Tianfu Communication already released its third-quarter report on October 29, while Yuanjie Technology and Zhongji Innolight will not disclose until before October 31. Rank all three by single-quarter gross margin: optical chips, components, modules — which segment does this round of price pressure hit first?

Tianfu Communication already released its third-quarter report on October 29, while Yuanjie Technology and Zhongji Innolight will not disclose until before October 31. Rank all three by single-quarter gross margin: optical chips, components, modules — which segment does this round of price pressure hit first?

The premise of this take is only half right: Source Photonics' (源杰科技) Q3 report isn't out yet — the latest is the August 29 half-year report, plus the October 13 Q3 earnings preview; Zhongji Innolight's (中际旭创) Q3 report was already disclosed on October 30. Ranking the three companies' Q3 gross margins side by side isn't possible right now.

What can be done: use actual Q3 figures for TFC (天孚) and Zhongji, and actual Q2 figures for Source Photonics, paired with what it said in its own preview. Conclusion first — this round of price pressure hasn't landed on any segment's gross margin yet; if it does land, modules get hit first, then TFC's active devices, with optical chips last.

Single-quarter figures are calculated like this: the Q3 report only has the cumulative figure from the start of the year through September, so single-quarter equals cumulative minus the half-year report cumulative, with operating cost subtracted the same way; the Q1 report is used directly.

No trace of price cuts in the Q3 reports

Zhongji's Q3 single-quarter gross margin was 46.61%, up 0.19 points quarter-on-quarter and up 3.82 points year-on-year; TFC's was 65.05%, up 0.39 points quarter-on-quarter and up 11.37 points year-on-year. Both are moving up.

Source Photonics has no gross margin figure for Q3, but its preview gave Q3 revenue of 6.5 yi to 7.0 yi and net profit excluding non-recurring items of 4.0 yi to 4.5 yi. Using these two ranges, the midpoint of the net margin excluding non-recurring items is about 63.0% (after stripping out equity investment income and government subsidies), higher than Q2's 59.1%; its own wording is "gross margin remains at a high level."

Zhongji's October 29 conference call was more direct:

Third-quarter prices were generally relatively stable versus the first half... prices for the next year are generally re-quoted. At present we judge industry pricing to be fairly rational, with no vicious competition.

In TFC's October 28 record, investors asked about the same thing:

Against the backdrop of a rising share of active business, the company's third-quarter gross margin still remained relatively stable.

All three issued a "recent production and operating conditions" announcement on the same day, October 23; Zhongji's said "the company and the industry sector it operates in have seen relatively large stock price fluctuations," and TFC's said "the market has recently paid relatively close attention to the industry the company operates in and its production and operating conditions." All three announcements said orders and deliveries are normal, and none mentioned price cuts, lost orders, or customers cutting orders.

Why modules get squeezed first: the higher cost as a share of revenue, the more it hurts under the same price cut

Module makers' operating cost is 53.4% of revenue, mostly purchased chips and devices; for Source Photonics this item is only 17.9%. If selling prices fall 10% and purchase prices stay flat, the three companies' gross margins don't fall by the same order of magnitude.

It's mainly the module segment that re-quotes directly with cloud providers. Zhongji said "prices for the next year are generally re-quoted"; by that account, what's executed within 2026 is the already-negotiated price, and the new price starts from 2027.

Its upstream is still tight: third-quarter prepayments were 15.67 yi, 11.7 times the start of the year; inventory was 238.78 yi, 88.3% more than the start of the year. With money tied up in materials, it has no room to push costs down. DSP "capacity is somewhat tight, prices are relatively firm" — also its own words at the call.

Devices aren't one segment: passive 72%, active 47%

TFC's half-year report splits devices into two: passive revenue 15.30 yi, gross margin 71.90%; active 12.60 yi, 47.03%. A 25-point gap — two different businesses.

The active segment's gross margin is almost identical to Zhongji's optical modules (46.59%), and cost as a share of revenue is almost identical too. When module makers push prices up the chain, this is what hurts first, not passive. Calculating active alone on a 10% cut, it drops 5.9 points, the same order of magnitude as modules.

In Q3, TFC's active share was still rising and its blended gross margin still rose 0.39 points, showing this segment hasn't been squeezed yet. Its half-year report says "short supply of certain materials has had some impact on the company's capacity ramp for active optical device products" — that's the real variable for this segment this year; once materials loosen, revenue and gross margin come back together.

Which quarter will show it

Compress this chain into one sentence: the pressure comes from cloud providers' 2027 re-quoting, passes first through module makers, then up to active devices, and only last to chips. The first full quarter where it can be seen is 2027Q1; TFC's active devices lag by a quarter, because its customers are the module makers. Optical chips are the hardest to transmit: Source Photonics says "downstream customer demand is strong, visibility is very high," and its phase-two capacity won't take shape until end-2026 to early 2027.

What moves first is the unit price in operating revenue; operating cost doesn't come down with it, so the result lands on gross margin. If 2027 brings a demand reversal rather than price cuts, what breaks first isn't gross margin but inventory write-down provisions — Zhongji's 238.78 yi of inventory is the largest single item in this chain.

Where I could be wrong

The strongest counterargument: in 2027, 1.6T is still in shortage, module makers hold allocation power, and prices simply can't come down; instead, chip-side capacity is released in a concentrated way in 2027 — Source Photonics' phase two plus domestic peers' expansion — and chips loosen first. There's also a cost-side squeeze: CW and EML are tight, chip makers raise prices, TFC's active segment and module makers' material costs rise first, and gross margin falls from the cost side.

This can't be ruled out. But through Q3, all three companies' numbers are still moving up, and Source Photonics' net margin excluding non-recurring items hasn't pulled back.

Where I could be wrong: if in 2027Q1 Zhongji's single-quarter gross margin rises rather than falls, while Source Photonics or TFC's active segment drops by more than a point first, that would mean the price pressure started upstream, and this ordering would have to be reversed.

Sources: TFC Optical Communication "2026 Third Quarter Report" 2026-10-29 (Announcement 2026-064); TFC Optical Communication "2026 October 28 Investor Relations Activity Record Form" 2026-10-28 (No. 2026-012); TFC Optical Communication "2026 October 29 Investor Relations Activity Record Form" 2026-10-29 (No. 2026-013); TFC Optical Communication "2026 Semi-Annual Report" 2026-08-19; Zhongji Innolight "2026 Third Quarter Report" 2026-10-30 (Announcement 2026-101); Zhongji Innolight "Investor Relations Activity Record Form 20261029" 2026-10-29 (No. 2026-009); Zhongji Innolight "Announcement on the Company's Recent Production and Operations" 2026-10-23 (Announcement 2026-100); Zhongji Innolight "2026 Semi-Annual Report" 2026-08-22; Yuanjie Technology "Voluntary Disclosure Announcement of the 2026 Third Quarter Earnings Forecast" 2026-10-13 (Announcement 2026-051); Yuanjie Technology "Announcement on the Company's Recent Production and Operations" 2026-10-23 (Announcement 2026-057); Yuanjie Technology "2026 Semi-Annual Report" 2026-08-29.