ZEINAB GABR

OpenAI, with an estimated $20 Billion annual revenue by December 2025, has made an astonishing $1.15 TRILLION in infrastructure spending commitments over the next five years.

This isn't just growth, it's a financial black hole.

The Unjustifiable Scale of Spending

The breakdown of the commitments is staggering:

- Nvidia: up to $100B (investment in OpenAI)

- AMD: $90B (GPU purchase, costing AMD $40B in stock warrants)

- Broadcom: $350B (custom AI chips)

- Oracle: $300B

- Microsoft: $250B

- Amazon & others: $50B

Total: $1.15 TRILLION.

To put that in perspective: the annual Capital Expenditure (CapEx) of ALL US companies combined is about $1.2T.

OpenAI's 5-year commitment is nearly 100% of the combined US annual CapEx. They are committing to outspend their major partners (MSFT, NVDA, AMD, ORCL, AMZN, etc.) combined CapEx by nearly 5x.

The Math Does Not Add Up

To cover the operating costs and maintain projected gross margins, OpenAI's revenue must grow from $12B in 2025 to nearly $983B in 2030. That is an 85x growth in just 5 years.

To meet its obligations, OpenAI essentially needs to become the largest, most profitable company on Earth.

And if the math doesn't work? OpenAI's CFO has openly suggested seeking a GOVERNMENT BACKSTOP (taxpayer protection) in case they can't meet their obligations. The terrifying implication: Too Big To Fail?

The Interconnected 'Ouroboros' Loop

The core issue is the financial engineering. The money is flowing in an "Ouroboros" (snake eating its tail) loop, making real revenue and cash flow untraceable:

1. Microsoft invests in OpenAI (via Azure credits).

2. OpenAI buys Microsoft capacity.

3. Microsoft buys Nvidia GPUs to provide that capacity.

4. Nvidia takes the cash and invests in OpenAI.

5. OpenAI uses Nvidia's investment to buy Nvidia chips.

6. The loop continues, creating 'fake' revenue flows.

This web of interconnected deals means: "The revenue of one company is the cost of another is the investment of a third." If one key pillar in this $1.1T structure fails, the ripple effect (contagion) could be massive.

Is This 2000 or 2008?

Compared to the GFC (2008): The risk is different. OpenAI's obligations are not being securitized and multiplied 15x by exotic derivatives. Big Tech has strong balance sheets.

Compared to the Dot-Com Bubble (2000): Major players today (MSFT, NVDA) are profitable. The Nasdaq P/E is 30x, not the peak of 60x.

The problem is the mix, as Charlie Munger famously said: "If you mix raisins with turds, well, you still got turds."

The AI bubble has strong "raisins" (real, profitable businesses) surrounded by "turds" (trillions in inter-company fake revenue, impossible spending commitments, and whispers of taxpayer bailouts).

The spending is insane, and the sheer amount of financial engineering is raising red flags across the board.

The Question: Can any economy sustain $1.1T in interconnected spend from a company with $20B revenue?