Nvidia reported fiscal second-quarter revenue of $96.22 billion Wednesday, beating Wall Street's $92.17 billion consensus by roughly $4 billion, and told investors to expect 70 percent revenue growth in fiscal 2028 — nearly double the 44 percent analysts had modeled. [1][2] Buried eleven items down in CNBC's live blog of the results, the same quarterly filing disclosed something new: for the first time, Nvidia's 10-Q lists "indebtedness" as a standalone risk factor, warning that rising obligations "could adversely affect" its financial condition and cash flows. [2]

The scale of the shift is stark. Nvidia's debt due within one to five years jumped to $15 billion in the new filing, up from $2.75 billion in the company's previous quarterly report — more than a fivefold increase in a single quarter. [2] The company also disclosed $33.5 billion in senior notes outstanding and an active $25 billion commercial paper program, the residue of what CNBC described as Nvidia's first bond sale since the AI boom began. [2] Wednesday's earnings extended this paper's Tuesday preview, which flagged Nvidia's own floor guidance sitting below the Street's number heading into the print; the actual results cleared both bars by a wide margin, even as the debt disclosure went unmentioned in that earlier coverage.

For the quarter ended July 26, Nvidia posted GAAP net income of $59.69 billion, more than doubling the $26.42 billion booked a year earlier, on gross margins that held at 75 percent for a second straight quarter. [1] Data center revenue — the business now driving 92 percent of total sales — came in at $89 billion, up 117 percent year over year and ahead of the $86.33 billion analysts had modeled, according to StreetAccount data cited by CNBC. [2] Diluted earnings per share reached $2.46 on a GAAP basis and $2.22 on a non-GAAP basis, versus a $2.10 analyst estimate. [2]

Chief executive Jensen Huang framed the results as evidence that the AI buildout has broadened beyond a handful of dominant labs. "This time last year, one lab alone was driving the buildout," Huang said in the earnings release. "Today, we have a golden age of new AI labs and startups, multiple frontier labs scaling in parallel, a thriving open-model ecosystem and physical AI coming online." [1] On the earnings call, Huang went further on Nvidia's own venture bets, telling analysts, "The only regret that I have is that I didn't invest more and sooner," and predicted OpenAI and Anthropic "will likely go public soon." [2]

The company's outlook for the current quarter also topped expectations: Nvidia guided to $108 billion in revenue, plus or minus 2 percent, against a Street consensus of $104.2 billion — and said the figure assumes zero data center compute revenue from China. [1][2] Chief financial officer Colette Kress told analysts that customer forecasts "point to our growth doubling next year," while cautioning that the guidance itself reflects supply constraints rather than a ceiling on demand. [2]

That supply story has a financing dimension the CNBC live blog treated as a footnote but which sits at the center of the new debt disclosure. Nvidia's supply commitments — chiefly memory procurement — more than doubled quarter over quarter, from $119 billion to $279 billion. [2] The company also disclosed that Amazon Web Services will buy 2 million Nvidia GPUs, with Huang telling CNBC's Jim Cramer that the deal additionally includes "millions of CPUs" built on Nvidia's new Vera chip. Kress told analysts that combined capital expenditure among the "top five hyperscalers" is projected to rise to $1.3 trillion next year from $800 billion in 2026 — the demand pool Nvidia is now borrowing against.

This is where the market's two live readings of Wednesday's results diverge sharply. CNBC's coverage treated the earnings as a conventional beat-and-raise story, threading in incidental color about SpaceX, Elon Musk and Nvidia's investment gains along the way, and noting almost as an aside that shares had fallen the day after each of Nvidia's prior four earnings reports despite beating estimates every time. [2] Finance commentators on X read the same filing as confirmation that Nvidia's growth now runs on borrowed capital. Ed Zitron, whose posts this paper has tracked through Nvidia's financing buildout, wrote that "its $674bn+ in FY28 revenues are contingent on hundreds of billions in debt," pointing to the jump in near-term obligations as evidence the company's forward guidance is underwritten by leverage rather than organic cash generation. The Kobeissi Letter ran a parallel calculation from the 70 percent growth guide, posting that the figure "implies $670+ BILLION in revenue next year, up +2,390% from FY2023," and noting the $108 billion third-quarter guide alone.

