Nvidia’s $450B Reversal Will Change The Way You View AI Stocks

Nvidia dropped earnings at 4:05 PM on November 20, 2025. The stock jumped 8% right away.

Within 18 hours, trading algorithms erased +$450 billion in value—purely by reading the report faster and more accurately than any human possibly could.

And what they picked up wasn’t bullish at all. It was a stack of financial red flags hiding in plain sight.

What the Algorithms Saw Before Analysts Did

Nvidia’s unpaid bills ballooned to $33.4 billion, up 89% in one year. The time it takes to get paid stretched from 46 to 53 days. Those seven extra days alone mean about $4.4 billion in cash that isn’t showing up when it should.

Inventory also jumped 32% to $19.8 billion, despite Nvidia insisting demand was off the charts. The last time they launched under similar conditions, inventory actually fell.

And to top it off, GPU spot prices have fallen 40% since August—another clear sign demand isn’t matching the hype.

The Dangerous Loop

The whole AI sector is basically trading promises. Nvidia funds xAI. xAI borrows billions to buy Nvidia chips. Microsoft funds OpenAI. OpenAI commits massive spending back into Azure. Oracle signs huge contracts that rely on Nvidia hardware. Almost nobody is paying in actual cash. Receivables just keep piling up.

With how many strange things we’ve been seeing in the markets lately, this might just be the beginning. We’re noticing similar trends across other AI-linked stocks.

Alphabet doesn’t exactly scream “Buffett stock.” High-growth tech, elevated multiples, and shifting business models aren’t typically his style.

That’s why this buy is likely the work of Berkshire’s younger portfolio managers, not Buffett himself.

What is Buffett’s doing, though, is far more important: Berkshire’s cash position has surged to $381.7 billion, now making up roughly one-third of the company’s market cap.

That’s unprecedented for a man who spent decades telling investors that sitting on cash is a losing long-term bet.

Even stranger, Berkshire reduced two of its largest holdings last quarter by $15.5B combined, and didn’t meaningfully increase any major positions. All signs point to the same conclusion: Buffett is playing defense.

Even after a brutal 25% drop, $PLTR is still the single most expensive stock in the entire S&P 500 — several times over.

Here’s the uncomfortable truth: Palantir trades at 99× sales today. The next closest company in the S&P 500 is AppLovin at 34× sales.

That means Palantir could fall another 66% and still rank as the priciest stock in the index by this metric. That’s how stretched things became.

The CEO has suggested “market manipulation” is causing the weakness, but that misses the core issue. The risk-reward profile is completely skewed toward risk, and investors are finally waking up to that reality.

The Rule of 40 is one of the cleanest ways to evaluate software businesses. It says that a company is considered exceptional if its Revenue Growth Rate + EBITDA Margin ≥ 40%.

On this chart, Verisign (VRSN) is positioned beautifully — massive EBITDA margins and stable growth push it well above the 40% threshold. But there’s a catch: the stock isn’t cheap.

Its valuation has stretched high enough that even Warren Buffett trimmed his long-held position, signaling that quality doesn’t always equal value.

Over the long run, stock prices tend to follow cash flow. That’s why Meta’s recent behavior is so strange.

Its operating cash flow has compounded at 25% annually, yet the share price has flattened and its valuation has dropped to the lowest multiple in Big Tech.

We’ve seen this movie before. In 2022, Meta suffered a 50%+ drawdown after a stretch of weak quarters and heavy spending.

Now? The reaction looks similar — but the fundamentals don’t. Meta actually beat Q3 earnings, and the only thing that dragged results down was a one-time tax charge, not business deterioration.

📰 Michael Burry launches Substack.

🚀 Saylor shrugs off delisting worries.

💻 Nvidia earnings — quick recap.

📉 Gold cools as rate cuts fade.

💊 Novo Nordisk sinks on trial miss.