Yesterday’s letter argued that the market’s problem right now isn’t growth, it’s the price of money, and Story 3 made the point through Broadcom: the numbers were excellent and the stock got sold anyway, because excellent wasn’t what the price already assumed. Amazon is the cleanest test of that idea we can find, because it just did the same thing in slow motion — a record high in August, a rejection off it, and a stock that has spent the whole year going nowhere versus an index it beat by nothing.
We’re writing it up because the chart answers the only question worth asking on this name right now: is this a pullback you buy, or a ceiling you wait out? Our honest answer is neither, not here. There is a level where we’d want to own it, and Friday’s close is about seven percent above it.
Where it closed
AMZN finished Friday at 258.51, down 0.15% on the day. It sits between the Bollinger Bands, under the 20-day, well above the 200-day. Four levels frame the whole picture.
Two bits of arithmetic worth having. The band is 22.29 wide (273.34 − 251.05), which is 8.50% of the middle band (22.29 / 262.20) and 8.62% of Friday’s close. That is a wide envelope for a mega-cap, and wide bands mean the last twenty sessions have not been quiet. And price sits 19.39 above the 200-day, which is 8.11% above the average itself (19.39 / 239.12) or 7.50% of the close, depending which base you measure from. Both are the same gap. We’ll use the 8.1% figure when we’re talking about how stretched the stock is above its trend, because the trend line is the thing you’re measuring against.
RSI is 49.55. That is the midline to two decimal places. It tells you nothing, which is itself the finding: the stock is neither oversold enough to bounce on relief nor overbought enough to be worth fading. The same panel had RSI at the 70 line in April and May and again briefly in July and August, and down near 25 in the February and March selloff. It has unwound all of it. There’s no edge in the oscillator here, and anyone telling you otherwise is reading a different chart.
What the daily chart is actually saying
This is a strong uptrend that has stalled, not broken. Over the three-year window the 200-day has risen steadily from roughly 110 to 239.12 and has never turned lower, and that’s the fact that outranks everything tactical below it. Price has dipped under it twice in that span and both times came back.
Inside that trend the last eighteen months have been violent. Price sold off through the 200-day in the spring of 2025 to around 165, the deepest damage on the chart. It recovered to roughly 240 by that autumn, dropped hard in February 2026 to about 190 where it undercut the 200-day again, and then ran almost vertically from April into May to roughly 280 to 285, riding the upper band the whole way. June gave a chunk back to about 230, where the 200-day caught it. August pushed to roughly 285 a second time and faded.
So the stock has tested the same ceiling twice in four months and failed at it twice. That is what the fade off 273.34 and the close under the 20-day is describing in miniature.
The monthly chart is the one that matters
Pull back to monthly candles and the August bar does something specific: it printed the all-time high near 285 and closed well off it, leaving an upper wick on the highest monthly bar in the stock’s history. September so far is a small red candle underneath it.
A rejection wick on a record high is not a top. Plenty of them get eaten a month later. But it is the market telling you where the sellers live, and it is the second time this year that roughly 285 has done the job.
Now look down, because that’s the part we care about. Roughly 240 is where the stock held in December 2025 and again in mid-2026. And the daily 200-day is at 239.12, right underneath it and rising.
That confluence at 239 to 240 is the whole letter. A horizontal shelf that held twice and a rising long-term average arriving at the same number is the highest-quality support this chart has to offer, and it’s the only level on the page we’d want to be a size buyer against. Everything between 240 and 285 is the market arguing with itself, and Friday’s 258.51 is roughly in the middle of that argument.
What the relative chart adds, and why it matters for premium
Tony’s second chart plots AMZN against SPY year to date. AMZN is up 10.92% on the year and SPY’s line finishes marginally ahead of it. After eight months, all that work for a dead heat.
