous conviction bet — often won’t look how you expect it to. It appears as a hundred tiny blips that no unusual-activity scanner will ever flag.

My point here is that unusual options activity is rarely the insider information we all dream about. Just like anything else in life, there's no silver bullet for this stuff.

You have to dig. Read the news. Find the catalysts.

You also have to understand volatility, term structures, and “kurtosis,” which is just a fancy way of saying how fat the tails are, or how much the market is paying up for the extreme outcome versus the boring one.

And you have to wait for confirmation. This is the part almost nobody does.

Volume tells you a contract traded. It does not tell you anybody owns anything — that volume could be an opening trade, a closing trade, one leg of a spread, a roll, or a hedge that unwinds by lunch.

The only thing that tells you a position actually exists is the next morning's open interest file. If open interest went up, somebody went home holding it. If it didn't, you were watching noise.

So that's the standard for all of the data below. Nothing makes this list on volume alone.

Now, are there rare circumstances where you'll see something that looks an awful lot like insider activity?

Sure. But they're rare, and they tend to happen in a way that almost nobody can act on in time.

That said, looking at whale positions and reading the tea leaves — if you know how to read them — can prove lucrative.

So let's try to read some of them today.

MCHP

Microchip Technology · $74.21 · Semiconductors · Earnings Nov. 5

THE TRADE: Monday, 11:30:02. One order, two legs. Sold 118,920 December $65 calls at $12.90. Bought 150,000 Oct. 16 $72.50 calls at $5.82. Just 17 minutes later, another 150,000. Roughly… $240 million.

THE TELL: That October strike held 96 contracts the night before. By Tuesday morning, it held 150,522. December open interest dropped 60,431, which confirms the first leg was a close, not a new short.

THE READ: So somebody was already long Microchip through December calls. They were up on it, and they decided to roll. They took the December position off and put the money into a nearer, higher strike. Less capital tied up, more leverage per dollar. Perfectly ordinary stuff. Except for one thing…

Microchip reports earnings on Nov. 5. The October contract expires on Oct. 16, three weeks before the print. Think about what that means.

This trader wants to own the move but has gone out of their way to not be there when management speaks. That's somebody who believes the stock drifts higher through the fall on its own and who has zero interest in finding out what the CEO says about it.

Analog semiconductors have spent two years being told the cycle bottomed. Microchip's last quarter beat and guided up. This is the largest single-name commitment anywhere in the American options market across the two days, and it's a bet on the drift, not the event.

And Tuesday? The name did $3.6 million…

Which is exactly what a real institutional order looks like when it's finished. They do it, and then they're gone.

DKS

Dick's Sporting Goods · $178.85 – $125.44 · Retail · Reported Aug. 25

THE TRADE: Monday 10:36, somebody bought 2,000 September $190 calls at the offer plus 3,000 September $200s. Tuesday 9:35:19, somebody bought 6,209 September $160 puts at the offer for $22.28.

THE TELL: Those Tuesday puts were already nineteen dollars in the money when they were bought.

THE READ: Let's take these in order, because it's a good lesson in what pre-earnings flow is actually worth.

Monday's buyer paid up for out-of-the-money September calls. Lifted the offer too, meaning they wanted them badly enough to pay the asking price rather than wait. On a screener Monday night, that's a bullish flag on DKS.

Tuesday before the bell, Dick's reported. The core business was honestly fine — comps up 4.9%, revenue $5.59 billion. The problem was Foot Locker. It’s always Foot Locker.

Foot Locker comps fell 3.6%, the unit lost about $32 million in the quarter, and management took its full-year outlook for that business from a $110–$150 million profit to a $40–$80 million loss. Company-wide EPS guidance went from $13.50–$14.50 down to $11–$12.

The stock opened at $141.65 and finished the day down 30.68%. Worst session in the company's history. Monday's bullish flag was worthless before the coffee got cold.

Now here's the part I actually care about.

Five minutes into that bloodbath, somebody bought six thousand puts that were already deep in the money. When a put is that far in the money, it's almost all intrinsic value. You're not paying for time or volatility; you're basically paying for short stock exposure with a defined risk. Nobody buys that structure as a hedge or as a lottery ticket.

They bought it because they looked at a 23% gap down and concluded it wasn't finished.

It closed at $125.44, down another 11.4% from where they bought.

And then look what happens overnight. September $130 puts up 4,787 contracts. September $130 calls up 4,619. $135 calls up 4,251, with the $145s and $150s sold against them. That's a position built around the stock moving, not around which way.

Somebody was short, got paid, and by Wednesday morning had stopped having an opinion about direction entirely. That's how this is supposed to be done.

