Nvidia is drawing Wall Street into the expansion of AI. Its stock prices dropped following the announcement.

      A consortium of financial firms is collaborating with Nvidia on a $500 billion funding package aimed at AI infrastructure. This group includes Apollo Global Management, Blackstone, BlackRock’s Global Infrastructure Partners, Brookfield Asset Management, Goldman Sachs, and KKR. The Financial Times was the first to report on the discussions, which Bloomberg later confirmed with sources familiar with the situation. A potential deal might be finalized as soon as Monday.

      The market reacted negatively to this news, with Nvidia's shares dropping by as much as 3.2% on Monday. By mid-afternoon in New York, they were trading at $219.01, reflecting a decline of 2.2% for the day.

      However, there is a significant caveat. The specifics of the funding package remain uncertain. Bloomberg's sources could not specify which projects or companies would be supported or the structure of the funding. They also could not confirm whether the $500 billion represents entirely new capital.

      This last point is particularly important, as Nvidia has already announced numerous commitments worth hundreds of billions across the AI supply chain this year. A package that simply consolidates existing commitments would be significantly different from one that introduces new funds.

      BlackRock and KKR declined to comment to Reuters, and Nvidia, along with the other firms, did not provide an immediate response.

      Nvidia has largely self-financed its growth this year, having engaged in discussions for a $250 billion backup for OpenAI to lease computing resources at a 10-gigawatt campus in Ohio, which is being developed by SoftBank subsidiary SB Energy. Additionally, it has explored financing terms for approximately $350 billion related to OpenAI's chip acquisitions and has expanded its partnership with South Korea’s SK Group to over $500 billion in mutual investments. Nvidia has also made a significant investment in Ilya Sutskever’s Safe Superintelligence.

      Overall, this pattern reveals a consistency: Nvidia guarantees the customer, the customer purchases the chips, and the revenue returns to Nvidia’s financial statements.

      The $500 billion deal's impact on share prices is based on a reasoning that investors find unappealing. This arrangement is referred to as circular financing, where a supplier financially supports a buyer, who in turn funds the supplier. Critics argue that this can artificially inflate demand and valuations within a sector until something collapses.

      Throughout the year, Nvidia has faced this critique. When it announced $750 billion in deals earlier in 2026, it did not receive a welcoming response from the credit market.

      To counter these criticisms, involving six external balance sheets can help dilute the financial burden beyond just the chipmakers. This also positions independent underwriters between Nvidia and the projects in question. However, this approach is only effective if the underwriting is genuine. Private credit and infrastructure funds are now more closely related to the AI trades than they have been in this cycle, with the BIS already noting similarities to credit structures prior to 2008.

      It's worth mentioning that these firms are not newcomers to the sector. Both Apollo and Blackstone constructed a $35 billion vehicle around Google TPUs. Morgan Stanley facilitated a $917 million loan secured by Lambda’s Nvidia GPUs. In September 2024, BlackRock, Global Infrastructure Partners, Microsoft, and MGX initiated the AI Infrastructure Partnership, targeting $30 billion in equity and potentially up to $100 billion including debt. Nvidia and xAI joined this partnership in March 2025, with Nvidia acting as a technical adviser rather than a capital contributor.

      At that time, Jensen Huang described this partnership broadly, stating that the global expansion of AI infrastructure would “benefit every company and country that seeks to achieve economic growth and address the world’s greatest challenges.”

      A $500 billion package would be about five times the size of that earlier initiative and would bring Nvidia significantly closer to the financing.

      The key figure to focus on is not the $500 billion. Major tech companies are expected to spend over $730 billion on AI this year. Nvidia itself returned to the U.S. bond market in June for its first debt issuance since 2021.

      In this context, a half-trillion dollars in arranged funding is substantial but not unrealistic. The critical issue, however, is more specific and complex: if the data centers are constructed but demand does not materialize, losses will be incurred by someone. Currently, it remains unclear who that party might be—whether it will be Nvidia, a pension fund, or a private credit investor who believed they were investing in infrastructure.