Micron Google SK Hynix - Stress Test Nobody is Talking About!

[00:00:00]Google reports earnings today after the close, and if you go on financial Twitter right now, you would think this is the single most important event of the quarter. It is not. Do not get me wrong, it matters. It is the first of the Magnificent Seven to report. It sets the tone, and the options market is pricing a 5 and 1/2% move in either direction. This is the real money, but Google today is the opening act. It is the support band. The actual headliner is 7 days from now, and almost nobody is talking about it because on July 29th, Microsoft reports, Meta reports, and SK Hynix reports. All on the same day. And if you do not know why those three names landing on one calendar day is genuinely enormous deal for the memory trade, for Micron, for Nvidia, and for the entire AI complex, then stay with me because the end of this video, I'm going to give you three specific things. One, the exact reason great earnings have been getting sold for three straight weeks, so you don't stop being surprised by it.

[00:01:04]Two, the three numbers to watch on every single one of these reports, and it is not revenue, and it is not earnings.

[00:01:11]Most people are watching the wrong line entirely. And three, the full calendar map of which companies move this trade over the next eight trading days, and what each one actually tells you. Now, get those three things, and you are reading this earning season like a professional instead of reacting to headlines like everyone else. Now, this is the unfair advantage because right now, most people are preparing for tonight by refreshing a stock ticker and hoping. Let's get into it. Before we talk about Google, we need to talk about a pattern that has been running for 3 weeks because if you do not understand it, you are going to misread every single one of these reports. Here it is.

[00:01:50]Companies are beating. There's no doubt about that, and the stocks are falling anyway. Micron reported the best quarter in the 47-year history. The stock is in deep drawdown. Taiwan Semi, the company that physically manufactures nearly every advanced AI chip on the planet, posted a record quarter with revenue up 36% year-over-year. The stock closed that week about 3% lower. ASML did not just beat, it beat and raised full-year guidance. The stock popped 3% and then gave it all back and finished the week down over two. And it is not just the chip names. Last quarter, the hyperscalers, Microsoft, Amazon, Alphabet, Meta, reported strong results across the board and their stock mostly went nowhere because investors were focused on one thing, capital expenditure. So, what is actually going on here? In a healthy uptrend, good news gets bought. A beat pushes the stocks higher because there are still buyers who need to get in, but when everyone is already positioned, when the trade is crowded and the room is full, a beat is not a catalyst anymore. It is an exit.

[00:02:59]It is liquidity for the people who want out. Now, the market has stopped paying companies for growth. It's has started asking them what a growth costs. And that shift in what the market cares about is the single most important thing happening in this entire sector right now. The market has essentially gone from the first date where everything you say is fascinating and your ambitions are adorable to the stage where it is asking how much you actually earn and whether you genuinely own that car, which brings us to Google. Here is what the street expects tonight. Revenue of about $117 billion, up more than 20% year-over-year.

[00:03:42]Earnings of about $2.90, up from 230 one year ago. And Google Cloud coming in around 22 and half billion, growing 65%. Now, that cloud number is the one that matters.

[00:03:56]Last quarter, clouds did 20 billion, up 63% and the CFO said on the call that revenue would have been higher if they had been able to meet the demand. That is the cleanest read in the entire report on whether AI infrastructure demand is real. But, here is where it gets strange and this is the part I need you to understand before tonight because the headline number is going to be nonsense. Google owns roughly 14% of Anthropic. At the end of the last quarter, that stake was valued off Anthropic being worth $380 billion.

[00:04:32]Then Anthropic raised money at the valuation of 965 billion, nearly tripled in one quarter. Now, under accounting rules, when a stake you own gets marked up like that, the gain flows straight through your income statement and if it were profit, Bank of America estimates that mark up alone is worth around $80 billion this quarter, which is why the earnings estimate is $8.38.

