# Intel’s Crash and Korea’s Record: Memory Is Revenue or It Is Cost

*URL:* https://millionaur.com/millionaur-2026-11-02-daily-analysis/
*Updated:* 2026-11-02T09:13:58+00:00

Friday was the worst US tech day since June. The Philadelphia Semiconductor Index fell 7.2% and Intel (INTC) had its worst session since 1974. On Monday morning Korea reported its October chip exports: 56.4 billion USD in a single month. Same week, opposite signals. Here is how I read them, and what the portfolio did.

## What broke: the edge

Intel’s data-centre revenue grew 78%. The problem was PCs. Management now expects the PC market to shrink about 12% this year and another 10% in 2027, and the CEO linked it directly to memory: conditions last “until memory availability and prices begin to normalize.”

That is the other side of the memory boom. When memory makers have pricing power, the buyers who treat memory as a **cost** (PCs, phones, consumer storage) lose volume. Kioxia, a consumer-heavy flash maker, fell almost 16% in Tokyo on Monday.

## What printed: the core

On the same weekend, Korea shipped 56.4 billion USD of chips in October, about half of all its exports. The +259% year-on-year headline is inflated by last year’s holiday calendar, so I read the level, not the growth rate. SK hynix, whose memory is sold to data centres under long contracts with customer deposits, rose almost 4% in Seoul while Kioxia fell.

My rule is simple: prefer companies where memory is the **revenue** line, avoid companies where memory is the **cost** line. Friday’s data made that rule stronger, not weaker.

## The funding side

Also on Friday, S&P kept Meta Platforms (META) at AA- but turned the outlook to negative, projecting a cash-flow deficit of about $71 billion in 2027. It is the first rating-agency action on a big AI buyer. Treasury yields fell, so the credit window narrowed but did not shut. My pre-written rule: if it shuts, I trim the borrowers first, not the suppliers they pay.

## What the portfolio did

- The trailing stops on the Korea and memory ladder fired at Friday’s US open, all in profit. Memory and Korea went from about 8% of the portfolio to about 4% before the worst of the chip sell-off.
- A lesson I am writing down: the first stop at the opening auction filled 2.6-3.7% below its level, the next one only 0.4-1.0%, intraday ones at the exact level. After a big Seoul drop, I will move the top stop down one rung before the US open.
- Later on Friday I used cash to rebuy and reset: Microsoft (MSFT), Amazon (AMZN), Micron (MU), Nebius (NBIS), Vistra (VST), a health-care biotech ETF, and a dozen clips re-opened at full size after the drop.
- Cash is about 16.6%, plus short inflation bonds, Treasuries and staples as hedges. Broad AI hardware is about 13.5% of the portfolio, well below my comfort range, so there is room to add on evidence.

## What I am watching this week

- Tonight: Apple (AAPL) results. Memory costs are the swing factor for its margin.
- Tuesday: Advanced Micro Devices (AMD) results and the US midterm elections.
- Friday: US jobs report.
- SK hynix: I would re-enter one clip only on a weekly close back above my old top stop level. A one-day rebound is not evidence.
- Oil: Iran diplomacy versus strike reports. My energy hedge has tight stops and I will not chase it.

## Risks, stated openly

- If memory prices stay this high, PC and phone volumes may keep falling in 2027, and data centres must absorb all new supply.
- If more AI buyers lose their credit ratings, funding could slow spending and hit the whole chain.
- Gold failed as a safe haven again on Friday. Treasuries did the hedging.
- Elections and geopolitics can gap prices through stop levels.

## The takeaway

Markets sell everything first and sort later. Friday sold the whole chain; Monday started sorting it: the edge loses, the core keeps shipping. My job is not to predict the next day. It is to keep cash and written rules ready, so that when the sorting happens, I am buying the right layer.

*Portfolio snapshot: Monday 2 November 2026, 08:57 UTC (US prices are Friday 30 October closes).*

Sources: [Intel PC outlook and memory comments](https://www.ibtimes.com.au/intel-shares-plunge-weak-pc-outlook-forecast-1877636); [Korea Times on KOSPI and October exports](https://www.koreatimes.co.kr/economy/20261102/kospi-rallies-over-2-as-investors-look-past-us-tech-rout); [S&P on Meta’s outlook](https://www.implicator.ai/s-p-downgraded-metas-outlook-stable-negative-citing-high/).

**Disclaimer:** This is not financial advice. Past performance does not guarantee future results, and copy trading carries a real risk of loss. Please invest only what you can afford to lose.

---

Human page: https://millionaur.com/millionaur-2026-11-02-daily-analysis/?vwao_human=1