Wall Street and NVIDIA Corp. are entangled in an escalating debate over the long-term value of artificial intelligence (AI) hardware, as lenders push for stronger financial guarantees on the semiconductor giant’s ambitious $500 billion financing initiative.

Institutional investors are signaling that NVIDIA’s advanced graphics processing units (GPUs) and related data center infrastructure may not hold the durable collateral value the chipmaker claims, according to banking sources and credit managers contacted for a Reuters report.

The pushback highlights a growing rift between Silicon Valley’s optimistic hardware valuations and Wall Street’s traditional credit metrics, threatening to complicate future funding rounds for AI developers seeking capital.

The dispute centers on NVIDIA’s strategy, announced in August alongside private equity titans Blackstone, Apollo, and KKR, to transform its computing power into an “investable infrastructure asset.”

The plan aims to mirror the aircraft leasing market, allowing AI developers to access chip power using GPU-backed loans while attracting independent institutional capital.

However, credit managers are questioning the assumption that these specialized chips can maintain commercial earning power over an extended period. NVIDIA CEO Jensen Huang has argued that the company’s GPUs have a useful life of up to a decade, yet asset managers remain skeptical about relying on fast-depreciating technology as core collateral.

“Wall Street is much more conservative,” said Tony Trzcinka, senior portfolio manager at Impax Asset Management, referencing NVIDIA’s 10-year revenue claims.

Lenders are particularly uneasy with NVIDIA’s proposed guarantee structure.

The company initially outlined plans where some deals would feature a residual value guarantee of no more than 25% — a measure intended to mitigate concerns regarding “circular financing,” where a vendor finance purchases of its own hardware. Bankers contend this level of protection is insufficient compared to recent AI infrastructure transactions.

To mitigate risk, financiers are asking NVIDIA for broader corporate backstops or requiring loans to be secured by long-term revenue commitments from investment-grade technology firms.

“NVIDIA would imply that the GPUs work well north of five years, and that actually has been proven to be true thus far,” said Andrew Chang, a director at S&P Global Ratings. “Yet we take a conservative view of the value of those chips.”

Despite the friction, market appetite for AI infrastructure debt remains robust. Deals currently in the pipeline, totaling tens of billions of dollars, are expected to feature restructured terms offering stronger credit enhancements. These upcoming transactions will likely blend chip collateral with direct customer contracts and backing from NVIDIA to satisfy investor demands for certainty.

An NVIDIA spokesperson defended the durable financial value of its hardware, noting that AI compute serves as a “productive, durable and fungible asset that can support long-term financing.”

The company added that independent financing partners continue to evaluate opportunities based on cash flow projections, customer commitments, and residual values, noting that “financing structures will vary as this market develops.”

The heightened scrutiny reflects broader caution among credit rating agencies and private capital lenders as hundreds of billions of dollars flow into data centers, power grids, and specialized hardware.

Analysts at Morningstar noted that complex financing structures, including vendor funding and private credit arrangements, are drawing increased comparisons to the speculative dealmaking of the dot-com era.

For now, financial institutions maintain that NVIDIA’s compute power cannot yet be underwritten like traditional physical assets such as commercial aircraft. As the market matures, Wall Street appears determined to enforce stricter underwriting standards before fully embracing AI hardware as a standardized credit asset class.