Nvidia's own balance sheet gives some shape to that debate without resolving it. Total liabilities rose to $91.29 billion as of July 26, from $49.51 billion at the start of the fiscal year, with long-term debt climbing to $32.37 billion from $7.47 billion over the same six months. [1] Total shareholders' equity, by contrast, grew to $228.98 billion from $157.29 billion, and the company still returned roughly $26 billion to shareholders in the quarter through buybacks and dividends, with $99 billion remaining under its repurchase authorization. [1] Nvidia is not a company running short of cash; it is a company that has begun, for the first time, treating its own leverage as material enough to warrant a dedicated warning to investors — a warning the CFO's own commentary framed defensively. "We want to be direct about this, rather than let it linger as an open question," Kress told analysts of the memory-driven margin pressure Nvidia expects to bottom out at 71 to 72 percent in the fiscal fourth quarter. [2]

The debt disclosure also lands alongside Nvidia's separately announced financing architecture: a set of partnerships unveiled this month with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR meant to mobilize more than $500 billion in third-party capital for AI infrastructure, an arrangement Morgan Stanley's credit desk flagged this week as directionally sound but "too early-stage, opaque, and sizable" to fully underwrite. [1][2] That financing web — vendor backstops, compute-for-equity swaps, hyperscaler debt raises Nvidia partially guarantees — is precisely what Zitron and other finance-focused X accounts have spent recent months mapping, and Wednesday's new risk-factor language is the first time Nvidia's own lawyers have used comparable words to describe it.

None of this changes what actually happened Wednesday: a company already generating $60 billion in quarterly net income posted a clean beat against consensus and guided materially higher than the Street expected. But the newly stated risk factor is a change in how Nvidia itself is willing to describe its financial position, and it arrived in the same filing as a guide implying, on Zitron's and Kobeissi's competing math, somewhere between $670 billion and $674 billion in fiscal 2028 revenue — a number that assumes the debt financing underpinning the AI buildout continues to be available on favorable terms. Nvidia shares jumped roughly 4 percent in the immediate reaction to the print, CNBC reported, even as the outlet separately noted the stock had fallen the day after each of the company's previous four earnings reports. [2]

The question of who ultimately pays for that buildout runs directly through Nvidia's own customer list, and CNBC's live blog raised it without resolving it. Amazon and Alphabet both reported negative free cash flow for the quarter, as did Elon Musk's Tesla and SpaceX; Meta, meanwhile, saw its cash generation shrink by roughly 90 percent. [2] Those four companies are among Nvidia's largest buyers, and their spending is increasingly financed rather than self-funded — Goldman Sachs projects hyperscaler AI infrastructure spending will reach $1.2 trillion in 2027, up from the $800 billion to $1.3 trillion range Kress cited for 2026 and next year combined. [2] Nvidia's own role in that financing chain has grown alongside it: the company disclosed a $21 billion stake in SpaceX this month, layered atop an earlier $10 billion investment in Musk's xAI, and confirmed the two companies are jointly developing an orbital data-center concept called Starmind. [2] Every dollar Nvidia invests in a customer's equity or backstops in a customer's debt is a dollar that shows up, eventually, as either investment income or investment risk on Nvidia's own books — and Wednesday's filing is the first time the company has asked investors to weigh that risk explicitly rather than infer it from press releases.

Memory costs compound the pressure from the other direction. Server DRAM prices rose 64 percent in the second half of last year and Trendforce projects a 260 percent jump in 2026, a shortage Nvidia's own purchasing has helped drive given its need for both high-bandwidth memory in its GPUs and conventional DRAM for complete systems. [2] Nvidia said it spent $145 billion in the first quarter alone to secure supply, and Kress acknowledged the company is "not immune to supply challenges" even as she expressed confidence in meeting the growth opportunity ahead. [2] Set against $279 billion in total supply commitments — more than double last quarter's $119 billion — the picture is of a company simultaneously financing its customers' compute purchases and racing to lock down the physical inputs, primarily memory, needed to build the chips those customers are buying. That is the operational reality underneath both the beat-and-raise headline and the new indebtedness warning, and it is why the debt disclosure, however far down CNBC ran it, may prove more durable news than the quarter it was filed alongside.

Nvidia's next quarterly report, covering the period through late October, will be the first test of whether the $108 billion guide — and the debt load now financing it — holds.

-- THEO KAPLAN, San Francisco