Loading today’s stories
Invest2 distinct publishers3 min readPublished
The first year-ahead forecast in Nvidia's history implies about $673 billion in fiscal 2028, roughly $103 billion above consensus, and hands every capacity planner a figure to hold it to.
The Investor · Invest desk
Compiled by The InvestorSomething wrong?How this is made
The forecast has an audience beyond the sell side. Huang said he published it because he can already see next year's compute requirements, and because he wanted to be consistent with the partners who supply the land and power that surround his chips [3]. That makes the 70% a scheduling instruction as much as a guide: a substation contract and a memory allocation both need a figure to size against, and there is now one figure in the record that every counterparty is reading.
It also makes it falsifiable. On the LSEG consensus of $396 billion for fiscal 2027, 70% growth puts fiscal 2028 at about $673 billion [4], which is $103 billion, or 18%, above the $570 billion the street had been modelling [5][1]. The bubble debate has been an argument about sentiment. It now has a quantity attached, and the quantity gets checked four times a year.
Anyone doing the checking should be careful about the base. Fortune's arithmetic on the same statement produces $690 billion to $700 billion [6]; CNBC's produces $673 billion [4]. The $17 billion to $27 billion between those two readings [3] is not rounding, it is disagreement about where fiscal 2027 actually finishes, and it is bigger than the annual revenue of most companies in the supply chain. Pick your base before you grade the result.
The more interesting number is the one Nvidia is declining to book. Kress said customers' forecasts pointed to the company's growth doubling next year [8], and Huang said demand is much greater than 70% while supply allows him to confidently deliver 70% [7]. Doubling $396 billion is $792 billion; the distance from $673 billion is about $119 billion of demand the company is telling you in advance that it does not expect to fill [4]. Huang's description of the position was that the entire supply chain is challenged and everybody is running flat out [9]. Memory is the named pinch [10].
Then there is the guarantee. Nvidia disclosed maximum gross guarantee exposure of $108.5 billion [13]. That is roughly 16% of the implied fiscal 2028 revenue [6] and about 13% more than the $96.2 billion the company just reported for a single quarter [11][7]. Kress's answer to the circular financing charge was that the frontier labs are limited by compute rather than by technology or customer demand, and that she expects them to become the largest technology companies in history [14]. She may be right. It remains an arrangement in which a shortfall in the buildout shows up on two lines at once.
The operational requirement behind the headline is plain enough. $673 billion averages about $168 billion a quarter, 56% above the $108 billion Kress guided for the current one [5][12]. Nvidia has cleared steps of that size before; the most recent took quarterly revenue to $96.2 billion, up 106% from a year earlier [11]. What is different is that the finish line has been published, by the company, in advance.
Ranked by verification strength, evidence, and original report placement.
CFO Colette Kress told investors that revenue growth in fiscal 2028 will reach 70%, against a 44% average analyst estimate, according to LSEG.
Huang said on the call: "It is the case that we've never forecasted, never guided to a year in advance."
Huang told analysts he provided the forecast because he has visibility into next year's needs for computing power now, and wanted to be consistent with partners who provide complementary data center inputs such as land and power: "we wanted to make sure that everybody has the same set of information."
The 70% preliminary expectation exceeded analyst projections for fiscal 2028 revenue of roughly $570 billion, or 44% growth from fiscal 2027.
Fortune calculated that Nvidia's projected growth, applied to fiscal 2027 revenue and expected revenue, would imply fiscal 2028 revenue of $690 billion to $700 billion, more than $100 billion over the $570 billion analysts had been modeling.
Huang, answering an analyst: "Even though our demand is much greater than 70%, our supply allows us to confidently deliver 70%."
Follow any of these and your For You feed starts watching them — no settings page required.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Strong on-record disclosure, weak independent verification
Two established business outlets covering the same call converge on the 70% figure, the supply-constrained framing and the broadening customer base, and one reports full quarterly detail ($96.2 billion, +106%, against $92.2 billion expected) plus the $108.5 billion guarantee disclosure. But the load-bearing element is company guidance about a year that has not happened, only one outside voice (S&P Global's Visible Alpha) appears in the cluster, and the two outlets derive materially different dollar figures from the same percentage - which caps evidence quality below the level the trailing numbers alone would earn.
Realized demand already at scale; the forecast year is unproven
Adoption of Nvidia's platform is measured, not aspirational: $96.2 billion of quarterly revenue up 106%, a $108 billion next-quarter guide at the top of the whisper range, non-hyperscale customers at about half of business growing 100% a year, and a financing program built specifically to let less creditworthy buyers take delivery. What is not yet adopted is the fiscal 2028 step-up itself - reaching the implied annual figure requires roughly $168 billion a quarter, about 56% above the quarter just guided.
Modestly overstated, but deliberately scoreable
The forecast is a company self-projection delivered after the close, and it arrives wrapped in superlatives - customers' forecasts implying a doubling, frontier labs becoming 'the largest technology companies in history', investments described as low-risk and high-reward. Downstream coverage then inflates the number further, with one implied range $17-$27 billion above the straightforward arithmetic. That pushes the gap positive. It is held near alignment rather than far above it because the trailing quarter is hard, management guided below the demand signal it cited, and the 70% figure is precise enough to be checked against actual results.
Heavily company-shaped disclosure with a direct market payoff
Nvidia broke a standing practice against year-ahead guidance and released the number after the close, with the stock rallying more than 4% in after-hours trading; the same call was used to pre-empt the 'circular financing' critique and to defend nearly $50 billion of frontier-lab investments, $108.5 billion of guarantee exposure and $500-billion-plus capital partnerships. Huang also said an explicit purpose was to coordinate suppliers of land and power around Nvidia's own numbers. Both articles are built almost entirely from management quotes, with one third-party analyst voice between them, so promotional incentive is high and independent counterweight is thin.
Confident on what was said, less so on what it implies
The utterances, the quarterly figures and the disclosures are firmly established by two independent outlets publishing within minutes of each other with direct quotes. Confidence is reduced by the unresolved divergence in implied fiscal 2028 revenue, the absence of any source outside the call other than one analyst comment, the truncated end of one source body, and the fact that the story's central assertion is a forecast that cannot be verified yet.
invest
Nvidia's pre-earnings re-rating is a credit story, not a chip story1 distinct publisher
product
The $500bn compute asset class rests on a depreciation curve Nvidia once denied1 distinct publisher
invest
Broadcom's $100B Chip Raise Turns AI Capacity Into a Credit Question1 distinct publisher
build
NVIDIA books $96.2bn at 75% margins and guides 74%: compute is not getting cheaper4 distinct publishers
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 26, 2026
1 article · August 26, 2026