The market paid 7.63% overnight for a growth rate Nvidia measured off its own supply

A growth rate with no base attached is an instruction to guess, and Nvidia handed the market 70% for fiscal 2028 without naming the dollars that 70% grows from ^[5]. What traded overnight was therefore arithmetic on a ratio: 1.70 over the 1.44 analysts had modelled works out to roughly 18% more fiscal 2028 revenue than consensus was carrying ^[17], a 26-point gap in headline growth ^[18] sitting on a base that will itself move when fiscal 2027 prints.

Run the sequence instead of the guide and the number gets less thrilling, or rather, it gets more interesting. The quarter just reported grew 18% on the one before it ^[3], and 1.18 compounded four times is 94% ^[19], so the 70% Nvidia has committed to is a deceleration from the pace it is currently setting. Which is exactly what you would expect from a company whose stated binding constraint is its own supply chain, Huang having said that demand is much greater than 70% while supply allows a confident 70% ^[7].

That distinction is the whole thesis, because a seller who cannot fill 100% of demand ^[7] never publishes a demand curve; it publishes a shipping record. The $96.2 billion quarter ^[2] against an implied year-ago quarter of about $46.7 billion ^[14] says output roughly doubled ^[8], not that customer budgets doubled, and four quarters at the current rate annualises to $385 billion ^[16] of hardware someone has to finance. Fortune reported the forecast being read as an answer to AI bubble critics and to what it called circular-financing doomsayers ^[12], and Melissa Otto of S&P Global's Visible Alpha put the market reaction at "Whoa, 70%" ^[6]. Using a supply-capped figure to settle a demand argument has one flaw: while the cap binds, strong demand and merely adequate demand produce the same revenue line.

Three ways this goes. Nvidia raises the fiscal 2028 number as capacity lands, in which case the supply story was conservatism and the demand is real. Or capacity lands and the guide is reaffirmed rather than raised, which is the first observable a sceptic gets, because reaffirmation with more supply available means the queue was shorter than advertised. Or 70% is delivered on schedule and the argument moves to who paid for it, which is where the circular-financing question ^[12] actually lives, on customer balance sheets rather than Nvidia's income statement.

This is probably wrong in one direction, and here is my view anyway: the price move discounts a demand certainty that the disclosure cannot contain, since the 7.63% after-hours gain ^[1] responded to a ceiling, not a floor. Note how little of it travelled. Nasdaq 100 futures were up 1.16% ^[9] against Nvidia's own 7.63% ^[1], about a seventh of the move ^[20], and the beat itself was $4.0 billion, or 4.3% above the $92.2 billion consensus ^[13] on the quarter, which is a good print rather than a discontinuity. What would prove me wrong is straightforward and dated: a raise above 70% accompanied by evidence that the incremental orders are funded out of customers' operating cash. Until then the market owns a factory estimate.