BE

Bloom Energy · $211.38 · Fuel Cells · Up ~150% YTD

THE TRADE: Monday 3:15 p.m., bought January 2028 $200 and $210 calls at $92.38 and $89.08. Tuesday 9:32, lifted 1,500 November $270 calls at $24.20. Tuesday 10:43, bought 300 June 2028 $300 puts and 600 June 2028 $100 puts.

THE TELL: $292.8 million across two days, more than Palantir. The November $270 strike went from 503 contracts to 1,978 overnight.

THE READ: Those January 2028 calls are what I'd point to first. They're struck way below the stock price, which means almost the entire premium is intrinsic value. You're paying roughly what the stock costs, minus the strike. There's barely any time premium in there at all.

That's what's called stock replacement. You get the same exposure as owning shares, you tie up less cash, and your downside is capped at what you paid. Institutions do this when they want to own something for years without posting the full capital.

Then look at what they wrapped around it: long calls way out at $270 in November for the upside, and June 2028 puts at both $300 and $100 underneath. Long the stock synthetically into 2028, with the floor defined and the tail covered.

On Aug. 12, Nebius picked Bloom's fuel cells to power an AI data center. Somebody is treating this as infrastructure now, not a trade.

NBIS

Nebius · $212.64 · AI Infrastructure

THE TRADE: Monday 10:41, bought 378 September 2027 $195 calls at $85.20. Monday 3:07, bought 1,500 December 2028 $75 puts at $21.50 and 300 December 2028 $320 calls at $96.40.

THE TELL: That September 2027 $195 strike held zero contracts before this trade. It closed the night at 650. The December 2028 $75 puts went from 7,031 to 9,541.

THE READ: And here's the other side of that same Aug. 12 handshake.

Same shape as Bloom, almost move for move: deep in-the-money long-dated calls for synthetic ownership, long-dated puts underneath for the floor, and a far out-of-the-money call for the dream scenario.

Now, a strike with zero open interest is one of the more useful things you can find. It means nobody, anywhere, held that contract before this order. There's no crowd to be part of and no prior position being rolled. Somebody sat down and decided that contract should exist.

Two counterparties to the same commercial deal, positioning identically, on the same two days. That's not a coincidence you can build a trade on by itself, really. But it tells you the deal is being taken seriously by people with real money and a multi-year clock.

SMR

NuScale Power · $9.79 · Small Modular Nuclear · Down 32% YTD

THE TRADE: Tuesday, between 1:22 and 1:33 p.m., somebody bought October $11 puts over and over: 1,595 contracts, then 1,497, then 1,321, then 1,213, 943, 855, 855. Every one at $1.97.

THE TELL: Open interest at that strike went from 12 to 20,136. The October $10 calls also went from 247 to 4,041.

THE READ: Remember what I said at the beginning about funds breaking orders into chunks? This is exactly that, and it's why I want you to look at time stamps and not just dollar amounts.

No single one of those prints is big enough to make anybody's unusual activity feed. The largest is under $320,000. But string them together across 11 minutes at an identical price, and you're watching one decision, executed carefully so as not to move the market against itself.

The backdrop: NuScale is down 32% on the year, reported second-quarter revenue of $75,000, announced a $750 million share sale, and has had analysts cutting price targets all month.

Somebody spent 11 minutes on Tuesday afternoon quietly buying downside on it.

XE

X-Energy · $18.74 · Small Modular Nuclear

THE TRADE: Tuesday 10:46, a 10,000-lot September $20 straddle crossed. Puts at $2.64, calls at $1.34.

THE TELL: Overnight: September $20 puts 2,261 → 12,264. September $20 calls 5,180 → 15,181. And the September $35 calls went from thirteen contracts to 23,587.

THE READ: A straddle is buying the put and the call at the same strike at the same time. It's a bet on movement. You make money if the stock goes somewhere, and you lose if it sits still.

So on X-Energy, somebody with ten thousand lots cares that it goes.

Now put the four power names side by side, because separately they're interesting and together they're an argument.

Money is going into fuel cells with multi-year duration and defined floors. Money is going into the downside — or at minimum the violence — of small modular nuclear.

Both of those are bets on AI power demand. One of them says the electrons show up in 2027. The other says they don't show up from a reactor.

MU

Micron · $911–933 · Up 206% YTD

THE TRADE: Tuesday, between 11:15 and 11:18 a.m., sixteen separate four-leg January 2028 call ladders. $155 million, 11,708 contracts, one expiry, three minutes.

THE TELL: $2.68 billion across the two days. Second only to the S&P complex itself.

