Dark pool / off-exchange activity — Tower Semiconductor

As of: 2026-04-29.

Confidence legend: ✓ verified-primary FINRA · ◐ aggregator-sourced · ⚠ inferred / estimate / unavailable

1. Off-exchange / ATS context

US equity trading is bifurcated between:

Lit exchanges: NASDAQ (TSEM’s primary listing), NYSE, NYSE American, BATS BZX/BYX, IEX, MEMX, MIAX (display orders, public price discovery).
Off-exchange / Dark venues: Alternative Trading Systems (ATSs) registered with FINRA, internalized retail flow at wholesalers (Citadel Securities, Virtu, Susquehanna, Two Sigma), and over-the-counter (OTC) bilateral block-trades.

For most actively-traded NASDAQ-listed mid-caps, off-exchange / dark-venue activity is approximately 35-45% of consolidated volume. TSEM, as a $21.4B mkt cap dispersed-ownership name with ~1.5M average daily share volume, fits the profile of typical NASDAQ-mid-cap routing patterns.

2. Dark-venue summary — pipeline status

The KB’s companies/tsem/data/STOCK_DARKPOOL.json is currently empty ({}) ✓ verified. When populated, it would contain:

Weekly FINRA ATS-reported share counts (FINRA disseminates weekly aggregate ATS volume by venue + symbol)
Top venue breakdown (UBS ATS, MS POOL / Morgan Stanley ATS, Credit Suisse Crossfinder [legacy], JPM-X, Barclays LX, Goldman SIGMA-X, etc.)
Total off-exchange % of consolidated volume
Anomaly detection (single-venue spikes signaling block-trading activity)

Status: This file documents the framework; specific FINRA ATS-reported numbers for TSEM are queued for ingestion in a future refresh pass once the data pipeline is wired up. The same gap exists for GFS — see ../../gfs/kb/06_market_data/darkpool_offexchange for the analogous treatment.

FINRA primary-source URL: https://otctransparency.finra.org/otctransparency/AtsIssueData (queries by ticker yield weekly-resolution share counts by ATS venue).

3. Inferred / estimated dark-venue characteristics

For TSEM specifically:

Metric	Estimate ⚠	Comment
Estimated off-exchange %	30-40% of consolidated volume	Consistent with NASDAQ mid-cap norms; possibly slightly elevated due to Israeli FPI ordinary-share routing nuances
Estimated retail-internalization %	10-18% of consolidated volume	Citadel / Virtu wholesalers handle retail order flow; lower than typical retail-heavy small-cap names
Estimated institutional-dark-pool %	18-25% of consolidated volume	UBS ATS, MS POOL, Goldman SIGMA-X likely the top venues for TSEM block flow
Block-trade frequency	Moderate	Dispersed-ownership profile (no controlling shareholder) means episodic-block-trading is less concentrated than for GFS (Mubadala secondaries)
Pension-fund cohort routing	Likely high-touch / agency execution	Israeli pension funds (Migdal, Harel, Menora, Phoenix, Clal) tend to use agency execution via large-cap brokers — Goldman / Morgan Stanley / Citi — for cross-border trade execution

4. The Israeli-pension-cohort flow signature

A distinctive feature of TSEM’s flow profile is the Israeli-pension-cohort dominance (~32% of outstanding held by Migdal + Harel + Menora + Phoenix + Clal — see ./institutional_holders). These holders typically:

Execute via agency block-trades rather than algorithmic small-order routing
Concentrate trading in NASDAQ pre-open / post-close auction sessions for cross-border timing alignment with Tel Aviv business hours (TA-USNYC time-zone offset is 7-8 hours)
Use VWAP / TWAP execution algos for large positions, often crossing in dark venues to minimize signaling

Read. This implies TSEM dark-venue activity may show:

Higher-than-typical block-trade-cluster signatures during Tel Aviv business hours overlap (~9:30 AM - 11:30 AM ET)

Modest tail in extended-hours / overnight auction crosses

Lower-than-typical retail-internalization % (since Israeli pension funds are not retail-platform-routed)

5. What dark-venue analysis would reveal

Signal	What it tells us
Dark % of consolidated volume rising	Institutional flow accumulating; retail % falling (or large-block executions clustering)
Single-venue concentration spike (e.g., MS POOL or Goldman SIGMA-X)	Specific institutional routing event; possibly signals a Migdal / Harel / Menora rebalancing or a US-active-fund position change
Wholesaler internalization % rising	Retail flow expanding (often around catalyst events — earnings, AI-narrative inflections)
Volume vs. OI signal	High dark-volume + low option-OI change suggests stock-only positioning rather than options-overlay
Anomalous extended-hours volume	Cross-border large-block clearance (Israeli-cohort timing)

6. Specific events worth monitoring via dark-venue lens

The following 2025-2026 catalysts are likely to have produced detectable dark-venue volume signatures:

