VLSI MARKET / EVIDENCE GUIDE

From a stock signal to the evidence.

A stock signal summarizes the topic of a collected official company release. Read it alongside price changes, then test your next question against earnings, memory, fab and supply-chain evidence.

01

Is revenue growth improving the business?

Compare the same company's quarterly revenue, gross margin and operating margin. Keep fiscal periods, reporting currencies and segment boundaries visible. A revenue increase alone does not establish pricing power or cash generation.

02

Does stronger HBM demand mean every memory price rises?

HBM, commodity DRAM and NAND have different specifications and supply chains. Separate the product generation, density, spot or contract basis, and observation period. Use a shared industry index as context, never as the transaction price of a specific chip.

03

Has announced fab investment become usable capacity?

Follow the sequence from announcement and construction to equipment installation, qualification and production. Keep wafer diameter, process node and period together. Announced investment is neither operating capacity nor realized foundry revenue.

04

Do equipment orders confirm a broad industry recovery?

Separate orders, backlog, shipments and recognized revenue. Then compare industry production and utilization at compatible periods. One supplier or one technology node cannot establish a recovery across every semiconductor segment.

05

Is a partnership an actual supply contract?

A technology collaboration, customer qualification and commercial supply contract are different relationships. Read the original wording, parties, product and dates. Undisclosed order volumes or customer allocations remain unknown.

06

How should a semiconductor forecast be evaluated?

Inspect source inputs, publication lags, assumptions, model version and out-of-sample error. Compare with a no-change baseline and preserve uncertainty intervals. A conditional scenario is not an observed fact or an independently validated prediction.

METHOD / 2026-09-17.1

How are the signals made?

  1. Use official company releases from the monitored VLSI newsroom collection and dated filings already verified in the earnings and IR library. General media reports and research commentary do not enter this signal feed.
  2. Match an issuer name or a company alias in the headline to a registered listing. Mentions in another issuer's release are labeled separately.
  3. Select the newest eligible release for each listing. Classify headline terms in this priority order: earnings, capacity & investment, partnership, memory, product, then company update. The first matching rule wins.
  4. Releases within 14 days are recent; older releases retain an older-release label. Future or invalid publication dates are excluded. Collection time never replaces publication time.

An empty signal means no eligible match in monitored sources, not that the company has no news. Rule-based topic classification does not assess economic materiality, bullish or bearish direction, causes of price moves, or analyst opinion.

Quotes use the existing Yahoo Finance provider path, with native currency and market timestamps on each row. Volume belongs to the quoted listing. Price levels across currencies are not ranked; missing market caps and analyst opinions are not invented.

If a company match or publication date is wrong, send a correction with the original source. gc@vlsi.kr

By VLSI Korea · Classification method updated September 17, 2026. Each announcement retains its original publication date.