THE READ: 16 identical structures in 180 seconds is more program than human. Somebody had a number to fill and an algorithm filling it.

Micron has sold out its 2026 production and is up 206% this year. It fell 7% on Monday.

And somebody used that dip to buy January of 2028.

I hate Micron here, for what it’s worth. You don’t need to take what these institutions are doing as gospel.

DRAM

Roundhill Memory ETF · $55.98 · Holds MU, SNDK, Samsung, SK Hynix

THE TRADE: Tuesday 11:20 a.m., crossed 6,500 March 2027 $60 puts at $13.03 and 6,500 March 2027 $60 calls at $10.17. About $15 million.

THE TELL: This fund reached $25 billion in four months, the fastest ETF launch in history. Overnight, the March $57 calls went from 15,507 to 34,554.

THE READ: This might be my favorite thing in the two days, and hardly anybody watches this ticker.

It's another straddle — puts and calls together, a bet on movement — except it isn't on one company. It's on an ETF that holds the entire memory industry in a single wrapper. Micron, Sandisk, Samsung, SK Hynix, all of it.

So somebody has looked at the loudest, most crowded, most argued-about trade in technology and said, “I don't know who wins this. I know it gets settled, and I know it gets settled hard, and I know it happens before spring of 2027.”

That's a different kind of conviction than picking a direction. Honestly, it might be a better one.

WPM

Wheaton Precious Metals · $159.63 · Gold & Silver Streaming

THE TRADE: Monday 10:51. 10,000 January 2027 $150 puts, lifted at the ask at $14.12.

THE TELL: Open interest at that strike the night before: forty-six contracts. No spread or financing leg. There wasn’t anything sold against it.

THE READ: Go back to the beginning for a second. Almost everything we talked about involves a second leg — the covered call, the bull spread, the bear spread, the butterfly. The reason those structures fool people is that the second leg changes the entire meaning of the first.

This trade has no second leg.

Somebody bought 10,000 puts, paid the offer to get them, and sold nothing against them. Across two full sessions and 19.5 million individual prints, this is the only outright, unhedged, paid-up bearish position of real size in the entire file.

Which makes the context wonderful. Gold is sitting near $4,300. Wheaton just posted record revenue. The stock rallied on Tuesday while these puts were being bought.

I don't know what they know. I just know this is the one trade in two days that can't be explained away as plumbing.

EWZ

Brazil ETF · $35.44 · Election Oct. 2026

THE TRADE: Tuesday 1:31 p.m., bought 500,000 November $43 calls at $0.63 and sold 522,500 November $45s at $0.38. 28 minutes later, another 136,500 at the $41 strike.

THE TELL: 1,458,188 contracts. The largest contract count of either session, by a distance.

THE READ: Now this is a bull call spread. (Buy a call, sell a further-out call above it, which means they genuinely think it goes up, but only so far.)

And look at where the ceiling is. The stock is at $35.44, they own the $43s, they've sold the $45s. That's a position that pays between roughly 16%–27% upside and then stops. Nothing above that.

Why would you deliberately cap yourself? Because you're not trading the asset. You're trading an event, and you've decided the event has a range of outcomes rather than an infinite one.

Brazil votes in October. Lula holds a narrow lead over Flávio Bolsonaro with about six weeks to go. Nov. 20 expiration lands after the first round and after any runoff.

That's somebody who thinks one specific election outcome is mispriced.

WBD

Warner Bros. Discovery · $28.89 · Merger Arb

THE TRADE: Accumulating January 2027 $15 calls, deep in the money, a couple thousand contracts a day: 5,464 - 6,784 - 9,067. Then Tuesday, October $28 puts went from 8,419 to 32,456.

THE TELL: 24,000 contracts of at-the-money downside, bought on top of the synthetic long.

THE READ: Deep in-the-money calls on a $29 stock, struck at $15. Same idea as Bloom: that's synthetic stock, the cheap way to own the deal consideration if this merger closes.

Paramount Skydance's $111 billion acquisition cleared the Justice Department back in June, and then got pushed out (potentially all the way to June 2027) under legal challenge.

So somebody is building ownership of the deal a couple thousand contracts at a time. And then on Tuesday, they went and bought a wall of puts right at the money.

Translated: I'll own this if it closes. I will not own it through the trial.

It's the least dramatic-looking thing in the two days, and I'd argue the most sophisticated.

AMGN

Amgen · $443.92 · Pharma · Earnings Nov. 3

THE TRADE: Tuesday 10:22, a cross of 4,125 October $460 puts at $29.75. September $440 puts also went 648 – 2,699.

THE TELL: Open interest at that October $460 strike: three contracts.

