Monday the argument was positioning; Tuesday it finally got numbers. Nvidia reports its fiscal second quarter — the period ended July 26 — after Wednesday's close against a grading sheet now precise enough to argue with. The company has guided revenue to $91 billion; Wall Street's consensus sits higher at $92.18 billion, which would be growth of 97 percent over the year-ago quarter. On earnings per share the gap repeats: management's implied figure is about $2.04 against a Street estimate of $2.09. [1]
Two structural details hide inside those figures. First, guidance assumes zero data-center AI chip sales to China, so the single most policy-exposed line in technology is already written out of the bar the stock will be judged against. The construction is asymmetric by design: any quarter that includes even partial Chinese revenue reads as pure upside against a zero baseline, while every licensing restriction Washington imposes has been pre-paid in the guide. Second, the next quarter's consensus of $103.9 billion means Wednesday's real verdict is not the reported quarter but whether the guide makes that number look conservative or heroic. [1]
The full-year arithmetic explains why so much hangs on one paragraph. Analysts model roughly $395.7 billion of fiscal 2027 revenue — an increase of about 83 percent over the prior year — and earnings near $9.02 a share, which prices the stock around 24 times forward estimates even after a fivefold market cap. Those are numbers that assume the buildout keeps compounding without a down quarter. Wednesday does not test a year of demand; it tests whether one more quarter of it remains believable at that multiple. [2]
History offers both camps ammunition. Nvidia has beaten consensus earnings in 22 of the past 24 quarters, yet the stock fell the day after each of its last four reports — declines ranging from 0.8 to 5.5 percent despite average beats of 4.6 percent. Beats, in other words, stopped paying years ago; only guidance pays now, and only the unscripted sentences in it. [1]
The analyst ledger tilts bullish on paper: 61 analysts carry a Strong Buy consensus and an average target of $305.41, roughly 43 percent above Tuesday's close of $213.05, with Raymond James raising to $352 and BMO holding $340 straight into the print. The low target on the same distribution is $180. [2] A market cap near $5 trillion makes that spread arithmetic worth more than any single rating.
Tuesday's tape did what Monday's slide predicted in reverse: indexes closed higher with semis leading, the S&P 500 up about a third of a percent to 7,677.28 and the Nasdaq up 0.7 percent to 26,151.30, Nvidia gaining 2.19 percent into the bell. [2][3][4] Read it as relief that the exam date has arrived, not as an answer key.
The make-or-break framing is honest for once. When one company's guidance paragraph functions as the index's earnings, a two-percent gap between guide and consensus is not a rounding error; it is the entire debate about AI capital spending compressed into a decimal.
-- THEO KAPLAN, San Francisco