Oracle's Credit Just Got Cut for Its AI Bet. Is Alphabet Next?

On July 9, 2026, S&P Global Ratings cut Oracle's credit rating to BBB-. That's one single notch above junk. Two weeks later, Moody's put six of the biggest AI spenders in the world — Microsoft, Amazon, Alphabet, Meta, Oracle, and CoreWeave — on notice, warning that "unprecedented" AI capital spending is starting to erode the credit quality that made these companies bulletproof borrowers for two decades.

Photo by Liz Henderson on Pexels

So here's the question every investor holding hyperscaler stock should be asking right now: is this an Oracle-specific problem, or is it the first crack in something bigger?

The number that should worry you: $785 billion

Moody's now projects combined hyperscaler capex will hit $785 billion in 2026, on its way to roughly $1 trillion in 2027 — about six times what these same companies spent in 2022. That capex is increasingly funded by debt rather than free cash flow. Direct debt across the six companies Moody's tracks has climbed to roughly $460 billion, and off-balance-sheet data center lease commitments have ballooned to $1.2 trillion, more than $820 billion of which is for leases that haven't even started yet.

None of that shows up as traditional debt on a balance sheet. Moody's treats it as debt-equivalent anyway, because a 15-year lease on a data center campus is a bill you have to pay whether or not the AI demand shows up to justify it.

Oracle: the cautionary tale, not the outlier

Oracle's downgrade wasn't a surprise so much as a preview. S&P now expects Oracle's fiscal 2027 capex to hit $90-95 billion, up from a prior estimate of just $60 billion, driven by data center builds tied to its OpenAI contracts. The company spent $55.7 billion on capex in fiscal 2026 against $32 billion in operating cash flow — meaning free cash flow came in at negative $23.7 billion for the year. S&P's forecast for fiscal 2027 is a free cash flow deficit of negative $42 billion, nearly double.

Total debt now sits at $167 billion, with leverage ratios expected to hit the mid-4x range next year — above what S&P considers safe for a standard BBB credit profile. The single biggest risk factor cited: OpenAI accounts for roughly half of Oracle's remaining performance obligations. If OpenAI's own path to profitability stumbles, Oracle doesn't have much diversification to fall back on.

That concentration risk is the real lesson here, more than the capex number itself.

Why Alphabet isn't Oracle — yet

Alphabet looks like the opposite case on paper. It's sitting on $242.5 billion in cash and marketable securities, long-term debt of roughly $98-100 billion (after a fresh $20 billion debt raise), and a debt-to-equity ratio of just 0.14. Moody's itself said Microsoft, Alphabet, Amazon, and Meta retain some of the strongest corporate balance sheets in the world, with leverage well below typical downgrade thresholds.

But the direction of travel matters as much as the starting point. Alphabet guided full-year 2026 capex to $195-205 billion — a number that's been revised upward three times this year, from $175-185 billion in February to $180-190 billion in April to the current range. Q2 2026 capex alone was $44.9 billion, and for the first time in the company's history, that pushed quarterly free cash flow negative, to roughly negative $5.9 billion. Alphabet is now tapping both debt and equity markets to fund the buildout, something it essentially never had to do before this cycle.

As I wrote in Google's Cloud Backlog Hit $514 Billion. The Growth Rate Just Slowed., the demand side of Alphabet's cloud story is still strong but decelerating at the margin. That combination — capex accelerating while backlog growth decelerates and free cash flow turns negative — is exactly the pattern that preceded Oracle's downgrade, just earlier in the cycle and with a much bigger cash cushion to absorb it.

Photo by stevepb on Pixabay

The variable that actually separates the safe hyperscalers from the risky ones

It's not the capex number. Every major hyperscaler is spending an eye-watering amount right now. The variable that matters is revenue concentration and diversification of the demand backing that spending:

• Oracle is building infrastructure that's roughly half dependent on a single customer, OpenAI, whose own economics remain unproven at scale.
• Alphabet is spending against demand spread across Search, YouTube, Google Cloud, and Gemini API usage — no single external customer can sink the thesis.
• Microsoft and Meta sit somewhere similar to Alphabet: internal product demand (Copilot, ad-ranking AI, internal infrastructure) rather than one anchor tenant.

That's why Moody's singled out Oracle for an actual downgrade while leaving the other five on a more cautious "watch the trend" footing. Balance sheet strength buys you time. It doesn't buy you immunity if the capex keeps compounding at this rate for another two or three years.

What would actually change my mind here

I don't think Alphabet is at Oracle-level risk today, and the cash pile backs that up. But there are two things I'm watching that would change my read: a fourth consecutive upward capex revision without a corresponding acceleration in Cloud backlog growth, and any sign that Gemini API/enterprise AI revenue is failing to convert bookings into recognized revenue at the pace guided. If free cash flow stays negative for two or three more quarters while capex keeps climbing, the "pristine balance sheet" argument gets weaker every quarter, not stronger.

The bond market is already pricing some of this in — Moody's noted that spreads across the hyperscaler group have been widening even for names that haven't been downgraded. Equity investors have mostly ignored that signal so far because the AI revenue story is still exciting. Credit markets are usually right before equity markets are.

This is not financial advice — always do your own research before making investment decisions.

My takeaway for today: if you own hyperscaler stock, stop looking only at capex guidance and start pulling up each company's free cash flow trend and customer concentration alongside it. That's the combination that told you Oracle was in trouble months before the downgrade actually landed.