THE READ: Three contracts. In one of the largest pharmaceutical companies on earth, with an enormously liquid options chain, that strike had three contracts on it.

$12 million showed up there in a single cross, seven weeks ahead of the Nov. 3 print.

Nobody hedges Amgen. Somebody just did.

QBTS

D-Wave Quantum · $19.18 · Quantum Computing

THE TRADE: January 2028 calls, accumulated across both sessions. The $3 strike went 214 - 1,857 - 3,676. The $8 added 1,888. The $10 added 1,278. Bought on the bid around $16.70.

THE TELL: $43.6 million. The most persistent accumulation of any small name in the file – and the only one on this list that built on both days.

THE READ: A $3 strike call on a $19 stock is about as close to owning the shares as an option gets. You're paying roughly sixteen dollars and change for something worth sixteen dollars and change. Almost no time premium.

Again, stock replacement. Somebody wants to own D-Wave into 2028 and would rather do it this way.

The company just reported bookings up 1,120% against a revenue miss and a wider loss. Which is either the setup of the decade or precisely what these things sound like right before they don't work.

I genuinely don't know. But whoever is buying has picked a side and given themselves two and a half years to be right.

SBET

SharpLink Gaming · $8.19 · Ethereum Treasury

THE TRADE: April 2027 $6 and $10 calls crossed twice on Tuesday.

THE TELL: Open interest went from zero to 15,125 at the $6 strike, and 39 – 15,042 at the $10.

THE READ: 15,000 contracts of a call spread on an $8 stock, built from nothing.

SharpLink holds roughly $3.8 billion of Ethereum, trades at $8, ETH is down 68% from its peak, and the company has been buying back its own shares. It's a levered proxy. If you want ETH exposure with a corporate wrapper around it, this is one of the vehicles.

Somebody wants that exposure through April 2027 and has capped it at $10.

Some Others…

CAPR · Capricor · $7.80: 7,500 December $17 calls crossed at $2.40 into a strike with zero open interest, plus 2,075 more at the $20s. A bet the stock more than doubles by December.

IRD · $4.32: 10,000 September $2.50 calls lifted at the ask. 621 - 10,828. The $7.50s went from seven contracts to 10,072.

ATI · $211.17: 3,968 October $185 puts bought on the bid Tuesday at 10:33. Prior open interest: three contracts. That's twice now this week, in a big industrial name.

SCHW · Schwab · $112.64: 8,000 December $115 calls crossed at $5.85. That strike went 2,365 - 10,390.

RKLB · Rocket Lab · $67.31: 7,000 January 2027 $45 puts crossed Tuesday afternoon, 1,331 - 8,337, with December 2028 $50 calls and $55 puts building underneath.

GLXY · Galaxy Digital · $24.43: One desk at 1:29 p.m. bought 4,695 October $30 puts, 4,695 September $30 puts, and 4,695 October $20 puts. All at the ask, nothing sold against any of it. About $6.9 million of pure downside.

CVNA · Carvana · $74.57: A single clip at 11:22 buying 3,350 January 2027 $74 calls, 2,000 $84 calls, and 4,390 $54 puts. Long the move, no opinion on direction, into next January.

TSEM · Tower Semiconductor · $211: A September 200/170 put spread crossed both days: 614 – 4,498 – 5,499. Expires before earnings.

CCJ · Cameco · $101.29: Opened Monday with a January 2027 110/115 put spread, +2,872 and +2,858 confirmed, then flipped to calls on Tuesday as the stock ran to $107.68. Somebody changed their mind in twenty-four hours.

FIX · Comfort Systems · $1,587: A five-leg September structure at 12:39 Tuesday: long the 1,700 puts, short the 1,800s and the 1,440s, short the 1,670 calls. A broken-wing butterfly with a covered call paying for it. Every leg confirmed overnight.

The Big Tape

The genuine hedging, when you find it, is at the index level. S&P 500 (SPX) combos struck 7,000/8,000 and 7,000/9,000 ran $350 million, $330 million, $212 million and $159 million on Monday alone.

So what's the picture? It’s not really fear, and it’s not really euphoria.

It's a market where the biggest players are quietly extending duration on the handful of things they actually believe in — memory, AI power, quantum, a Brazilian election — building those positions with defined floors and defined ceilings, and renting their downside from the index instead of selling a single share of anything.

Meanwhile, 53% of all new premium sits inside two months. The crowd is still trading the week.

And there's one soft spot anywhere in the entire curve, on both days: eight to 13 months out. Put-to-call ratios of 1.10 and 1.05 — the only two readings above parity on the board.

That window lands next summer.

I don't have a good story for why yet. But it's the thing I'll be watching.