[00:04:58]While the rest of the street sits at 290. So, tonight Google could print a number that looks like a 200% earnings beat and almost none of it would come from Google selling ads or cloud service. It would come from a spreadsheet being updated because a private company did a funding round. It is a bit like neighbor's house getting a higher valuation and your bank calling to congratulate you on your pay rise.

[00:05:24]Lovely news. You did not actually earn anything and before anyone says this is theoretical, it already happened. Last quarter, Google reported $5.11, which read as a 94% beat. Strip out the 37.7 billion of marks to market gains and the adjusted number was $2.62 against the consensus of 263.

[00:05:49]Now, the quarter that got reported as a blowout was on an operating basis and one cent miss. So, tonight, ignore the EPS headline, go straight to cloud growth, the backlog, and the capex guys.

[00:06:02]That is the business. Everything else is accounting wearing a costume. And honestly, if I could book my house going up in a value as personal income, my year-end numbers would look absolutely spectacular. My bank balance would look exactly the same, but the headline would be tremendous. Right, now the bigger picture because this is the part nobody's walking you through. Google is one hyperscaler. There are five that matter, and they all report inside the next eight trading days. Today, you get Alphabet. Tomorrow, Intel. The next week, the whole thing detonates.

[00:06:36]Wednesday, July 29th, Microsoft reports, Meta, Lam Research reports, Arm, Qualcomm, Vertiv, and on that exact same day, we get SK Hynix reporting. Now, if you do not follow the memory space closely, that name might not mean much to you right now, but let me fix that because SK Hynix might be the single most important company in the AI supply chain that most retail investors have never looked at. SK Hynix holds 56.4% global market share in high bandwidth memory. HBM is the memory that gets physically bonded onto Nvidia's GPUs.

[00:07:14]They are the primary HBM supplier to Nvidia. They are also two in DRAM and number two in NAND. Last quarter, their revenue grew 198% year over year. Their gross margin was 79.3%.

[00:07:29]Their operating margins were 72%.

[00:07:32]Those are not semiconductor margins.

[00:07:34]Those are software margins on physical silicon. The stock is up 515% over the past year, and it pulled back 10% over the past month before bouncing 14% on Tuesday. So, here is why July 29th is the biggest day of this earning season. On that one day, you get two of the largest buyers of AI infrastructure telling you how much they intend to spend, and you get the single largest supplier of the memory that goes into it's telling you what demand actually looks like on the order book. The demand side and the supply side of the same trade. Reporting within hours of each other, and then it keeps going. Thursday the 13th, we get Apple and Amazon. The following Wednesday, August 5th, we've got SanDisk and Western Digital, which are the storage side of the same story.

[00:08:24]This is not one earnings report. This is a two-week stress test of the entire artificial intelligence trade. So, if you were planning a nice quiet fortnight away from the screens, I would gently suggest rescheduling, or at minimum picking a hotel with the genuinely good Wi-Fi. So, what do you actually watch across all these reports? Because it is not revenue, and it is not earnings.

[00:08:47]Wolf Research put it perfectly this week. Revenue and earnings are likely to not matter as much as the amount of capital spending completed in the quarter, and the guide for the rest of the year. Let me read that again. The spending guide matters more than the profit. Three numbers, write them down.

[00:09:07]Number one, capital expenditure.

[00:09:10]The hyperscalers are on track to spend somewhere between 660 and 700 billion dollars this year, and Goldman Sachs estimates roughly 75% of that is directed at AI. Wolf expects them to raise those numbers again over the coming weeks, and here is the connection most people miss. Hyperscaler capex is not an abstract number. It is literally Micron's revenue. It is SK Hynix's revenue. It is Nvidia's revenue. When Microsoft says it is spending more, that money physically flows into memory and chips, which is why Wolf made the point that higher capex guidance may pressure the hyperscalers stocks, but could reinvigorate the semiconductor memory and storage suppliers. Two sides of the same coin moving in opposite directions.

