Pre-Market Briefing on Oct 12: Lumentum +5.22% on Optics Sold Out to Early 2029; TSMC Reports This Week

All three major indexes closed higher on Friday, October 9: the Dow at 51,654.95 for a 0.83% gain, the Nasdaq at 27,366.17 for 0.64% and the S&P 500 at 7,811.54 for 0.59%, taking back Thursday's selloff on AI valuation concerns. Rates, though, reached their highest since April 2002: the 10-year closed at 5.27%, up 3.8 basis points, after touching 5.3645% intraday, while WTI settled at $91.85. The stock in focus is Lumentum (LITE), which closed at $1,103.36 for a 5.22% gain, adding about $4.3B of market value to reach $85.8B and sitting just 3.5% below its 52-week high. The driver is the CEO saying publicly that optical components are effectively sold out through early 2029; six months ago the same line said 2028, so the visibility window slid about a year in half a year. The company gave only that date, and the order value it cannot take plus the new capacity figures are left for later results and calls. Today, October 12, there is no US economic data and no major earnings, with the bond market closed for Columbus Day while equities trade as usual; the main event of the week is TSMC's call before Thursday's open. The data in this article is as of the October 9 US close, and all times are UTC.

Unit: %, single-day change on October 9; all three rose, so the bars are drawn from zero. The Dow gained the most at +0.83%, with the Nasdaq at +0.64% and the S&P 500 at +0.59%.

 

Three indexes moving the same way by similar amounts says this was a whole-market rebound rather than one sector's story. Thursday fell on AI valuation concerns; on Friday oil steadied and technology rebounded. The same set of worries was priced twice in two days, in opposite directions.

 

When indexes move together by similar amounts, look for the cause one level up. Three indexes with completely different compositions moving as one usually means the driver sits in what they share: risk appetite, or rates.

 

Rates, however, went the other way. The 10-year closed at 5.27%, up 3.8 basis points, after touching 5.3645% intraday, the highest since April 2002. TSMC (TSM) closed at $453.31, down 1.02%, falling on a day the Nasdaq rose.

 

A day when both stocks and yields rise deserves a question mark: is this conviction, or just a bounce off the previous day's fall? With the discount rate above 5%, anything whose valuation rests on distant cash flows reacts harder to any revision in the revenue outlook.

 

Unit: %. The card sets the basis for these six: optical networking and high-speed connectivity names grouped by hand. Ciena sits at one end on +5.60% and Astera Labs at the other on −1.39%, 6.99 points apart.

 

🔴 One thing first: the stock in focus was not the biggest gainer in this set. Ciena rose 5.60%, 0.38 points more than Lumentum's 5.22%. Four of the six rose and two fell, averaging +2.69%, while the Nasdaq communications equipment field they belong to read −0.18% that day. A negative field average with a large individual gain is moving against the group; but three other names in the set also rose, so this was not a solo move. The card's own conclusion puts it best: what moved was this group, not the whole industry.

 

Moving against the group and moving alone are two different claims with two different tests. The first needs only a negative field average; the second needs no second name rising. Skip the second question and a group's move gets written up as one company's story.

 

The dividing line is clear too: Fabrinet in contract manufacturing fell 0.22% and Astera Labs in connectivity chips fell 1.39%, neither joining in. What rose was the optical networking leg. The tightness that day sat in the components themselves, not in the firm that builds them and not in the chips beside them.

 

What moved Lumentum is the CEO saying publicly that optical components are effectively sold out through early 2029; six months ago the same line said 2028, so the visibility window slid about a year in half a year, while new capacity takes three years to reach customers. ⚠️ The company gave only that date; the order value it cannot take and the new capacity figures are left for later results and calls.

 

Unit: 0–100, against peers and its own one-year history, measured over a one-year window rather than the single session. Trend position is highest at 96.0, with peer rank at 80.0, relative strength 77.0 and volatility control 58.6, while valuation heat is lowest at 26.3.

 

The five axes are far apart, and the highest and the lowest are not describing the same layer. Trend position at 96 says the price already stands at the top of its own one-year range: it is just 3.5% below the 52-week high. Valuation heat at 26.3 says the industry field it sits in is not priced hot. One is a price position, the other a valuation level, and both can be true at once.

