# Millionaur Daily Analysis — 2 October 2026: Washington Just Repriced the Laser

*URL:* https://millionaur.com/millionaur-2026-10-02-daily-analysis/
*Updated:* 2026-10-02T08:41:01+00:00

**On a flat day for the indices, one corner of the portfolio jumped: optics.** Coherent (COHR) gained more than 10% and Lumentum (LITE) almost 8%.

Three things landed together. A bipartisan group of US senators introduced a bill to keep Chinese-made optical transceivers out of sensitive federal systems. A large Wall Street research house started coverage of the sector with positive ratings. And Coherent launched a new integrated-optics platform, citing more than ten customer engagements for co-packaged optics.

## Why this matters (and what it is not)

Every AI data centre moves its data as light. The scarce part is not the module, it is the laser inside it and the indium phosphide wafer the laser is grown on.

Per recent analyst commentary, any broader US restriction would likely start with the next generation of modules, not the ones shipping today. So this is not an order shift next quarter. It is a slow tightening of an already tight laser supply, which is a pricing story with a long fuse.

That is why I own the layers rather than one name: lasers and modules through Coherent and Applied Optoelectronics (AAOI), the wafer through Sumitomo Electric, test equipment through Viavi (VIAV), glass and fibre through Corning (GLW), and a photonics ETF on top. Together about 5.6% of the portfolio.

What I am not doing: buying more after a 10% day. A policy headline is a reason to protect a position, not to chase it.

## Two corrections to yesterday’s post

1. I wrote that Korea “sold” the Micron (MU) report. That was true in the morning. By the close Seoul had reversed and finished about 2% higher, helped by record September exports: $120.9bn in total, with chips at $60.3bn, up 263% on the year.
2. My headline said Micron told us when the shortage ends. It did not. Micron said it cannot yet see when supply and demand rebalance. Late 2028 is when extra cleanroom space becomes available, and some other capacity arrives in 2027. The thesis stands, the wording was too strong, and you deserve the precise version.

## What changed in the portfolio

- The memory rule was executed: one more position in the Roundhill Memory ETF (DRAM) at the US open, as written before the report. Memory and Korea are now about 10% of the portfolio and closed for new adds.
- New position: Corning, one small clip, bought before the optics move.
- One stop fired: X-FAB closed with about +7% profit, filled almost exactly at the stop level.
- New protection: stops now sit under the Japanese component makers (TDK, Murata, Taiyo Yuden, Sumitomo Electric) after gains of 13 to 15%, and under ON Semiconductor (ON). Several existing stops were raised.
- Cash is about 13%. Together with short-term inflation-protected bonds, roughly 16% of the portfolio is built for a world of high yields.

## What I am watching

- The US jobs report today. Expectations are for roughly 98,000 new jobs and 4.1% unemployment. Wages matter most: factory input prices just rose much faster than expected and oil is back above $100 a barrel.
- The US 10-year yield, which touched 5.34% yesterday, the highest since 2002, and a dollar at a 17-month high. Both pressure gold miners, copper and uranium, which are the weak part of the portfolio right now.
- Broadcom (AVGO), down after news that it will lend up to $42bn to an AI lab to lease its chips. At about 20 times forward earnings it is the networking name I would rather add to than the ones that just jumped.
- Samsung’s preliminary results, expected next week.

## Risks, stated openly

A bill is not a law, and optics stocks are expensive after this move. Western laser makers cannot replace Chinese supply quickly. Lower-quality corporate bond yields have risen sharply in five weeks, and AI spending is increasingly financed with debt. If credit tightens, the whole AI supply chain feels it, including this portfolio. What would change my view on optics: restrictions that hit today’s products and disrupt data-centre builds, or signs that cloud companies are slowing orders.

## The lesson

Good days test discipline as much as bad ones. Own the bottleneck before the headline, protect it after, and correct yourself in public when the wording was wrong. Steady beats exciting.

*Portfolio snapshot: 2 October 2026, 06:06 UTC. US lines carry the prior close.*

Sources: [Benzinga on the optical transceiver bill](https://www.benzinga.com/trading-ideas/movers/26/10/62117250/lumentum-coherent-rip-higher-as-washington-targets-chinese-optical-transceivers), [The Korea Times on September exports](https://www.koreatimes.co.kr/economy/20261001/koreas-sept-exports-hit-record-1209-bil-on-robust-chip-sales), [Micron prepared remarks](https://s25.q4cdn.com/621799436/files/doc_financials/2026/q4/Q4-FY26-Prepared-Remarks.pdf).

This is not financial advice. Investing and copy trading involve real risk of loss, and past performance does not guarantee future results.

---

Human page: https://millionaur.com/millionaur-2026-10-02-daily-analysis/?vwao_human=1