AI infrastructure hides potential “explosion” risks? Oracle: a force majeure notice ignites market fears...
Earlier today, Oracle’s stock fell by more than 7% intraday and ultimately closed down 3.47%. The spark was a force majeure notice: New Mexico rejected the company’s data center construction plans. But the weight of this matter goes far beyond an ordinary negative headline.
First, the incident itself. Oracle issued a formal notice to the Project Jupiter developer, arguing that the planned 2028 commissioning of a $300 billion data center to provide compute capacity to OpenAI is no longer possible.
After the news broke, Oracle dropped more than 7% at one point. Blue Owl Capital fell 3.6% in tandem, while Bloom Energy slid 3.1%.
Four reasons the market is highly sensitive:
1. A severe mismatch between scale and leverage.
Project Jupiter’s total investment is about $165 billion, while Oracle’s latest financial report discloses total debt of $129.5 billion. When a single project’s investment exceeds the company’s entire debt, any delay isn’t just a project issue—it becomes a balance-sheet problem.
2. The financing chain can’t hold up anymore.
According to reports, the roughly $18 billion syndicated loan tied to this project has seen the lead bank’s quoted price sink to 89 to 91 cents on the dollar, a discount of nearly 10%. Distribution has stalled for more than six months. The CDS spread also jumped to a record high, with the 5-year tenor reaching 221.78 basis points at one point. The credit market has reflected concerns earlier and more honestly than the stock market.
3. This isn’t an isolated case. The IPO of SB Energy, another anchor project under the “Stargate” program, has already been delayed due to similar setbacks. Once Oracle issued this notice, it may further undermine confidence there. If OpenAI’s compute capacity gap forms, the expansion of products will face a real, tangible hit.
But perhaps the Oracle incident is only a side episode. AI development won’t stop, and the risks for highly leveraged companies shouldn’t be ignored. As for “chip-making” companies that sell shovels, the risk is lower—revenue is clearer, and Shu Qin also remains more optimistic on a long-term basis. We discussed AMD on the 22nd—it has already broken to a new high of 640. Congratulations to everyone! For the short term, I think it may be reasonable to trim some positions on strength~
Earlier today, Oracle’s stock fell by more than 7% intraday and ultimately closed down 3.47%. The spark was a force majeure notice: New Mexico rejected the company’s data center construction plans. But the weight of this matter goes far beyond an ordinary negative headline.
First, the incident itself. Oracle issued a formal notice to the Project Jupiter developer, arguing that the planned 2028 commissioning of a $300 billion data center to provide compute capacity to OpenAI is no longer possible.
After the news broke, Oracle dropped more than 7% at one point. Blue Owl Capital fell 3.6% in tandem, while Bloom Energy slid 3.1%.
Four reasons the market is highly sensitive:
1. A severe mismatch between scale and leverage.
Project Jupiter’s total investment is about $165 billion, while Oracle’s latest financial report discloses total debt of $129.5 billion. When a single project’s investment exceeds the company’s entire debt, any delay isn’t just a project issue—it becomes a balance-sheet problem.
2. The financing chain can’t hold up anymore.
According to reports, the roughly $18 billion syndicated loan tied to this project has seen the lead bank’s quoted price sink to 89 to 91 cents on the dollar, a discount of nearly 10%. Distribution has stalled for more than six months. The CDS spread also jumped to a record high, with the 5-year tenor reaching 221.78 basis points at one point. The credit market has reflected concerns earlier and more honestly than the stock market.
3. This isn’t an isolated case. The IPO of SB Energy, another anchor project under the “Stargate” program, has already been delayed due to similar setbacks. Once Oracle issued this notice, it may further undermine confidence there. If OpenAI’s compute capacity gap forms, the expansion of products will face a real, tangible hit.
But perhaps the Oracle incident is only a side episode. AI development won’t stop, and the risks for highly leveraged companies shouldn’t be ignored. As for “chip-making” companies that sell shovels, the risk is lower—revenue is clearer, and Shu Qin also remains more optimistic on a long-term basis. We discussed AMD on the 22nd—it has already broken to a new high of 640. Congratulations to everyone! For the short term, I think it may be reasonable to trim some positions on strength~
