OpenAI likely needs ~$100 Billion+ in revenue by ~2029 to avoid default on their obligations.

• OpenAI committed to spend $300 Billion on datacenter capacity from Oracle over a 5 year period, starting in 2027.

• While exact terms of the deal are not public, Oracle's revenue forecasts and RPO(most of the RPO growth is OpenAI), as well as common sense would suggest that the contracted spend ramps up over the 5 year period.

• If it doesn't ramp, splitting it evenly over 5 years would still be $60 Billion a year

• Microsoft is entitled to a 20% cut of OpenAI's revenue until their capped profit threshhold is reached. Estimates put this cap at $130 Billion.

• ChatGPT already has 770 Million weekly users; most of their growth will need to come from enterprise implementations, which generally use the API as part of integrations with other software tools.

• OpenAI already has a contract with Microsoft to use Azure exclusively for their API through 2030. This means that even if OpenAI sees significant enterprise growth via AI integrations, they have to pay Microsoft for this capacity, they cannot use the capacity they paid Oracle for.

• With fierce competition from X.AI, Anthropic, Gemini, etc, OpenAI has limited ability to raise prices without losing market share.

• OpenAI has a $10 Billion commitment to Broadcom for custom chips in 2026.

• OpenAI faces challenges retaining top talent, with Meta offering $250 Million pay packages to poach top researchers.

• OpenAI is already burning $8 Billion a year even before all of these new obligations.

Despite operating at a loss, OpenAI will likely be able to raise some funding to close the gap. But there are risks they face:

• Microsoft has been unwilling to adjust their past agreements(they are very lucrative), putting OpenAI's for profit conversion at risk, and jeopardizing a Softbank funding round(which is contingent on for profit conversion)

• Right now OpenAI has a $500 Billion valuation based on the employee share sale.

• Investors will want to believe their funding will keep the company afloat, and that the company is growing enough to justify the valuation. For example, if OpenAI is at a $50 Billion annual burn rate and isn't doubling their revenue every year, they will struggle to raise $50 Billion to close the gap.

• The largest equity raise in history was Saudi Aramco at ~$25 Billion. The largest bond sale, Verizon at $49 Billion(which had a lot of assets that they could back the bonds with and strong cash flows).

• Investors are going to be very nervous about these huge obligations. Softbank is a big outlier that is known for highly speculative bets like WeWork. But they only have so much capital they can deploy.

• Realistically, it is highly unlikely that there is sufficient liquidity for OpenAI to raise more than $50 Billion via equity a year every year.

• Term loans or revolving credit facilities, if even attainable, would be very high interest rate and limited capacity given their high obligations and negative cash flows.

The concern here is that such an event would have a significant impact on most of big Tech:

• Oracle with their $300 Billion in revenue forecasts from OpenAI that might not materialize

• Microsoft with their lucrative contracts with openAI for API exclusivity, rev share, etc.

• Google likely losing OpenAI as a cloud customer, because they will have plenty of capacity from Oracle/Microsoft they are obligated to buy.

• Nvidia/Broadcom when orders of hardware fail to grow as expected

• Amazon if AWS faces demand pressure due to Oracle needing to sell their capacity to other buyers

• Google/Meta if Openai starts selling ads for users on free plans as a desperation play to raise revenue.