Nvidia withstands AI credit jitters as Broadcom, Oracle shares slide

Rising borrowing risks across the artificial intelligence industry are weighing on valuations, but stronger earnings forecasts have helped Nvidia outperform its peers

Nvidia has proved more resilient than Broadcom and Oracle as investors grow increasingly concerned about the debt and financing costs associated with the artificial intelligence boom. Although credit-risk indicators have deteriorated across all three companies, Nvidia’s stronger earnings outlook has helped support its share price while its competitors have suffered steeper declines.

In an opinion article published by MarketWatch on 10 October, Michael Kramer, founder of Mott Capital Management, examined the divergence between equity valuations and credit markets. His analysis found that rising credit-default swap (CDS) spreads and contracting price-to-earnings (P/E) ratios have affected Nvidia, Broadcom and Oracle, but differences in expected earnings growth have helped explain their contrasting stock-market performance.

The figures cited in Kramer’s analysis cover the period from 2 June to 8 October 2026 and suggest that investors are becoming more selective about which companies they believe can convert heavy AI investment into sustainable earnings.

Credit-default swaps are financial contracts that protect a borrower failing to meet its debt obligations. When CDS spreads widen, the cost of that protection rises, generally indicating that investors perceive greater credit risk. A widening spread does not, however, mean that a company is about to default.

According to Kramer’s analysis of LSEG data, Nvidia’s five-year CDS spread increased from approximately 40 basis points on 2 June to 84 basis points on 8 October. Over the same period, Broadcom’s spread more than tripled, rising from 41 to about 132 basis points, while Oracle’s climbed from roughly 151 to 252 basis points.

The movements reflect concerns about the financing requirements of the AI industry, where companies are committing substantial sums to chips, data centres and computing infrastructure. The scale of investment has raised questions about whether future revenue and cash flow will justify the costs.

Oracle and other large cloud-computing companies are expanding their infrastructure to accommodate AI workloads. Broadcom supplies semiconductor and networking technology used in AI systems, making its prospects partly dependent on continued investment by major technology customers.

Nvidia occupies a different position in this ecosystem as a leading supplier of AI accelerators. Its sales can benefit from increased spending on AI infrastructure, but the company is not immune to the possibility that customers may eventually slow their investment.

The clearest difference in Kramer’s analysis is the pace at which analysts have raised their earnings forecasts.

Between 2 June and 8 October, analysts’ rolling 12-month earnings-per-share estimates for Nvidia increased from $10.48 to $14.18, a rise of more than 30%. Broadcom’s estimates rose from $15.83 to $19.23, an increase of just over 20%, while Oracle’s climbed from $8.24 to $9.36, or approximately 14%.

These estimates help explain why falling valuation multiples have not translated into the same share-price outcome for all three companies. When expected earnings rise faster than a share price, a stock’s forward P/E ratio can decline even if its market price remains relatively stable.

Nvidia’s forward P/E ratio fell from about 21 to 16 over the period. Yet the increase in projected earnings helped offset the effect of that lower valuation multiple, allowing its shares to hold up better than those of Broadcom and Oracle.

Broadcom’s forward P/E ratio contracted from around 30 to 18.7. Its shares fell approximately 25% over the period. Oracle experienced a steeper adjustment: its forward P/E ratio dropped from nearly 30 to 14.5, while its shares declined by more than 40%, according to the figures cited by Kramer.

The comparison suggests that markets are not treating all AI-related companies equally. Investors appear to be weighing not only the cost of financing expansion but also the expected pace of earnings growth and the ability of each business to generate returns from its investment.

The credit-market pressure raises a wider question about the sustainability of the current AI investment cycle. Major technology companies are spending heavily on data centres, computing capacity and specialised chips, while some are increasingly turning to debt and external financing to fund expansion.

Reuters reported on 8 October that large companies were raising substantial sums in capital markets to finance AI infrastructure, adding to concerns that corporate borrowing linked to the technology boom is becoming a significant force in global finance.

The potential risk extends throughout the supply chain. If cloud providers or other major AI customers reduce capital expenditure, chipmakers and equipment suppliers could face weaker orders. Companies whose valuations depend on continued rapid growth may also face pressure if earnings forecasts are revised downwards.

For Nvidia, rising earnings expectations currently provide a buffer, but they do not eliminate exposure to a slowdown in AI investment. Broadcom and Oracle face similar uncertainties, although their business models, capital requirements and sources of revenue differ.

Kramer identifies two possible paths for the market. If credit concerns ease while earnings forecasts remain strong, valuation multiples could recover. If analysts begin cutting forward earnings estimates, however, companies already experiencing lower P/E ratios could face additional pressure.

Neither outcome is guaranteed. CDS spreads are indicators of perceived credit risk rather than direct forecasts of share-price movements, while earnings estimates can change as demand, financing conditions and investment plans evolve.

The divergence between Nvidia, Broadcom and Oracle therefore offers a broader lesson about the AI boom: rising demand for the technology does not necessarily translate into identical returns for every company involved. Investors are increasingly distinguishing between businesses that can support their valuations through earnings growth and those whose investment commitments may place greater pressure on cash flow and financing.

Sources: Michael Kramer, MarketWatch, “Why Nvidia’s stock is dodging the AI credit scare that is crushing Broadcom and Oracle” (10 October 2026); LSEG market data cited in Kramer’s analysis; Reuters, “Morning Bid: Sovereign bonds shouldered aside as AI takes their turf” (8 October 2026).