The announcement of a strategic alliance between Nvidia and OpenAI marks a new milestone in the race for leadership in artificial intelligence. Nvidia has committed to investing up to $100 billion to finance the deployment of at least 10 GW of capacity based on its GPUs, using the new Vera Rubin architecture, which will replace Grace Blackwell as the technological spearhead. The first phase will enter operation in the second half of 2026.
This collaboration adds to other significant initiatives by both companies: OpenAI has already worked with Microsoft, Oracle, and the Stargate consortium, while Nvidia has intensified its strategic alliances, including investments in Intel and agreements with players in France and Saudi Arabia.
From a financial standpoint, although the contract details are still unknown, the construction of a 1 GW data center implies GPU investments in the tens of billions of dollars, which could be reflected in Nvidia’s results in the coming years.
Preference for GPUs and Signals in Market Sentiment
This agreement, together with the previous deal between OpenAI and Broadcom, reinforces the perception that GPUs continue to be preferred over ASICs for inference tasks, even once models are trained. However, Monday’s stock market reaction—a moderate drop in Nvidia shares—reveals that investors are beginning to interpret this news as subtle warnings.
Behind this skepticism lies the business structure: Nvidia has been investing in cloud-based startups that rely on its GPUs. This “circular relationship,” where the provider funds its own customers, now shows a scale that unsettles some analysts.
Valuations and Systemic Risks: Is History Repeating Itself?
Although current valuations of big tech companies have not yet reached the levels of the dot-com bubble or the Nifty Fifty of the ’70s, they remain demanding. The recent rally in fixed income has allowed investors to maintain long positions in these companies without discomfort, as it has not negatively affected multiples. But the environment warrants attention.
One of the key risk areas lies in the review of tariffs enacted under IEEPA (International Emergency Economic Powers Act). According to The Washington Post, institutional investors are buying rights to legal claims for the payment of these tariffs at $0.20 per dollar. This anticipates a move: if the Supreme Court rules that the charges were illegal, the Treasury could be forced to reimburse up to $130 billion.
Impact of IEEPA: Two Paths to Instability
If the Court overturns the tariffs enacted under IEEPA, the impact on the markets could come through two channels:
Volatility and Regulatory Uncertainty
The removal of the current framework would open a period of uncertainty. Although most of the affected countries have already accepted the new tariffs—which rose from 2.5% in 2023 to nearly 19%, according to the Yale Budget Lab—the Trump administration would resort to other mechanisms such as Section 232. In fact, it has already announced tariffs of 25% on heavy trucks and 30% on upholstered furniture. Restrictions on semiconductors manufactured outside the United States are also under consideration, which would force companies to double their domestic capacity.
Significant Fiscal Loss
The Treasury would lose about $500 million in daily revenue—funds that markets already factor in as partial offsets to the fiscal deficit generated by the OBBA plan. The immediate effect would be a rise in long-term bond yields, which would compress equity valuations in the short term.
Yields, Correlations, and Sensitivities
In this context, the correlation between fixed income and equities is at peak levels. This means that any upward movement in interest rates has greater potential to negatively impact stock prices, especially in an environment of elevated multiples.
The market continues to price in an accommodative Fed: the yield curve projects a terminal rate below 3% by December 2026. This has limited the downside in the 10-year bond yield, currently anchored around 4%.
The key to rate direction lies in labor data. Differences within the FOMC over whether to implement one or two additional cuts before year-end will be resolved—if inflation expectations remain stable—based on employment trends.
Labor Outlook: Mixed Data, Mixed Reactions
The decline in immigration and the slow normalization of the labor market after the pandemic complicate projections. This structural disruption makes it difficult to apply historical models reliably.
The market is particularly sensitive to any data that could disrupt the current balance. Statistics such as those released this Thursday will be key, as they could trigger volatility in the short end of the curve and define its steepening.
A deterioration in the labor market, if interpreted by the Fed as structural rather than cyclical, could prompt more aggressive cuts in the short term. This scenario would benefit tech stocks, which have historically thrived in environments with negative real rates and moderate but sustained growth expectations.