[00:09:58]Number two, free cash flow. This is the pressure point. Bank of America projects hyperscaler free cash flow falls from 191 billion in 2025 to 19 billion this year to -26 billion next year. Negative.

[00:10:14]These are the most profitable companies in human history and the AI build is projected to take their combined free cash flow below zero. That is why Alphabet raised $85 billion in equity, including money from Berkshire Hathaway.

[00:10:29]When companies this profitable start raising capital and issuing debt to fund spending, that is not a red flag on its own, but it is absolutely something you watch.

[00:10:39]And number three, evidence of return.

[00:10:41]Not promises, evidence. Cloud revenue growth, backlog conversion, AI products actually generating revenue. HSBC say they will be specifically looking for whether customers are managing their AI costs by shifting to cheaper models and how rising hardware prices affect spending plans. That is the question underneath everything. Capex, cash flow, returns. That is the checklist. Run it on Google tonight, on Microsoft and Meta next Wednesday, on Apple and Amazon next Thursday. Three numbers, that is genuinely all you need. And you do not need 14 indicators and a chart with so many lines on it that it looks like someone sneezed while holding a highlighter. Now, let me be extremely practical because this is where people lose money. Do not trade into these prints. I know how tempting it is. You have got a view. You are confident and there is a catalyst on the calendar, but understand what you are actually doing.

[00:11:39]You are making two separate guesses at once. First, you are guessing what the number will be and second, you are guessing how the market will react to that number. And as we established at the top of this video, companies have been beating and falling for three straight weeks. Getting the number right has not been enough. That is two coin flips stacked on top of each other and you need both to land. I do not put capital on that. I flip coins for who takes the bins out, not for my portfolio. So, here is the thing I want you to really absorb because it is the whole message of this entire video. The three to five-year bullish case on AI and memory can be completely correct and you can still lose a substantial amount of money over the next three months.

[00:12:27]Those two statements are not in conflict. Demand growing for decades does not mean the stock goes up in a straight line for a decades. It never has, not for any stock in any cycle ever.

[00:12:39]Short-term volatility over this earnings run is close to a certainty. You have got five hyperscalers, the largest HBM supplier on earth, and the entire storage complex all reporting inside two weeks into a tape where hedge funds have been dumping technology at the fastest pace in a decade and where the Magnificent Seven ETF is up 2% this year while the S&P is up 10.

[00:13:05]Now, the difference between the people who come out of this fine and the people who blow up their accounts is not who has the best thesis.

[00:13:13]Everyone has roughly the same thesis.

[00:13:15]The difference is who can read the market as it actually happens instead of arguing with it. So, the process is simple. Let the numbers print, strip out the noise like tonight's anthropic markup, look at capex, cash flow, and evidence of return, and then watch the reaction because the reaction tells you more than the number ever will. If great news gets sold again, that is a market that wants lower. If great news finally gets bought and holds, that is your first real signal the tide has turned.

[00:13:45]And none of that requires you to be a genius. It requires you to be patient, which is significantly harder when you think about it because being a genius is at least entertaining. Patience mostly looks like sitting on your hands doing nothing while your group chat loses its mind. So, the number is data, the reaction is the truth. So, here is where we are. Google tonight is the opening act, July 29th is the main event with Microsoft, Meta, and SK Hynix on the same day, and the whole thing gets stress tested through to August 5th. I'm going to be covering every one of these, the read through to Micron to Nvidia to the whole memory complex, and what each print actually means rather than what the headline says. So, make sure to subscribe, turn that bell on because those reports are landing back-to-back. They're going to move this sector hard in one direction or the other, and being 24 hours late to understanding what happened is genuinely the difference between reacting and being positioned. We do not guess here.

[00:14:46]We read the data, we watch the reaction, and we let the markets tell us when it is ready. Process over prediction, data and price action always. This is educational, non-financial advice.

[00:14:58]Always do your own research and manage your risk. Mr. Market.