The Truth Behind Oracle's Stock Plunge: The $10 Billion Data Center Deal and the $250 Billion Lease Bomb
The reason for the volatility in Oracle's stock price is not so much about its performance, but rather a shift in the 'funding narrative' regarding who is supporting AI data center expansion, under what conditions, and at what scale. Three points were linked together: (1) the departure of a key partner from the 1GW Michigan facility, (2) the expansion of leases (future commitments) that are easily kept off-balance sheet, and (3) the concentration of AI demand centered on OpenAI and suspicions of 'circular trading'.
1. 'Partner withdrawal from a $10B deal'—the surprise the market hates most
The spark for this incident was the report that, regarding the 1GW-class data center plan in Michigan, Blue Owl Capital, considered the largest financial partner, would not be supporting the $10 billion project.Oracle explained that 'the project is on schedule' and that 'as a result of choosing the optimal equity partner from among competitive options, it was not Blue Owl this time'.
The issue is not the 'delay,' but the association that 'there must be a reason why the lender or investor is pulling out.' Doubts were brought to the forefront in the program as well, with questions like 'Is it becoming difficult to raise funds?' and 'Are the lenders seeing something?'
2. Data centers: From 'buying' to 'leasing'—the expansion of leases amplifies anxiety
This is the core of the argument. Traditional CAPEX was a straightforward expenditure: 'buying computers, buying cables, building data centers.' On the other hand, in recent years, the movement to 'lease' rather than 'own' data centers has been spreading.
While explaining this as a structure that 'easily becomes off-balance sheet,' it was mentioned that the commitments Oracle promises to pay in the future are 'around $250 billion over the next 20 years.' In other words, it is not just immediate capital investment, but a massive bundle of fixed costs extending into the future. This is what has suddenly raised investor risk sensitivity.
3. Oracle is a 'newcomer to hyperscalers': Execution risk becomes a premium
Oracle is historically a software company and a latecomer to the massive data center construction race. While expressing expectations on the ground that 'this could be a transformative historical build-out for the entire industry,' it was also pointed out that 'many things need to go according to plan'.
As a comparison, it was explained that 'even if Microsoft has $100 billion in commitments, the market is unlikely to be surprised, but Oracle's spending is a large ratio relative to the company's size.' In short, even with the same amount of money, the perception changes depending on 'stamina' and 'experience.' Oracle is in a position here where it is easy to be viewed with suspicion.
4. OpenAI x Cloud x Investment: Is circular trading 'evil' or inevitable?
The discussion connected to the story that OpenAI is negotiating to raise at least $10 billion from Amazon, and in the past, has had a $30 billion, 7-year computing resource contract with Amazon. Because this looks like a 'cycle' of 'investment -> purchasing cloud services with that money -> returning as revenue,' it was bluntly stated that 'creative financing will be scrutinized'.
On the other hand, the investor side had a tone that 'in a new technological frontier, it is natural for companies to team up and create an ecosystem. This is something to watch, but not an immediate rejection.' What was emphasized here was not the form of the funding, but whether the final demand is strong.
5. Investor's map: AI 'space exploration' is not a winner-take-all game
An impressive metaphor was the view that 'AI is a space race, but there are many races on many moons.' There is room for multiple players to grow simultaneously, not just NVIDIA, but also Google's TPUs, Amazon's in-house chips, and AMD.
So how should we view Oracle? Supporters stated, 'The demand signal is strong, and building now is a once-in-a-generation opportunity. The market may be overreacting.' The points of concern for the skeptics are: (1) what future AI demand will look like in 2028–2029, (2) how certain the return on investment centered on 'training' is, and (3) the risk of customer (especially OpenAI) concentration.
Conclusion
This news is the moment when the 'bet on Oracle transforming into an AI infrastructure company' began to be re-evaluated by the market in terms of two points: funding (who supports it) and accounting debt perception (lease commitments). Investors have entered a phase where they are looking not at the 'number of data centers,' but at the quality of contracts (customer diversification, duration, unit price) and the quality of funds (balance of debt, leases, and equity).
