The 4-Day Bleed: A Confession from the Edge

Four. Consecutive. Red. Days.

The S&P 500 has posted four straight losses. The Nasdaq is down five out of six. Bitcoin crossed below $90,000 and clawed back up like a drunk trying to remember where they left their car keys. You know that feeling? When you're watching something collapse and you're not sure if you should laugh or start taking antacids? That's been Tuesday through Friday.

The noise is unbearable.

Every financial news outlet is running some variation of the same headline wrapped in different fonts: "NVIDIA Stock Slides Ahead of Earnings." "Markets Wait for the Earnings Verdict." "Tech Sell-Off Intensifies." They're all code for the same thing. Nobody knows what happens tomorrow after the close. And knowing that is starting to crack people's skulls.

Here's the problem: everyone—and I mean everyone—is treating Nvidia's earnings like it's the Super Bowl, the State of the Union, and the lottery drawing combined. Options markets are pricing in an 8.5% move. That's a $320 billion swing in market value. For one company. One earnings call. Analysts have adjusted their estimates for Q3 revenue up to $54.9 billion, up to $55.2 billion depending on which outlet you trust. The consensus is now above what Nvidia itself guided. Think about that. Wall Street thinks the company is going to beat its own forecast.

And if it doesn't? If the number comes in at $54.5 billion instead of $55 billion? The stock will crater. Not because the result is bad—it's incomprehensibly good in any reasonable universe—but because it missed the imaginary mark traders painted on an imaginary wall.

This is how delusion works when the only thing holding it up is the shared belief that it's not delusional.

The Insider Exodus

Peter Thiel's fund just sold its entire $100 million Nvidia stake. SoftBank offloaded $5.8 billion of it. These aren't retail FOMO people. These are the people who are supposed to know something. They've been sitting at the table watching the math get worse, and now they're walking.

SoftBank's situation is particularly grim—they're liquidating positions to fund their own AI bets. They're not selling because they don't like Nvidia anymore. They're selling because they're drowning in their own capex and need liquidity. That's not bullish. That's desperation.

But here's what nobody's willing to say out loud: if the smart money is exiting and the dumb money is entering, what exactly are we building here? A monument to belief or a Ponzi scheme where the exit velocity matters more than the actual fundamentals?

The Tell Nobody Wants to See

Bank of America's Fund Manager Survey came back this month and it's screaming. Investors are holding 3.7% in cash—the lowest level possible. They're at the most overweight in equities since February. Most bullish on commodities since September 2022. That's the smell of people who've already shoved all their chips onto the table. When everyone is leaning one direction and there's barely any powder left on the sidelines, that's not a market condition. That's a configuration for a flush.

Michael Burry—the guy who saw the 2008 housing collapse coming—showed up with a massive bearish position on Nvidia and Palantir last week. He's claiming that companies are playing accounting games with data center depreciation. Maybe he's right, maybe he's paranoid. But he didn't get famous for being wrong about when things are going bad. His presence on this trade is a warning light that's been flashing orange.

The Catch-22 Nobody Can Solve

So here's the bind they're stuck in, and it's beautiful and horrible at the same time. If Nvidia crushes earnings—comes in with guidance that looks amazing, demand that looks unstoppable—then everyone will read that as validation that the AI spending will never stop. Which means valuations might actually compress because "yeah, okay, the story checks out but that doesn't mean you get to own it at 25x forward revenues."

But if Nvidia misses or guides slightly lower? If there's even a whiff of deceleration? Then it confirms the bubble fears and we're off to the races downward. There is no acceptable outcome. The only scenarios are "good but disappointing" or "bad."

A Catch-22 with a market cap of $4.6 trillion. That's the asymmetry traders are about to live inside for the next 24 hours.

The Actual Numbers (If You Care)

Nvidia is guiding for $54 billion in Q3 revenue. Consensus has moved it to $54.9 billion. Data center, which is 85% of revenue, is expected to be around $49.5 billion. The company has grown revenue 56% year-over-year. Adjusted EPS is expected to be $1.25 or $1.26 depending on which analyst you ask. All of these numbers are astonishing by any measure that existed before the AI craze started.

The company also has a $500 billion order book for Blackwell and Rubin infrastructure through 2025-2026. That sounds insane until you realize it also means there's a hard ceiling on visibility. Wall Street needs to believe the spigot never turns off. The moment it does—even slightly—the composition of that order book becomes a liability instead of an asset.

Nvidia's gross margins are 72%. Net margins are 52%. Return on equity is over 100%. On paper, this is the most impressive commercial machine in history. On paper.

The Real Stakes

This isn't about Nvidia being a good or bad company. It's about what Nvidia represents. The company is 8% of the S&P 500. Mega-cap tech is 40% of the index. That concentration is a feature of our market now, and it's also the thing that will break it when the mechanics shift.

Consumer discretionary stocks have already lost 3% in November. Williams-Sonoma is down more than 6%. Tesla down 5%. That's the smaller boats already taking on water. The big boats—Apple, Microsoft, Google—haven't moved much, but that's because everyone's too focused on Nvidia to notice what else is falling.

Retail earnings tomorrow morning—Target, Lowe's, TJX—will tell us if the consumer actually has any ammunition left. But after a 39-day government shutdown, delayed economic data, and a labor market that's been sending mixed signals for months, we're essentially flying blind.

And tomorrow night, Nvidia reports and the whole thing either gets another boost of confidence or starts the unraveling.

The Thing Nobody Wants to Admit

We're in a regime now where sentiment is doing 80% of the heavy lifting. The fundamentals are real—Nvidia is making money, the demand is there, the technology is game-changing—but the valuation is a collective agreement that this goes on forever. And collective agreements break when the first person convinces themselves it might not.

That person might be Jensen Huang's next sentence. It might be a Q4 guidance number that's too conservative. It might be a single analyst on the call asking the wrong question.

Or it might be that nothing goes wrong and we bounce off this dip and climb another 15% and nobody ever remembers the 4-day bleed in November 2025 because we were all too distracted by whatever comes next.

But tonight, in the quiet before the storm, the odds of all three scenarios feel exactly the same.

Tomorrow at 4:20 p.m. Eastern, we find out which world we're actually living in.

Until then, the best advice I can give is this: watch the numbers, sure. But more importantly, watch the how. Watch if management sounds confident or hedged. Watch if they're excited about 2026 or already lowering expectations. Watch the implied guidance buried inside the guidance.

The market doesn't move on truth tomorrow. It moves on interpretation. And interpretation moves faster than reality ever could.

Stay sharp. Tomorrow's going to hurt someone.