‍#每周来晒 #美股财报季

Why I’m most focused on Nvidia’s earnings—its performance will determine whether the Nasdaq correction is over

Meta’s earnings-driven move has just played out: revenue +26%, EPS beat expectations by 13%, but free cash flow turned negative for the first time since the company went public. The stock was hammered 8% over two days, dragging the Nasdaq down with it. Now the market is focused on one question: Can AI companies earn back the hundreds of billions they pour in every year? It seems the Nasdaq won’t find a floor until that question gets a satisfactory answer, and I think Nvidia could well be the “savior”!

🔥 Why NVDA in particular—not Microsoft or Meta?

The logic here is actually very simple:

The AI bull market over the past two years has essentially been a chain of transmission—cloud providers believe in AI → pour money into data centers → money flows into Nvidia’s pockets → NVDA’s results explode → the stock doubles → the Nasdaq rises along with it.

Now Meta has put a crack in the first link of that chain: free cash flow has turned negative, capex guidance has risen instead of falling, and it plans to issue new shares to raise funds. The market immediately starts reasoning backward—if the companies spending the money are starting to feel the pain, can orders for the company collecting that money (NVDA) still keep growing?

Microsoft can’t answer that question (it’s spending money too), and Meta can’t answer it (no matter how profitable its ad business is, it can’t fill the hole left by AI). Only NVDA can answer it, because it holds the industry’s strongest cards: hyperscalers’ actual orders, procurement schedules, and capital spending plans for the next 6–12 months.

So the November 18 earnings call is, on the surface, about Nvidia’s own earnings, but in reality, the entire market will be using a magnifying glass to put the whole AI capex chain through a “stress test.” If it beats expectations and keeps raising its guidance, that will tell all investors: Relax, capex is still accelerating; Meta is just an isolated case. If it falters or gets timid with its Q4 guidance, then Meta’s chart won’t be an isolated case—it will be the first domino, and capex expectations for Microsoft, Google, and Amazon will all have to be revised down. The Nasdaq correction will be far from over.

📋 On earnings night, keep your eyes on these 5 numbers (everything else is noise)

No need to get overwhelmed by the jargon in the news. For Nvidia’s earnings, just watch these metrics:

① Actual revenue vs. $108 billion guidance ** 🎯 The company’s own Q3 guidance is $108 billion ±2% (that is, $105.8–110.2 billion), and this figure explicitly excludes data center compute revenue from mainland China. Data center revenue was $89 billion last quarter, so it needs to rise another 21% sequentially just to reach this number. The consensus estimate is around $110–112 billion, slightly above the company’s guidance—** in other words, merely “meeting” guidance won’t cut it; it needs to beat $112 billion for the stock to rally after hours.

② Q4 guidance—this is tonight’s real trump card 💣 Don’t focus on the Q3 number; what Wall Street really cares about is the next-quarter guidance Jensen Huang gives. For FY27, NVDA itself has said the second half will accelerate. If Q4 sequential growth drops to the single digits, or guidance merely “meets expectations,” the market will immediately interpret that as “the Blackwell cycle has peaked, and Rubin hasn’t taken over yet.” A 20%+ after-hours drop would be no joke. Conversely, as long as Q4 guidance comes in at $118–120 billion or higher, along with a line like “orders are booked through 2027,” the Nasdaq could come right back to life that night.

③ Can gross margin hold above 74%? 📉 The company’s non-GAAP gross margin guidance is 74% ±50bp. This figure is even more sensitive than revenue—if gross margin falls, the market will instantly imagine three scenarios: ramp-up costs as yields improve during the Rubin transition; another H20 impairment in China (it already took a $4.5 billion hit last quarter); or forced discounting to win customers. If gross margin falls below 73.5%, the stock will get hammered after hours no matter how much revenue beats expectations.

④ What exactly will it say about China? 🇨🇳 This guidance once again assumes “zero data center compute revenue from China.” Analysts will certainly ask on the call: How are talks going on new H20 licenses? Is China revenue truly zero, or is there some limited revenue? This is an asymmetric option—if it says, “We’ve regained some licenses,” that’s a pure incremental upside surprise, and the stock will open sharply higher; if there’s another impairment charge, that’ll be salt in the wound.

⑤ Major-customer concentration & inventory-cycle signals 🏭 Listen for two details: first, whether the top four customers’ share of data center revenue continues to rise (the higher the concentration, the bigger the shock if one hyperscaler cuts orders); and second, the sequential changes in data center inventory and accounts receivable—if both pile up, that suggests channel stuffing, and the 2027 story will need to be discounted.

🌊 The ripple effect on U.S. tech stocks: two scenarios

After earnings night, there are only these two possible paths—nothing in between:

🟢 Scenario A: Beat + raised Q4 guidance + gross margin holds above 74% This amounts to an official stamp of approval that “AI capex is still accelerating, and Meta’s kind of anxiety is unnecessary.” Nasdaq futures will surge that night, NVDA itself will race to $230–240, and the rally will pass down the whole chain like a game of hot potato—Broadcom, TSMC, AMD, Micron, SanDisk, Corning, Coherent, and other optical-module and memory players will all rebound together. More importantly, when Microsoft, Google, and Amazon report earnings later, the market will automatically find a reason to excuse their high capex: “Look, NVDA orders are still growing, which means there’s someone to take your investment.” This correction will bottom out that very night, November 18.

🔴 Scenario B: Revenue barely meets expectations, but Q4 guidance is conservative, or gross margin falters Then Meta’s problem won’t be Meta alone—the market will immediately settle on a new narrative: “hyperscalers’ AI capex has already exceeded real demand; NVDA has peaked first, and the whole chain has peaked.” NVDA itself will be down 10% or more that night, Nasdaq futures will plunge, and Broadcom, TSMC, Micron, and the rest will crash along with it. Then you’ll see a very familiar scene: over the next two weeks, every tech giant reporting earnings will see its stock rise and then fall just like Meta’s whenever it raises capex guidance—because the market no longer believes in the logic that “investment = future revenue.” That would be the real halfway point, and the Nasdaq correction would go on to test its annual moving average.

🎯 My position: leaning bullish

Honestly, NVDA at $210 and a 26x P/E already reflects a considerable amount of pessimism about “AI capex peaking”—the move from $243 to $210 in October was an emotional selloff triggered by Meta. I’m inclined to think the actual Q3 numbers will probably be positive (around $112 billion); the real deciding factor is Q4 guidance.

If Jensen Huang keeps saying “Rubin orders are sold out, and we don’t have enough capacity for 2027”—then tonight will be a golden buying opportunity. Don’t hesitate below $200; this is the second chance to get in this year.

If he suddenly changes tack and starts talking about “customers being more rational and inventories healthier”—ignore how good the revenue looks and cut your position after hours. That’s the standard language of a cycle peak.

Here are the key levels to keep in mind (before earnings): first support below is the $200 round-number level; if that breaks, look to $190 (the September low of $208 has already been tested once; $190 is the base from June this year). First resistance above is $222 (the October 21 rebound high); only a move back above that would count as a bottom. Strong resistance is $235–243 (the October highs).

Let me put it more strongly: this NVDA earnings report may determine not only the Nasdaq’s fate, but also whether the “AI narrative” can continue to lead the U.S. stock market. As long as the standard-bearer is standing, tech stocks are still in a bull-market correction; if it falls, that’s a rerun of the 2000 dot-com bubble—except this time, nobody wants to be the last one left standing on the stage. $NVDA $NVDA