But the path was not a dead heat. By eye off that chart: AMZN gapped down in early February to about −10%, bottomed near −15% in late February, bottomed again with the index in early April, then overtook SPY and peaked near +18% in mid-May. June and July gave the entire lead back to roughly −3%. Then in early August it gapped from about −3% to +22% in a matter of days, the single biggest move on the chart, and faded through the rest of the month.
That’s a swing of about 33 points versus the index between the February low and the May high (18 − (−15)), round-tripped to level. The beta on the quote card is 1.48.
For anyone who sells premium, this is the section to read twice. A 1.48 beta and two gap events this year is not a quiet stock, whatever the last few weeks of drift look like. What the path has actually done is the thing to size against. We have no chain in front of us, so we are not going to tell you what AMZN options are priced at today. What we can tell you is that this name swung 33 points against the index and gapped twice this year, in February and again in August, and a stock that has done that once can do it again on a Tuesday. The correct response to a record like that is smaller size than the regime baseline would otherwise allow, and better strikes, not more contracts.
What the fundamentals add, briefly
We’re not going to relitigate the quarter. Three things from the July 29 call matter to the chart.
AWS is accelerating, not decelerating. Revenue of $42.2 billion in Q2, up 36.7% year over year, a $169 billion annualized run rate and a $496 billion backlog. The run rate ties to the quarter: 42.2 × 4 = 168.8. Company-wide, revenue was $200.6 billion, up 20%, with operating income of $27.5 billion, up 43%.
The bill for it went up. Management raised 2026 cash CapEx to roughly $220 billion from roughly $200 billion, spent $53.1 billion of it in Q2 alone, and told everyone plainly that heavy near-term spending will pressure free cash flow until the data centers are earning. The platform’s model rating flags negative trailing-twelve-month free cash flow. We weren’t given a free-cash-flow figure, so we’re not printing one.
The stock is priced somewhere around 20 to 21 times earnings. The quote card says 20.5; the model’s own write-up says roughly 21.6. We’re giving you both rather than picking the one that suits the argument.
That is the Broadcom problem wearing a different suit. The growth is real, the spending to produce it is enormous, and the only open question is whether the price already assumes it works. A model on Tony’s platform reiterated a $311 target on August 1 with an Outperform score of 79. That is a screener model, not a person and not Wall Street, and it’s five weeks old. One more piece of arithmetic: 311 / 1.2030 = 258.52, so the 20.30% upside it advertises is being measured off roughly Friday’s price.
And the single most useful fact for anyone structuring an option trade: the next earnings report is October 21, 2026.
The two scenarios
Up. Reclaim 262.20 and hold above it, and the 20-day stops being resistance and starts being a floor. The next stop is the upper band at 273.34, 5.74% above Friday, and then the roughly 285 high, about 10% up. Getting through 285 on the third attempt would be a genuine breakout on the monthly chart, not a bounce.
Down. Lose 251.05, the lower band, and there is very little between there and the shelf. Roughly 240 and the 200-day at 239.12 are where the buyers showed up twice before, about 7% below Friday’s close. That is where we’d rather be a buyer, and we’d say that out loud rather than pretend we like it here.
Everything between 251.05 and 262.20 is noise, and Friday closed inside it.
How a GYP member would express this
This is analysis, not an alert. We are not opening anything in Portfolio 1, 2 or 3 on the back of it, there is no trade card below, and no Greeks changed today. If we do put something on, it comes as its own alert with the fills, like everything else we publish.
Here is how the framework reads it.
Sell into the shelf, don’t chase the middle. The whole method is to get paid to be a buyer at a price you actually want, and 258.51 is not that price. The 239 to 240 confluence is. On our own regime table, the VIX at Friday’s 14.58 close (per this morning’s letter) sits in the normal bucket, not the low-volatility one. Normal means standard sizing and the ordinary 45-to-21-day cycle. It does not come with the extra distance the low-volatility rule demands. The shelf sits roughly 7% below Friday’s close, and that figure is measured off a level read by eye. So this isn’t two independent rules agreeing. It’s a level worth waiting for that happens to sit right at the edge of the band the low-volatility rule would ask for if we were in it, and we aren’t. Without a chain in front of us we can’t say what delta that shelf is. What we can measure is distance, and the rulebook screens on that as well.
