Nvidia's chief financial officer told analysts this week that shipments of chips approved for sale in China made up less than 1 percent of the company's data center revenue last quarter — and that the company's guidance for the current quarter assumes zero data center compute revenue from China at all. [1] The disclosure came in the same earnings call this paper covered Wednesday for a separate reason: Nvidia's quarterly filing listed "indebtedness" as a standalone risk factor for the first time, with debt due within five years jumping more than fivefold in a single quarter. The China numbers are the other half of a story CNBC's live blog covered as a straightforward beat-and-raise: Nvidia is simultaneously locked out of its largest addressable market and squeezed by rising costs at home, even as the headline results sailed past Wall Street's estimates.
Nvidia received a green light from the US government late last year to sell certain older chips into China, but Chief Financial Officer Colette Kress said in written commentary that it remains unclear to what degree Chinese authorities have actually approved the imports. [1] Sales have stayed negligible as a result — the sub-1-percent figure Kress disclosed this week — despite the regulatory clearance that was supposed to reopen the door. Nvidia's guidance for the quarter now underway does not merely discount China; it assumes the market contributes nothing at all to data center revenue, a planning assumption that treats the world's second-largest economy as effectively absent from the company's near-term math. [1]
The cost side of the ledger tells a related story. Nvidia's gross margin held at 75 percent for a second straight quarter, but the company guided that figure down to 74 percent for the current quarter and said it expects margins to bottom out at 71 to 72 percent in the fiscal fourth quarter — a decline Kress attributed largely to memory prices. "We want to be direct about this, rather than let it linger as an open question," she told analysts, describing a memory shortage she said is being driven in significant part by the AI buildout itself. [1] Nvidia's own supply commitments, chiefly for memory procurement, more than doubled in a single quarter, from $119 billion to $279 billion. [1]
Neither fact undid the headline numbers. Nvidia reported $96.22 billion in quarterly revenue against a $92.17 billion Wall Street consensus, and guided to $108 billion for the current quarter against a $104.2 billion estimate — a clean beat on both counts, and the reason CNBC's coverage read as an unambiguous win. [1] Data center revenue alone reached $89 billion, up 117 percent year over year, and now accounts for 92 percent of total sales. [1]
What the earnings call actually described is a company managing two problems that don't show up in the same paragraph of most coverage: a geography it cannot yet monetize despite regulatory clearance, and a supply chain whose costs are rising faster than Nvidia's own pricing power can fully absorb without giving up margin. Both facts were disclosed voluntarily, in Kress's own commentary, rather than extracted under pressure — which makes it notable that neither made it into the lead of the day's dominant earnings narrative. Nvidia's stock still rose roughly 4 percent on the print. [1] The company that just told investors it is planning around zero China revenue and thinning margins is, for now, still being read by the market as an unambiguous growth story.
-- DAVID CHEN, Beijing