# Zero of Eight: Why the Week&#8217;s Buy Orders Wait for a Reversal Day

*URL:* https://millionaur.com/millionaur-2026-11-22-daily-analysis/
*Updated:* 2026-11-22T10:14:27+00:00

**Sunday outlook.** After a week in which the S&P 500 fell 8.6% and the chip index 17%, the most useful thing I can do is decide now what I will do on each outcome, so that Monday’s headlines do not decide for me. Earlier posts: [Saturday’s week in review](https://millionaur.com/millionaur-2026-11-21-daily-analysis/) and [the same review on eToro](https://www.etoro.com/posts/e1ea6b00-e58a-11f1-8080-80003f1d7899).

## Where the portfolio stands

- About 40% in cash. No trades since Friday evening.
- Chips and AI hardware are a small slice, well below my usual range. That is deliberate.
- Utilities, staples and Treasuries did the hedging last week. Gold and silver did not: they were sold with everything else.

## Has the turn started? My 8-point checklist says no

Every Sunday I count eight signs that a sell-off is ending. Today the count is zero, with one unknown.

- Volatility is in panic mode: the VIX (35.9) sits above its 3-month version (30.9), a sign of near-term fear.
- High-yield bonds are still drifting lower.
- No breadth thrust, the dollar is still strong (DXY 103.8), and bitcoin ETFs were still seeing outflows.
- Friday’s chip sell-off closed near the lows, not with a reversal.

That does not mean “sell”. It means every new entry waits for a daily turn inside its zone.

## The week and my branches

- **Monday, Asia.** The first session after Friday’s chip rout. If Taiwan and Korea open lower but close well off the lows on heavy volume, that is the exhaustion I want to see. Then [$TSM](https://finance.yahoo.com/q?s=TSM) gets one position on a daily reversal in its 295–315 zone (around its 100-week average), during the US afternoon, never in the opening auction. A weekly close below about 285 cancels the idea.
- **Monday, US.** Two pre-written moves: trim one silver position ([$SLV](https://finance.yahoo.com/q?s=SLV); the weekly rule fired on Friday), and swap one inflation-linked bond position ([$VTIP](https://finance.yahoo.com/q?s=VTIP)) for one 7–10 year Treasury position ([$IEF](https://finance.yahoo.com/q?s=IEF)). Last week the risk was growth, not inflation, and Treasuries were what actually hedged.
- **Power equipment.** Same rule: [$ETN](https://finance.yahoo.com/q?s=ETN) sits on its 200-week average (about 307.5). A daily reversal earns one position; a weekly close below it ends the idea.
- **Wednesday.** US inflation (PCE), GDP and orders, all bunched before Thanksgiving. If core inflation runs hot and the 10-year yield closes a day above 5.10%, I pause all new AI buys.
- **Thursday.** US markets closed for Thanksgiving.
- **Friday.** Half day. Oil’s weekly close decides one commodity trim.
- **Bitcoin.** [$BTC](https://finance.yahoo.com/q?s=BTC) is near 60,000, but tonight’s weekly close will likely be lower than last week’s. My rule buys the first recovery sign, not a price: a capitulation reversal day, a higher weekly close, or a first week of ETF inflows. None of those yet.
- **[$NVDA](https://finance.yahoo.com/q?s=NVDA)** stays on hold until its quarterly filing (around 4 December) answers the accounting question reported on Friday.

## The long-term trend: AI infrastructure 2027–2030

**The level keeps rising.** McKinsey expects global data-centre demand to grow about 22% a year to roughly 220 gigawatts by 2030, about 70% of it AI.

**The growth rate does not.** Big-tech capex is expected near 1–1.2 trillion dollars in 2027, but on Goldman Sachs’ path its growth slows from about 100% this year to about 54% in 2027 and 12% in 2028. Stocks trade the growth rate.

So I sort the chain by which revenue survives a slowdown:

- **Power and grid:** paid per gigawatt connected, with multi-year order books.
- **Foundry and lithography:** contracted capacity into 2027.
- **Memory:** real, but the most cyclical layer. Small and disciplined.
- **The buyers (the big platforms):** slower capex growth means more free cash flow. A 2027 idea, not a today idea.

How I build it: in steps over 2027, on weekly zones and daily turns, not on breakouts.

## The next 12 months, three paths

1. **Last hike or two, then a peak in yields (about 40%).** Cash goes back to work in thirds around the second hike.
2. **Higher for longer (about 25%).** Cash stays high, only contracted-revenue companies.
3. **Growth scare (about 35%, up from 25%).** The 10-year fell from 5.12% to 4.91% in a week while stocks fell. More Treasuries, then quality on the turn.

## Risks, openly

A disorderly Monday in Asia, a hot inflation print, and stop-losses close to price on my utility positions. If those fire, the cash goes to Treasuries first. Stops are meant to fire; I do not move them to dodge a fill.

## The process

The hardest part of a week like this is not the losses. It is staying patient while every chart looks cheap. The plan is written; the market now has to show its hand.

*Snapshot: Friday 20 November 2026 closes; portfolio data Sunday 22 November 09:50 UTC; bitcoin as of Sunday morning.*

Sources: [McKinsey, the cost of compute](https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-cost-of-compute-a-7-trillion-dollar-race-to-scale-data-centers) · [McKinsey, AI power and data-centre capacity](https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/ai-power-expanding-data-center-capacity-to-meet-growing-demand) · [CNBC on Nvidia’s losing streak](https://cnbc.com/2026/11/20/nvidias-7-week-losing-streak-is-longest-since-2008-financial-crisis.html) · [The Block on bitcoin ETF outflows](https://theblock.co/news/markets/2026-11-14-bitcoin-etfs-lose-1-billion-for-the-second-week-in-a-row-as-crypto-funds-except-one-continue-bleeding-440337)

*This is not financial advice. Investing and copy trading involve real risk of loss, and past performance does not guarantee future results. Always do your own research.*

---

Human page: https://millionaur.com/millionaur-2026-11-22-daily-analysis/?vwao_human=1