The cycle matters more than the strike here. Our entries go on around 45 days to expiration and come off at 21 days, no exceptions. From Friday, the October monthly expiration on October 16 is 42 days out, which is right in that window, and it lands five days before the October 21 earnings print. That is a full cycle of premium with no earnings event inside it. The September monthly on September 18 is only 14 days out, too short by our own rules, and it carries CPI on Friday the 11th and the Fed on the 15th and 16th. Short of earnings does not mean short of risk.
Into CPI week, prefer defined risk. This morning’s letter said we’d rather be a buyer of cheap volatility than a seller of it ahead of the number, and nothing about this chart changes that. If you want AMZN exposure before Friday, a put spread caps what a gap can do to you, and this stock has produced two gaps this year that a naked short strike would have felt. The cost is a smaller credit. That is the right trade when volatility is ordinary and the calendar is not.
On what a short put actually ties up. We don’t have a per-name margin figure for AMZN from Tony’s platform, so we’re not going to make one up. What we can tell you is the arithmetic on the other side: in a cash account, a short put at a 240 strike requires 240 × 100 = $24,000 of margin required on cash account per contract, which is also the sizing number and close to the worst case. In a margin account the requirement is a fraction of that. Those are two different numbers for the same trade, and knowing which one your account uses is the difference between holding through a bad week and being forced out of one. The 240 strike here is hypothetical, chosen to illustrate the level. We have no options chain, no implied volatility and no premium data for AMZN in front of us, so there is no credit quoted anywhere in this letter and there won’t be one until there’s a fill.
Trader Take
No edge at 258.51, real edge at 240. RSI at the midline, price between the bands, the 20-day 1.43% overhead and the lower band 2.89% underneath. That is a coin flip with a spread on it. The reason to have this chart open is that the stock sits roughly 7% above the best support it owns, and waiting for that level is the trade. Reclaiming 262.20 changes the near-term picture; losing 251.05 starts the trip to the shelf. Sizing beats cleverness on a 1.48-beta name that has gapped twice this year.
Investor Take
AWS accelerating into a $220 billion capital plan is the entire investment case and the entire risk, and they are the same sentence. Amazon is up 10.92% year to date and has still not beaten the index it round-tripped 33 points against. If you own it, you have already been paid nothing for a great deal of turbulence, and that’s before you ask whether roughly 20 to 21 times earnings is right for a company deliberately suppressing its own free cash flow to build data centers.
That is the connection to this morning’s letter. When the price of money goes up, the market discounts far-off cash flows harder, and Amazon is now a company whose cash flows are further off than they used to be by management’s own choice. Nothing about that says sell. It says the entry price matters more than it did two years ago, and 240 is a materially better entry than 258. The 200-day has risen from roughly 110 to 239 over three years and has not turned down once. Let it come to you.
Disclaimer
This is not investment advice. Nothing in this letter is a recommendation to buy or sell any security. Grow Your Pile and Squared T Capital publish what we do in our own accounts for education. Options involve substantial risk and are not suitable for every investor. A short put can require you to buy stock at the strike price and can lose more than the premium collected. Past results do not predict future results. Do your own work and consider speaking with a licensed advisor about your circumstances.
Past performance is not indicative of future results, and results shown are those of our own accounts and are not representative of any subscriber’s results. Read the Characteristics and Risks of Standardized Options before trading.
Technical levels quoted here are drawn from single charts at a single point in time and are not predictive. They are as of the Friday September 4, 2026 close and will move.
Tony Battista and Tony Rihan Grow Your Pile
Every options trade in all three portfolios is published, win or lose, with entry, exit and running P&L at growyourpile.com.