To The Moon

$Oracle(ORCL)$  You have put your finger on the key issue: Oracle has turned itself into a leveraged macro proxy on AI infrastructure and OpenAI. Let us break this down carefully.

---

1. Where Oracle stands now

Price action

All-time high: about US$345.7 in early September 2025. 

Latest close: about US$198.7 on 21 November 2025, a drawdown of more than 40 percent from the peak. 

Despite that, total return is still roughly +20 percent year to date, because the run-up before September was so strong. 

Balance sheet and AI capex

Around US$104 billion of debt outstanding, including US$18 billion in bonds already issued. Oracle is currently spending more on AI and cloud build-out than it generates in operating cash, relying on external financing. 

Banks are preparing an additional US$38 billion debt package to fund data centres in Texas and Wisconsin that support AI workloads for Oracle and OpenAI (Stargate and related projects). 

Some analysts project that Oracle’s net debt could approach US$290 billion by 2028 if this build-out continues at present pace. That is a projection, not a fact, but it is part of what markets are reacting to. 

In other words, the equity has already repriced sharply, but the company is not cheap in a conservative sense once you factor in leverage and execution risk.

---

2. What the CDS spike actually means

CDS spreads have roughly doubled

Articles tracking the market note that Oracle’s 5-year CDS has risen from a stable range into about 105–110 basis points, roughly double where it traded earlier, and well above the North American investment-grade index around 50–60 basis points. 

This is why you see headlines such as "CDS costs have surged 2x" and "Oracle CDS becomes a hedge against an AI crash". 

Default risk in context

Using standard CDS maths and a 35 percent recovery assumption, a 5-year spread of about 108 bps implies an annual default probability of roughly 1.7 percent and a cumulative 5-year default probability around 8 percent. 

That is elevated for a large investment-grade software company, but it is nowhere near "imminent default" territory. For comparison, highly stressed names like CoreWeave trade around 600–700 bps and carry 40 percent type 5-year default odds. 

How to interpret it

The CDS market is signalling that Oracle’s balance sheet has become meaningfully riskier than a typical blue-chip tech name.

It is also being used tactically: banks and investors with large long AI exposures (Microsoft, Nvidia, OpenAI ecosystem) can buy Oracle protection as a macro hedge on an AI spending bust, since Oracle is one of the most levered names to the build-out. 

So the CDS spike is a warning light about leverage and concentration, not proof that Oracle is about to blow up.

---

3. Will Oracle continue to decline?

No one can give you a binary answer, but we can look at the drivers and skew.

Bearish forces

1. Leverage and ratings pressure

Fixed-income research has started to highlight that Oracle’s AI capex and contract commitments far exceed what its free cash flow can comfortably support, leading to heavy reliance on borrowing. Some analysts have warned that its credit rating could drift toward BBB-, just one notch above high yield. 

Any downgrade or negative outlook from the rating agencies would increase funding costs and could weigh on the equity further.

2. Execution risk on AI revenues

The debt is effectively a front-loaded bet that long-term contracts with OpenAI, CoreWeave and other AI tenants will fill these data centres at healthy margins. 

If AI workloads or pricing disappoint, Oracle would be stuck with a very heavy capital base and interest burden, which would compress returns on equity and free cash flow per share.

3. Market positioning and sentiment

After the "AI darling" narrative earlier this year, there has been a sharp reversal as investors realised how capital intensive the strategy is. Recent coverage explicitly frames Oracle as a test case for whether the trillion-dollar AI build-out can be financed without creating a credit crunch. 

Supportive forces

1. Core business is still real and profitable

Oracle still runs a large legacy database and applications franchise with recurring revenues and decent operating margins. AI is additive on top of that, not the only business line. 

2. Valuation no longer euphoric

With a drop of over 40 percent from the peak and earnings still positive, a substantial part of the "AI dream" premium has already been removed. If the company can show a credible path where AI capex translates into contracted revenue and free cash flow, the stock can stabilise or even re-rate.

3. Systemic support for AI infrastructure

Governments and hyperscalers are deeply embedded in the AI build-out. It is plausible that demand for compute remains strong over the medium term, even if there are periods of overbuilding and pricing pressure.

My read

Short to medium term, downside risk remains elevated, because:

Credit markets have clearly re-priced Oracle as a higher-risk AI financing vehicle.

There are additional debt transactions still to be completed for data centres.

The company will be under pressure to prove that AI contracts support its capex and debt load.

So while a sharp further collapse is not inevitable, the path of least resistance is still choppy with a real possibility of new lows if:

CDS spreads widen further,

a ratings agency moves toward a downgrade, or

any AI partner scales back commitments.

---

4. "Will the OpenAI bubble be paid for by Oracle?"

This is the heart of your question. A fair way to frame it is:

> Who absorbs the cost if AI infrastructure is overbuilt relative to sustainable profit: the AI model companies, the cloud landlords, or the credit investors?

How Oracle is exposed to OpenAI

Oracle is tied into the Stargate and related mega-data-centre projects that provide compute to OpenAI and others, with a large chunk financed via bank loans and bonds. 

Oracle effectively plays three roles:

1. Cloud infrastructure provider and operator.

2. Long-term counterparty to AI customers, often with take-or-pay style capacity commitments.

3. Borrower on a very large stack of project and corporate debt.

If the OpenAI economics disappoint, Oracle cannot simply walk away from the interest payments and capex already committed. The equity and subordinated creditors are the shock absorbers.

Why Oracle looks more exposed than some other big techs

Compare with peers:

Amazon, Microsoft, Alphabet, Meta are also issuing substantial AI-related debt or using structured project finance for data centres. 

However:

They have larger and more diversified revenue bases, with very profitable cloud businesses already at scale.

Meta, for example, used an off-balance-sheet project finance structure for Hyperion, so a lot of the debt sits in a joint venture, not directly on Meta’s own balance sheet. 

Their net cash or modest net debt positions give them more room to absorb mistakes.

By contrast, Oracle:

Is starting from a heavier leverage position and is pushing it further. 

Has less scale in its cloud business compared with AWS or Azure and is betting aggressively on catching up through OpenAI and similar contracts.

So yes, if there is an "OpenAI bubble bill", Oracle is one of the most obvious candidates to pay a significant share of it, in the form of:

Lower equity valuation and multiple compression,

Higher funding costs,

Potential rating pressure, and

Reduced flexibility for buybacks or other shareholder-friendly actions.

That said, the burden would not be Oracle’s alone. Banks, bondholders, project finance vehicles and perhaps some governments would also share the pain if AI capex has been overdone.

---

5. How to think about Oracle from an investor’s point of view

If you strip away the noise:

Oracle is no longer a conservative "database stock".

It is now a levered AI infrastructure bet with meaningful credit sensitivity, used by the market as a hedge against an AI downturn.

For a portfolio:

Position sizing and risk management matter more than fine valuation arguments.

Oracle makes more sense as a high-beta satellite position rather than a core long-term compounder, unless you have very high conviction that:

1. OpenAI and similar AI tenants will deliver sustainable, profitable demand, and

2. Oracle can convert that into strong free cash flow after interest.