Reading conviction: what $48M in NVDA calls actually tells you
A walkthrough of one notable flow alert — how to read the data on screen, what to do with it, and the four questions to ask before acting.
You open the Smart Money digest. The top entry reads:
NVDA · $48.2M · 950 calls May 17 · 2.0× OI · bullish bias
Five seconds of information. The next ninety seconds are the ones that matter. This post is what to do with them.
What the headline says
Five things in one line. They are, in order of importance:
The dollar size. $48.2M is large in absolute terms. It is large enough that it’s almost certainly institutional — a retail trader does not deploy $48M into a single options strike unless they are about to make the news for entirely different reasons. The dollar size is what makes this notable, not the contract count, not the volume.
The strike. Calls at 950 on NVDA spot of, say, 928. That’s a 2-to-3 percent out-of-the-money strike. Slightly aggressive but not lottery-ticket aggressive. Not a 1000 strike screaming ‘this is gambling.’ Not a 920 strike that’s basically a synthetic long. The strike implies someone with a directional view but enough conviction to want delta gearing, not just stock exposure.
The expiry. May 17. If you’re reading this on May 5, that’s a 12-day option. Twelve days is short but not 0DTE-reckless. Twelve days is ‘something is going to happen in this window.’ The earnings call, a product announcement, a macro print — the buyer thinks something specific catalyses inside this window.
The vol/OI ratio. Volume to open interest of 2.0x. This is the part that filters out noise. A 2.0x ratio means today’s volume was double the existing open interest. New positions were opened at a rate that materially expanded the total interest in this strike. This is not existing holders trading among themselves. This is fresh conviction.
The bias. Calls, bought, on size. Read: bullish. (If this had been a put buyer or a call seller of comparable size, the bias label would flip.)
The four questions before you act
Before this becomes a trade, four questions.
One: does this match anything you already think?
If you have a long NVDA thesis already logged, this is corroboration. If you’re flat NVDA, this is a possible signal. If you’re short NVDA, this is a meaningful warning. The first response to an unusual flow alert is never to copy it — it is to read it against your own existing thesis.
If you don’t have a thesis on this name yet, this signal alone is not enough. Institutional flow is one input. The setup, the regime, the earnings calendar, the macro context — all of those exist before the flow signal and they all matter equally.
Two: what’s the calendar?
A 12-day option on a megacap during a known earnings window is one trade. A 12-day option on a megacap with no scheduled news is a different trade. Mantis surfaces the next earnings date alongside the symbol, so this question gets answered fast: NVDA reports May 22. The May 17 calls expire before earnings. Someone is positioning ahead of earnings, not into the print itself. They expect the run-up move, not the post-print move.
That subtlety matters. A buyer positioned for the run-up is betting on sentiment building. A buyer positioned through earnings is betting on the print itself. The 17-strike, 17-expiry combination is a specific kind of bet — and one you can either disagree with or align to.
Three: where’s the regime?
Open the Macro Regime gauge. If it reads 72, the macro tape is risk-on and breakout setups have a tailwind — a bullish NVDA flow signal aligns with the regime and the conviction signal is amplified. If it reads 28, the macro tape is risk-off and you are looking at someone fighting the tape with size — interesting, but the burden of proof on the trade is higher. You’d want a closer expiry, smaller size, or both.
The regime doesn’t tell you whether to take the trade. It tells you what kind of size to take.
Four: do you already have NVDA exposure?
If you’re long stock, an aligned flow alert is a hold signal — maybe even an add — but it’s not necessarily a new options buy. If you’re long the stock plus already long calls, you are now overweight a single name and a concentration check matters more than a new entry. Risk Manager will flag this for you before you act.
If you’re flat, this becomes a setup candidate. Not yet a trade.
What the trade looks like, if it happens
If your answers line up — your thesis aligns, the calendar makes sense, the regime cooperates, your existing book has room — the trade is still not ‘buy the same option they bought.’ Copying the institutional position is rarely the right move. They have a different cost basis, a different time horizon, a different exit framework, and a different size relative to their book than you do.
The signal tells you ‘something is moving here.’ The trade is your own setup, sized to your own book, with your own stop. Maybe that’s stock. Maybe it’s a slightly-different-strike call you find more attractive. Maybe it’s a spread that captures the directional view at lower premium cost. Maybe it’s waiting one more day to see if the price action confirms the flow signal.
Whatever the trade is, log it in the journal at entry time with the flow signal noted in the context — Mantis attaches the smart-money tag automatically so you can look back later at how your flow-driven trades performed as a category. That category-level analytics is where the discipline compounds. Some traders are good at reading flow. Some are bad at it. Most don’t know which they are because they never tag the trades.
What the signal is not
The signal is not ‘NVDA is going to 1000 in two weeks.’ Institutions buy options and lose money on them all the time. The signal is one input. Strong input — large dollars, clear bias, new positioning — but one input.
The signal is also not always actionable. Some days the digest reads cleanly aligned with your existing thesis on a name you already hold. Action: hold. Some days the digest contradicts your existing thesis. Action: re-examine the thesis. Some days the digest is on a name you don’t follow and have no edge on. Action: ignore.
About a third of the alerts in any given week become decisions. Two-thirds get acknowledged and dismissed. That’s the right ratio. If you are acting on every alert, you are over-trading. If you are acting on none, you are not paying enough attention.
What the next ninety seconds buy you
A retail trader without this tool is reading about that $48M NVDA flow in a newsletter sent the following Tuesday, after the move has happened. The information is the same. The asymmetry is the timing.
The Smart Money digest costs you ninety seconds a day. Most of those ninety seconds end in ‘nothing to do here.’ The rare days they end in a trade are the days the subscription pays for itself for the entire year.
Try Mantis free for 14 days.
