OpenAI: Can OpenAI Deliver on Its $1.4 Trillion CapEx Commitment?

Over the past two months, we’ve seen way too much AI bubble talk, and the market has clearly been pricing in AI bubble concerns during this period.

You can see it clearly in stock prices - names with exposure to AI TAM or AI semi TAM like NVDA, TSM, AVGO have been flat for 2-3 months. Yet from the supply chain side, we hear about order increases across the board almost monthly. Just this week, NVDA added more optical module orders.

Meanwhile, weaker segments like NeoCloud are getting hit even harder by the AI bubble narrative. Since NeoCloud takes on extension risk for AI semis and CSPs, falling stock prices could also reduce their financing capacity, creating a downward spiral. Going long NeoCloud CDS has become the Street’s best idea - and long CDS positions are triggering quant funds’ sentiment indicators to further reduce positions or short NeoCloud.

Look at ORCL - it’s already back to pre-OpenAI $1.4 trillion announcement levels.

In contrast, names without AI TAM exposure like storage, TPU, AMD are performing well. Their typical characteristic is gaining share within overall AI CapEx, growing from small to large bases. This reflects that the market still values increasingly optimistic supply chain data but wants to avoid AI TAM-exposed names. More alpha hunting than beta.

In this report, we want to review the evolution of CapEx attitudes, break down OpenAI’s $1.4 trillion, identify core metrics to prove or disprove the AI bubble thesis, and examine progress at the eye of the storm - OpenAI.