Nvidia (NVDA) stock is currently trading near $225, up nearly 19% in 2026 and close to 15% from its late-June low. Though that’s hardly weak, it is noticeably more sluggish than last year, when shares surged about 39%, according to Yahoo Finance data.
The Nvidia story is incredible to say the least, with veteran analysts like Jim Cramer describing its remarkable influence as:
“The barometer for what might be as much as a third, some would say a half of the economy.”
A lot of the hesitation we’re seeing of late has less to do with AI demand and more to do with “vendor financing”.
Simply put, investors worry that Nvidia is effectively funding AI customers who buy its chips, exposing the company to credit losses, making demand look a lot more circular than organic.
That debate exploded in late July amid reports that Nvidia may commit around $250 billion to an OpenAI data-center project in Ohio, as WSJ reported.
Additionally, financing for up to $350 billion in chip purchases was discussed.
Then, on August 10, Nvidia revealed a Wall Street-backed platform targeting more than $500 billion in third-party AI infrastructure capital, according to an Nvidia press release.
On Monday, August 17, it added another wrinkle to that by agreeing to guarantee $105 billion tied to OpenAI’s Ohio campus, as reported by Reuters.
Nevertheless, in a note shared with me, Bank of America veteran analyst Vivek Arya says the market is pricing that risk too aggressively and remains bullish despite rising off-balance-sheet exposure.
Carl Court/Getty Images
Bank of America says Nvidia’s risk is already priced in
Bank of America argues that the market’s probably discounting a far worse outcome than the financing commitments justify.
More Nvidia:
- Nvidia just made a move Wall Street wasn’t ready for
- Nvidia just locked down deal that changes AI race
- Nvidia stock is doing something it hasn’t done in years
Analyst Vivek Arya reiterated a Buy rating and $350 price target, implying nearly 55.5% upside from the $225.01 price used at the time of reporting.
The bank’s core bull case rests on Nvidia becoming more than a chip supplier.
Arya and his team argue that the company continues to secure virtually all scarce inputs required for the AI infrastructure buildout.
That includes everything from chips and land to energy and data-center shells, among other things. Nvidia’s healthy GPU rental rates and tremendous cash generation make that strategy attractive.
It does, however, create a valuation problem. If Nvidia continues to finance its own ecosystem, it becomes incredibly tough for investors to value each dollar of free cash flow equally.
Consequently, BofA splits its cash generation into two.
Around 50% is treated as stockholder cash, returned through share buybacks and dividends, while the rest is assumed to support ecosystem investments and financing.
BofA analysts argue that even if it heavily discounts the cash Nvidia might use to support its AI partners, the stock’s still cheap.
Let’s look at a simple example to understand the argument better.
Imagine Nvidia generates roughly $100 of free cash flow:
- $50 of that accrues to shareholders through buybacks/dividends, which BofA normally values at a 48x peer multiple.
- $50 used for AI investments/backstops, as they carry a lot more risk; BofA is valuing them at just 24 times.
Even if investors give that riskier portion 50% less valuation credit, it yields a valuation of 36 times 2027 free cash flow. Nvidia currently trades around 18x, which suggests the market is pricing in far too much risk.
Though Nvidia has committed an enormous $300 billion to AI investments and backstops, the bank expects nearly $470 billion in free cash flow across 2026 and 2027, pointing to a ton of financial firepower.
BofA sees better economics behind Nvidia’s OpenAI backstop
Perhaps the most striking part is the economics behind Nvidia’s massive $105 billion OpenAI guarantee.
Though on paper it looks massive, BofA stresses that Nvidia isn’t guaranteeing the entire loan to the ChatGPT makers. In fact, it is backstopping the residual value linked to leases at SB Energy’s Ohio data center, limiting its losses if the project runs into trouble.
Moreover, the potential sales associated with that guarantee are huge.
According to BofA, the initial 4.25 gigawatts Nvidia is backing will entail roughly 1.5 million GPUs, generating $150 billion to $200 billion of Nvidia revenue per hardware generation.
At a 50% free-cash-flow margin, that translates into $75 billion to $100 billion in FCF, which approaches the size of the $105 billion guarantee.
According to BofA, this relationship might represent $600 billion of Nvidia sales, $300 billion of FCF through 2030, and 12 to 16 gigawatts of compute, against a $105 billion backstop.
Another major protection is that Nvidia’s GPUs can be effectively moved and reused elsewhere, with CUDA making the chips useful across a wide range of customers.
What Nvidia investors should watch before earnings
For Nvidia stock investors, BofA’s take switches up what they should monitor, but it doesn’t break the bull case.
The bank believes investors are applying a substantial discount for the potential that Nvidia’s ballooning commitments to its AI customers will lead to balance-sheet problems.
That makes Nvidia’s August 26 fiscal Q2 2027 earnings report doubly important. Wall Street is expecting a normalized EPS of $2.08, a GAAP EPS of $2.06, and a revenue of $91.96 billion, according to Seeking Alpha data.
For perspective, that marks a massive step up from fiscal Q2 2026, when Nvidia posted normalized EPS of $1.05 on sales of $46.74 billion, representing 55.6% year-over-year sales growth.
Also, expectations have been rising, with 32 positive EPS revisions over the past 90 days compared to just 3 cuts ahead of the print.
Related: 5-star analyst sets alarming SpaceX stock price target
Bank of America is expected to soar past Q2 estimates by $3 billion to $4 billion, with sales of nearly $94 billion to $95 billion, guiding Q3 sales toward $107 billion to $108 billion, above sell-side expectations.
That said, for investors, it’s imperative to listen closely for disclosures around guarantees, residual-value commitments, and off-balance-sheet exposure.
Arya expects management to provide much more detail on those commitments, as well as on gross margin resilience, AI model economics, and demand for Vera Rubin.
Nonetheless, the bullish evidence is still incredibly strong. BofA points to powerful GPU rental rates, constricted compute availability, and Nvidia’s tremendous free-cash-flow generation as evidence for strong underlying AI demand.
Interestingly, despite memory-cost inflation, the bank still sees margins holding around 73%-74%, viewing financing and memory concerns as overblown.
For perspective, according to Seeking Alpha data, Nvidia’s levered free cash flow has shot up from $4.5 billion in fiscal 2023 to $58.1 billion in fiscal 2026, a whopping 1,182% increase over three years.
For existing stockholders, that goes against selling mainly because financing headlines look alarming. For new buyers, earnings offer a better test than simply chasing the upside case laid out by BofA.
The warning sign is obvious: if Nvidia’s financial commitments continue to outpace AI demand, it would put substantial pressure on both growth and its fortress-like balance sheet.
From a technical standpoint, Nvidia’s near-term setup still looks mostly constructive according to Barchart data.
The stock has held up remarkably above its 20-, 50-, 100-, and 200-day moving averages, underscoring a relatively strong underlying uptrend.
Momentum is getting stretched somewhat, though: the 14- and 20-day stochastic readings are near 89%, currently in overbought territory, while raising the odds of consolidation or a pullback. Moreover, relative strength around 56-57 remains mostly healthy rather than extreme.
However, with ATR near 3% (average daily price volatility), investors should be ready to stomach the elevated day-to-day swings heading into earnings.