Clever accounting in the AI race

In 1999, Lucent Technologies had a problem. Demand for telecom equipment was slowing, but Wall Street expected growth. So Lucent invented demand. It lent billions to startup carriers who used the money to buy Lucent gear. Revenue soared. The stock followed. Then the carriers defaulted because there were no actual customers using all that fiber. Lucent wrote off $7 billion. Cisco lost 89 percent of its value. The Nasdaq fell 78 percent.

Nvidia studied that playbook and fixed the flaw.

The flaw was debt. Loans can default. Equity cannot. So Nvidia invests in its customers instead of lending to them. It took a stake in CoreWeave, which uses the capital to buy Nvidia chips. It committed $100 billion to OpenAI, which will use it to buy Nvidia chips. When the strategy works, Nvidia books the revenue. When it doesn't, Nvidia owns equity in companies that own Nvidia hardware. There is no failure mode that doesn't benefit Nvidia.

But the real innovation is subtler. Nvidia signed an agreement to purchase any unsold CoreWeave capacity through 2032. Read that carefully. If demand for AI compute disappears tomorrow, Nvidia has contractually guaranteed to buy its own chips back, laundered through a cloud provider. The company has eliminated demand risk by becoming the buyer of last resort for its own products.

This is not financing. It is the financial equivalent of a perpetual motion machine. Money leaves Nvidia's balance sheet as "investment," passes through CoreWeave or OpenAI, and returns as "revenue." The chips move. The cash circles. Wall Street sees growth.

And now the model is mutating. Last week OpenAI announced an equity stake in Thrive Holdings, a firm created by Thrive Capital, one of OpenAI's largest investors. The purpose is to embed OpenAI into Thrive's portfolio companies.

This is the next logical step. When you've perfected buying your own chips, you start buying your own customers. Thrive invests in OpenAI. OpenAI invests in Thrive's subsidiary. That subsidiary's companies adopt OpenAI. Adoption metrics rise. Valuation follows. No independent decision was made anywhere in the chain.

This is not fraud. It is not illegal. It is also not demand in any meaningful sense.

The unsettling question is not whether AI is a bubble. It is whether we can even know. Markets require feedback. Customers decide if something is worth buying. That signal propagates. When supplier, customer, and investor collapse into the same rotating cast, the feedback loop breaks. The measurement system has been captured by the people being measured.

Lucent's error was crude: debt that could default, losses that were visible. The modern architecture leaves no trace. Equity instead of loans. Ownership instead of exposure. Guaranteed buybacks instead of uncertain demand. When you own the customer, there is no default. There is only a company using your product exactly as designed.

That is the innovation here. Not in chips. In the accounting for demand.

Author

Ai Base Network (ABN), ABN ASIA was founded by people with deep roots in academia, with work experience in the US, Holland, Hungary, Japan, South Korea, Singapore, and Vietnam. ABN Asia is where academia and technology meet opportunity. With our cutting-edge solutions and competent software development services, we're helping businesses level up and take on the global scene. Our commitment: Faster. Better. More reliable. In most cases: Cheaper as well.

Feel free to reach out to us whenever you require IT services, digital consulting, off-the-shelf software solutions, or if you'd like to send us requests for proposals (RFPs). You can contact us at [email protected]. We're ready to assist you with all your technology needs.