← All notes

FIELD NOTE / 002

NVDA Options Tape: Defensive Flow and Late Dip-Buying

NVDA tape — desk read

NVDA’s tape was decisively bearish and defensive through 15:30, followed by meaningful structured dip-buying near the lows. The closing flow softened the damage, but it did not reverse the day’s larger bearish signal.

NVDA fell from $227.36 to $217.51, down $9.85 or 4.33%. It traded as high as $229.10 and as low as $216.91, closing only $0.60 above that low.

Tape overview

  • 233,182 reported trades
  • 1.96 million contracts
  • $1.063 billion gross option premium
  • Calls: $719.7 million, or 67.7% of premium
  • Puts: $343.5 million, or 32.3%
  • Explicit spread premium: $263.0 million
  • Clearly classifiable single-leg delta pressure: −3.49 million shares
  • Exact-deduplicated result: −3.27 million shares
  • Clearly bearish premium share: 59.8%
  • Execution side unknown for 57.5% of gross premium

The raw and deduplicated results tell the same story. This was not a conclusion created by duplicate prints.

The high call-premium share is misleading by itself: a substantial amount of the call activity was call selling, particularly in longer-dated expirations.

How the session developed

Session NVDA move Classified delta pressure Read
09:30–10:00 −$2.12 −474,000 Immediate defensive tone
10:00–12:00 −$3.83 −625,000 Selling broadened
12:00–14:00 −$2.02 −1.22 million Bearish pressure accelerated
14:00–15:30 −$2.14 −1.32 million Most forceful part of the selloff
15:30–16:00 +$0.22 +146,000 Constructive offset near the lows

The close recovered only about 4% of the full-day negative delta estimate. It was genuine dip-buying, but not enough to establish that the bearish campaign had been reversed.

What drove the bearish reading

The pressure was spread across the term structure:

  • August 31: −550,000 delta, dominated by afternoon put buying.
  • September 18: −546,000, combining call selling and put buying.
  • January 15, 2027: −391,000, driven primarily by substantial call selling.
  • October 16: −331,000.
  • September 2: −311,000.
  • September 25: −258,000.
  • September 4: −242,000.
  • October 2: −220,000.
  • March 2027: −215,000.

That breadth matters. This was not merely traders buying a few front-week crash puts. The tape also showed investors selling longer-dated upside exposure.

The largest bearish outright flows included:

  • August 31 $220 puts: approximately −396,000 raw delta. Purchases substantially exceeded sales. This contract was duplicate-sensitive, but remained strongly bearish after deduplication at approximately −253,000.
  • January 2027 $200 calls: −293,000; about $19.7 million sold versus $5.0 million bought.
  • September $220 calls: −165,000; roughly 4,405 contracts sold versus 1,090 bought.
  • January 2027 $220 calls: −121,000.
  • March 2027 $235 calls: −121,000, with approximately 2,523 sold and only ten classified as bought.
  • September 2 $222.50 puts: −108,000, heavily purchase-dominated.
  • October 2 $190 puts: −57,000, including nearly 5,900 bought—clear tail-hedging demand.

Major bullish offsets

There were several important counterflows:

  • September 9 $240 calls: approximately +281,000 delta, with 32,037 contracts classified as bought against 4,427 sold. This looks like concentrated upside reloading, though the contract hit the export’s 1,000-print cap.
  • January 2027 $190 puts: +156,000, primarily put selling.
  • January 2027 $260 calls: +133,000, heavily bought.
  • December $195 puts: +76,000 from put selling.
  • December $215 calls: +73,000 from call buying.
  • August 31 $217.50 calls: +39,000, concentrated from 15:42 through the close.

These flows support the idea that some participants treated the late-day lows as an opportunity. They do not outweigh the broader defensive activity.

Reconstructed spread activity

The most significant bullish package was a January 2027 risk-reversal candidate:

  • Approximately 5,000 January $260 calls bought around $7.65.
  • Approximately 5,000 January $190 puts sold around $7.80.
  • Near-flat economics, producing roughly a $75,000 net credit.
  • Estimated package delta: +255,000 shares.
  • Executed primarily around 15:30, with small related prints extending toward 15:50.

This is a strongly bullish structure: upside calls financed by selling downside puts. One principal slice was recovered through exact timestamp-and-size matching rather than an exchange-declared strategy condition, so it should be viewed as a strong package candidate rather than absolute proof.

A second important campaign was a September 4 $195/$207.50 bull-put spread:

  • Lower $195 puts bought.
  • Higher $207.50 puts sold.
  • Approximately 11,381 matched contracts.
  • Roughly $800,000 net credit.
  • Estimated delta: +130,000 shares.
  • Executed in numerous slices between 12:59 and 13:54.

That looks like structured bullish support well below the prevailing stock price.

Several large packages could be linked but not reliably assigned a direction:

  • June 2027 $220/$260 call package: approximately 2,001 contracts and $10.7 million gross premium.
  • September 4 $222.50/$227.50 calls: approximately 5,295 contracts. The lower call traded on the sell side, but the upper leg was midpoint/unknown; it may be a bearish call spread or roll-up, but the evidence is insufficient to force that conclusion.
  • February/March 2027 $205 put calendar: approximately 4,600 matched contracts and $12.3 million gross premium; direction unknown.

Bottom line

My best diagnosis is:

A broad institutional defensive session—front-end put buying plus substantial long-dated call selling—followed by late structured dip-buying near the session low.

The closing risk reversal, bull-put spread, call purchases and put sales suggest sophisticated buyers began accepting bullish risk around $217–$219. But the day’s −3.49 million classified delta estimate, the breadth across expirations, and the persistent long-dated call selling make the overall tape bearish.

I would treat the late bullish structures as evidence of a possible support attempt, not yet evidence of a durable reversal. Confirmation would require the next tape to show reduced put demand, cessation of long-dated call selling, and continued call buying above the late-day strikes.

Data caveats: the feed contained 19,089 exact duplicate-excess rows and 81 contracts at the export’s 1,000-print limit. The broad bearish result survives exact deduplication, though individual contract totals—especially the August 31 $220 puts—should be interpreted with those limitations. The feed contains no open-interest transition data, so opening versus closing activity cannot be proven.