← back to the pulse
pulse note nvda

michael burry is buying nvidia puts. our kill criteria says he might be right.

we track 6 kill criteria on nvidia. one just triggered. four are in the watch zone. here is the scorecard.

xvary research desk · march 25, 2026 · ~3 min read

Michael burry is back. The man who shorted the housing market before 2008 is now buying put options on nvidia and palantir. His thesis: the AI trade is a bubble, credit stress is building ($1.2 trillion in credit card debt), and the market is pricing perfection in a world that does not deliver it.

That is not a new take. What is new is that our data agrees with part of it.

the scorecard. we track 6 kill criteria on nvidia. these are the specific, measurable conditions that would break the bull thesis. not "regulatory risk" or "competition" — named thresholds with current values. here is where they stand today:

1. free cash flow yield: DANGER. threshold: 3.0%. current: 2.4%. nvidia is generating less free cash per dollar of stock price than our minimum. at current prices, the stock is not paying you enough in cash generation to justify the valuation. this is the one burry is right about.

2. gross margin: WATCH. threshold: 65.0%. current: 71.1%. comfortable — but narrowing. nine percentage points of cushion sounds like a lot until you realize it was fifteen points eighteen months ago. the direction matters as much as the distance.

3. operating margin: WATCH. threshold: 55.0%. current: 60.4%. same story. safe today. the trend is what you watch.

4. revenue growth: WATCH. threshold: 43.6% YoY. current: 65.5% YoY. still well above. but the market is pricing acceleration, not deceleration. if growth slows to 44% — still extraordinary by any normal standard — the thesis is under pressure because the multiple assumes much more.

5. ROIC: SAFE. threshold: 50.0%. current: 70.3%. the moat is intact. nvidia still generates exceptional returns on invested capital. this is the strongest pillar of the bull case.

6. EV/Revenue: WATCH. threshold: 15.0x. current: 19.8x. trading at nearly 20 times revenue. that is not cheap. but it was 30x a year ago, so the compression has already started.

the honest assessment. burry is right about one thing: nvidia's valuation does not leave room for error. a free cash flow yield below 3% means the stock price has outrun the company's ability to generate cash. that is a real concern, not a vibes call.

But the full picture is more nuanced than "nvidia is a bubble." the company still has 70% ROIC, 71% gross margins, and 65% revenue growth. Four of six kill criteria are in the watch zone — approaching their thresholds but not there yet. The thesis is stressed, not dead.

what would make burry wrong. if nvidia's next two quarters show revenue growth above 43% and gross margin holds above 65%, the thesis survives. the market will forgive a low FCF yield if the growth engine keeps delivering. growth covers a lot of sins in valuation.

what would make him right. if gross margin drops below 65% for two quarters — meaning competition or pricing pressure is real — and revenue growth decelerates below the market-implied 43% pace, the multiple compresses hard. at that point, burry's puts print.

The next earnings report is the decision point. Nvidia reports in late may. That is 60 days away. Both sides of this trade are betting on the same filing. The difference is which line they are watching.

key takeaways

  • Six named kill criteria on NVDA: one hit (FCF yield below floor), four in watch (margins, growth, EV/revenue), ROIC still in safe territory in the article's scorecard framing.
  • The bull survival path in the piece is continued revenue growth above ~43% YoY with gross margin holding above ~65% across the next two prints; misses compress the multiple story.
  • Late may earnings is treated as the decision catalyst — both directional trades are keyed off the same filings and line items.

faq

What part of burry's thesis does the xvary data agree with in this note?

Valuation pressure tied to cash generation versus price: the write-up flags sub-3% free cash flow yield vs a 3% threshold as the breached criterion — i.e. Less room for execution error at prevailing multiples.

Does one red criterion mean the bull thesis is dead?

The article argues stressed versus dead: several other thresholds are still in a watch band (not fully breached) and ROIC is cited as still strong — the next earnings are framed as the test of whether growth and margins hold the line.

What numbers does the verdict say to watch on the next print?

Gross margin remaining above about 65% and revenue growth remaining above about 43% YoY — those are the lines called out for validating or challenging the bull case in the narrative.

nvda · full deep dive → nvda · snapshot →