The $105 Billion Loop: Why Nvidia and OpenAI Can’t Stop Financing Their Own Demand

Posted on by

(SeaPRwire) –   By: Oliver Hawthorne

Nvidia is guaranteeing $105 billion for OpenAI’s Ohio data center. This figure arrived after two rounds of cuts, dropping from $250 billion. The market reacted instantly. Nvidia shares fell 4.5% when reports of the larger guarantee leaked in July. Investors understand what the companies do not say outright. The AI industry is now financing the infrastructure that creates its own demand.

The facts on record are precise. Nvidia disclosed an aggregate payment obligation capped at $105 billion in an SEC filing tied to Monday’s announcement. The Wall Street Journal first reported the August 14 cut to less than $120 billion, then confirmed the final settlement at $105 billion. The campus will eventually reach 8 gigawatts of computing capacity. Each gigawatt requires roughly $50 billion to $60 billion in total spending, according to Nvidia’s investor materials. OpenAI will lease from SB Energy for up to 20 years. Nvidia invested $1.5 billion in SB Energy itself. The first 800 megawatts are expected to become available in 2028, largely using existing AEP infrastructure. Further development will require new power plants and new transmission lines.

The commercial loop is what matters. Nvidia is simultaneously the world’s dominant AI-chip maker and a financier of the infrastructure that purchases its chips. The company partnered with six major financial institutions last week to launch compute-financing platforms targeting more than $500 billion in third-party funding. SEC staff guidance issued in July concluded that certain data-center debt falls outside Dodd-Frank securitization rules, making it easier to mobilize outside capital rather than carry the exposure on Nvidia’s own books. This is not new. Nvidia has invested in AI companies and data center operators for years. The Ohio deal simply makes the cycle visible at scale. Jensen Huang described the arrangement as securing long-lived infrastructure for Nvidia compute, enabling OpenAI to deploy productive AI factories upgradeable with each new generation. What he did not describe is who pays when the models do not generate enough revenue.

The fundamental question is whether this ecosystem can produce outside revenue. OpenAI must reinvest future revenue to fund its buildout. Microsoft, Oracle, and SoftBank are part of the same web of partnerships. The circular financing structure reduces Nvidia’s direct financial exposure, but it also means the entire supply chain depends on revenue that has yet to materialize at the scale required. The market already priced in discomfort. The $105 billion figure is lower than the $250 billion alternative. That reduction signals risk recognition, not confidence. The Ohio campus sits on a former uranium-enrichment site. Approximately 9.2 gigawatts of natural-gas generation is planned to support it. SoftBank and SB Energy are expected to invest billions more in regional power infrastructure. All of this capital is committed. The revenue stream it depends on remains unproven. Nvidia and OpenAI will lease, invest, and guarantee. The question is whether anyone outside the loop will ultimately foot the bill.

Author bio: Oliver Hawthorne is a Principal Correspondent permanently stationed at an international technology review, covering semiconductor markets and AI infrastructure investment.