A bullish call after a decisive quarter

Bank of America maintained its Buy rating on Nvidia and a $350 price target in an August 25 research note, published before the chip designer released its latest quarterly results. The call was notable not simply because it anticipated another strong earnings report, but because it placed balance-sheet commitments at the centre of the investment debate.

Nvidia’s results on August 26 gave the bullish case further operational support. Revenue for the second quarter of fiscal 2027, ended July 26, reached $96.2 billion, an 18% sequential increase and 106% rise from a year earlier. Data Center revenue, Nvidia’s principal AI-related business, was $89.0 billion, up 117% year on year. GAAP net income rose 126% to $59.7 billion, while gross margin was 75%.

Management also forecast third-quarter revenue of $108 billion, plus or minus 2%, while stating that this outlook assumes no Data Center compute revenue from China. The guidance indicates that Nvidia expects demand for AI infrastructure to remain unusually strong even without contribution from that market.

For Bank of America, however, the immediate earnings beat was only part of the story. Its central argument is that investors have attached a substantial risk discount to Nvidia shares because the company is becoming more involved in funding the infrastructure and customers that support demand for its own systems.

From component supplier to ecosystem financier

Nvidia has traditionally been valued principally as a supplier of high-performance processors, networking equipment and software. That model generates significant cash when customers purchase systems outright. The emerging concern is that the company is also taking financial exposure to projects and businesses that will deploy its technology.

Bank of America estimated Nvidia’s capital commitments at about $300 billion, including direct equity investments alongside guarantees, residual-value arrangements and other forms of support. These are analyst estimates rather than a single Nvidia-reported balance-sheet line item, and the different commitments do not carry identical risks. An equity investment can lose value; a guarantee or backstop may create an obligation only under specified circumstances.

Still, the direction of travel is clear. Nvidia disclosed in August that it would provide credit support related to land, power and building capacity at SB Energy’s PORTS-Pike technology campus in Ohio. OpenAI is intended to be the customer for the planned 8-gigawatt facility, while Nvidia said it would invest $1.5 billion in SB Energy. The project illustrates how AI infrastructure increasingly depends on securing power, sites and construction capacity, not merely on procuring chips.

This can be strategically rational. Large AI campuses require multi-year coordination across utilities, data-centre developers, cloud providers, financiers and model developers. Nvidia’s financial support may help remove bottlenecks and accelerate deployments of its computing platforms. It may also create a deeper ecosystem around its hardware and software.

But the approach complicates the valuation question. If a company supports customers or projects that subsequently buy its products, investors may ask how much reported demand reflects independent end-user economics and how much is enabled by vendor-linked financing. The issue is not necessarily that the revenue is invalid. Rather, it is whether investors should apply the same valuation to every dollar of cash flow when a portion may be needed to support future capacity commitments.

Why Bank of America sees the risk as manageable

Bank of America’s argument is based on scale and cash generation. The firm estimated that Nvidia could generate roughly $469 billion in free cash flow across calendar years 2026 and 2027. On that view, substantial commitments are more manageable than the headline totals initially suggest, especially if AI demand remains robust and the related assets retain value or can be deployed elsewhere.

Nvidia’s quarterly financial performance supports the premise that it has considerable capacity to fund investment and shareholder returns at the same time. During the latest quarter, it returned approximately $26 billion through share repurchases and dividends. It had about $99 billion remaining under its share-repurchase authorisation at quarter-end.

The bank has suggested that a larger proportion of future free cash flow could be returned to shareholders. That could matter for the valuation case independently of revenue growth. Buybacks reduce the share count over time and may provide investors with clearer evidence that Nvidia’s cash generation is not wholly being recycled into the AI ecosystem.

The $350 price target should nevertheless be viewed as an analyst opinion, not a forecast with certainty. Its implied upside was calculated from Nvidia’s share price when the note was issued on August 25; it will change as the market price changes. Price targets also depend on assumptions about growth, margins, capital intensity and the valuation multiple investors are willing to pay.

The issues investors will need to follow

Nvidia’s results demonstrate that current AI infrastructure spending remains powerful. The company’s third-quarter outlook and 75% gross margin suggest that demand has so far outweighed concerns over costs and execution. Yet the balance-sheet debate is unlikely to disappear after one quarter.

Several factors will determine whether Bank of America’s assessment proves persuasive:

  • The extent and terms of future guarantees, investment commitments and purchase arrangements.
  • Whether funded projects attract durable end-customer demand rather than relying on continuing financial support.
  • The pace at which data-centre capacity, power generation and networking can be delivered.
  • Nvidia’s ability to sustain margins as infrastructure investment broadens and competitors pursue more of the AI computing market.
  • The balance it strikes between ecosystem investment, acquisitions, dividends and repurchases.

The key analytical distinction is between productive strategic financing and financial exposure that obscures underlying demand. In the favourable scenario, Nvidia uses a portion of its financial strength to solve infrastructure constraints, expands the market for accelerated computing and preserves its technological leadership. In the less favourable scenario, commitments become costly if AI projects are delayed, customers face funding pressure or equipment utilisation falls short of expectations.

For now, Bank of America is taking the first view: Nvidia’s earnings power and free-cash-flow potential outweigh the risks embedded in the commitments. Nvidia’s August 26 results strengthen that case, but the quality, scale and transparency of the company’s future financing arrangements will be as important as the next revenue figure.

Sources