 

Read a radar by asking what each axis measures before deciding whether they can be added up. Standing high on price while the industry is not expensive is the complete reading; either line on its own gives a skewed one.

 

One more detail worth noting: volume ran only 1.07 times its average, so turnover did not visibly pick up. The supporting name is Credo Technology (CRDO), up 2.21%, which makes the high-speed cables and SerDes chips between servers and switches; Nasdaq semiconductors averaged −2.04% the same session, so it was one of the few that rose.

 

 

Today, Monday, brings no US economic data and no major earnings, with the bond market closed for Columbus Day while equities trade as usual. Two dates matter this week: September CPI and core CPI at 12:30 UTC on Wednesday, and TSMC's results call at 06:00 UTC on Thursday, before the US open, running about 90 minutes.

 

TSMC is worth taking apart early because it is the only large-scale source of leading-edge foundry capacity, with the newest chips from Nvidia, Apple and AMD all on its lines. Its margin tells you how expensive leading-edge production is right now, and its capital spending prices the whole AI compute chain.

 

But revenue is already in before Thursday: TSMC reports monthly, so the three months of the third quarter add up to NT$1,494.2B, up 50.9% year on year (25Q3 was NT$989.9B), a quarterly record. At the 32 New Taiwan dollars per US dollar the company's own guidance assumes, that is about $46.7B, already above its own $45.8B guidance ceiling.

 

Unit: %, the 26Q2 share of wafer revenue by node, not of total revenue. 5nm is 33%, 3nm 30%, 10nm and above 23%, 7nm 11% and 2nm 3%; 7nm and finer together make up 77%, the bulk of the revenue.

 

The real variable hides in the smallest bar. 2nm is only 3% of wafer revenue, yet it is the one node now ramping. A new node carries the heaviest depreciation and the lowest yield in its first year of volume production, and that is where the company's 26Q3 margin guide of 65% to 67%, below last quarter's 67.7%, comes from. ⚠️ That 65% to 67% is the company's own guidance, not realised results.

 

A new line worth 3% of revenue can set the margin for the whole company. A revenue share measures how much is being earned now; depreciation and yield measure how much is being spent now. The two do not move together, and the market usually prices the second.

 

So Thursday reads in three steps. First, revenue is known and is not the event. Second, watch whether the margin holds 65% to 67%, with the gap coming from that 3%. Third, read the 2027 capital spending basis: with the 10-year at 5.27%, the market is shifting from how fast AI capex grows to how credible it is. The same logic applies to today's stock in focus: Lumentum's order visibility now reaches early 2029, but new capacity takes three years to arrive, and the margin is what answers the gap in between.

 

A: No. Ciena rose 5.60%, 0.38 points more. Four of the six rose and two fell, averaging +2.69%.

 

A: Against the group, but not alone. Moving against the group needs a negative field average, and Nasdaq communications equipment read −0.18% while this name rose 5.22%, 5.40 points ahead. Moving alone needs no second name rising, and three others in the set rose. The card puts it best: what moved was this group, not the whole industry.

 

A: No. It is a statement about capacity visibility, not money already booked. Whether it turns into profit depends on three things: whether the date keeps sliding out, when new capacity arrives (a new laser plant takes about three years to ship), and whether the margin follows.

 

A: The margin and the capital spending. Revenue is disclosed monthly, so the third quarter landed at NT$1,494.2B, up 50.9%, about $46.7B and above the company's own guidance ceiling. What is not public is the margin: the company guided 65% to 67%, under last quarter's 67.7%.

 

A: Because a new node carries the heaviest depreciation and the lowest yield in its first year of volume production. A revenue share measures how much is being earned; depreciation and yield measure how much is being spent. They do not move together, which is how a 3% node pulls the whole company's margin guide down.

Disclaimer: This article is compiled and written by the MEXC RealStocks team. The data in this article is based on the closing of the US stock market on October 9, 2026. The content is a compilation of public market information, and individual stocks are publicly discussed targets, which do not represent the recommendation or opinion of MEXC and do not constitute any investment advice. More US stock content: @MEXC | @Alpha_MEXC | @MEXC_Research