Date	Event	Expected dark-venue signature ⚠
2025-11-12	Tower CPO foundry technology press; AI-photonics premium ignites	Dark-volume spike from US institutional accumulation; pension-cohort begins building incremental positions
2025-11-14 to 2025-12-15	Short-interest acceleration from 1.80% → 2.67%	Dark-venue activity from short-side institutional accumulation; cover-side from longs
2026-01-27	-16.9% single-day drawdown	Block-execution dark venue activity from institutional rebalancing
2026-03-11	LWLG-Tower PH18 development agreement announcement	Dark-volume spike on US-hedge-fund positioning
2026-03-19	+17% single-day rally on 7.94M shares (largest move since Intel-deal-announcement)	Significant lit + dark + retail-internalization flow; would be the cleanest single-day study sample
2026-04-17 / 2026-04-20	All-time high $226.45 close / $228.73 intraday	Likely pension-cohort distribution + retail-momentum-buying; classic top signature
2026-04-21 to 2026-04-29	-15% cooldown to $191.54	Profit-taking by Senvest-cohort + position-trimming by short-term momentum holders

When the FINRA-ATS data pipeline is wired up, retroactively studying these dates would build the flow-signature playbook for future catalyst events.

7. Limited public data — what we cannot say (yet)

Without FINRA ATS data ingested:

Cannot calculate exact off-exchange % for TSEM
Cannot identify top venues for TSEM share routing
Cannot detect anomalous block-trading days
Cannot triangulate dark-volume + short-interest + Form 4 (when the post-HFIAA Form 4 stream activates) for full institutional-flow read
Cannot distinguish pension-cohort-rebalancing signatures from US-hedge-fund-rotation signatures

This is a partial-data file rather than a complete analysis. As the data pipeline expands, more substantive analysis becomes possible.

8. Comparative context — GFS and POET dark-venue profiles

Issuer	Mkt cap	Estimated off-exchange % ⚠	Distinctive flow feature
TSEM	$21.4B	30-40%	Israeli-pension-cohort agency execution; cross-border timing
GFS	$33B	35-40%	Mubadala secondary-block events; small-float concentration
POET	<$1B	40-50%	Microcap retail-skewed; extended-hours volume non-trivial
TSMC ADR	$1T+	30-35%	Mega-cap deep liquidity; global routing
MRVL	$80B	35-40%	US-domestic large-cap; institutional-typical

TSEM’s flow profile is closest to GFS among foundry comparables but without the Mubadala-secondary block events that periodically punctuate GFS’s flow signature. TSEM’s flow is more “smoothly institutional” — consistent with the dispersed-ownership profile.

9. Forward catalysts driving expected dark-venue activity

Window	Catalyst	Expected flow-signature lens
May 2026	Q1 2026 6-K print	Pre-print accumulation + post-print distribution; volume spike across both lit and dark
2026-Q2 / Q3	First post-HFIAA Form 4 filings	Limited direct flow impact (insider sales typically <0.5% of daily volume) but signals institutional trading-window activity
H2 2026	First named PH18 customer / LWLG productization announcement	Major dark-venue spike on positioning around the binary outcome
2026-2027	OFC / ECOC / SC conference cycle	Routine dark-venue activity around demo-and-press cycle
CY 2027 forward	Potential M&A re-emergence (no current signals; speculative)	Distinctive “merger-arbitrage” dark-venue signature on any M&A-related rumor or filing

10. Open items / backfill queue

FINRA ATS data ingestion — wire up STOCK_DARKPOOL.json data pipeline to populate weekly ATS-by-venue volume. Use FINRA OTC Transparency portal https://otctransparency.finra.org/otctransparency/AtsIssueData.
Block-trade tracking — Bloomberg / aggregator block-trade reports for TSEM during high-volume catalyst days (especially 2026-03-19 and 2026-04-17/04-20).
Retail-internalization estimation — limited public availability; some indication available via FINRA Rule 605 reports.
2025-11-12 CPO press dark-venue signature backfill — once FINRA data is ingested, retroactively study the AI-photonics-premium ignition event flow.
2026-04-17/20 ATH-window block-trade study — was there a single-day block-trade signature consistent with pension-cohort distribution at the ATH? Would be a major analytical insight.
NYSE TRF / NASDAQ TRF data — Trade Reporting Facility data captures non-ATS off-exchange volume; complementary to ATS-reported data for full off-exchange picture.

Sources

FINRA OTC Transparency — https://otctransparency.finra.org/otctransparency/AtsIssueData — primary ATS-volume source.
FINRA Rule 605 reports — wholesaler order-execution-quality reports; some retail-internalization signal.
companies/tsem/data/STOCK_DARKPOOL.json — currently empty {}; pipeline pending.
CBOE / NASDAQ public reporting for general off-exchange % characterization.
GFS dark-pool reference implementation — ../../gfs/kb/06_market_data/darkpool_offexchange.

Cross-references

Stock price history — volume context for high-flow catalyst days
Institutional holders — Israeli-pension-cohort agency execution context
Short interest history — short-side flow context
Insider transactions log — post-HFIAA Form 4 (zero filings as of 2026-04-29)
Options chain — options-overlay flow vs. underlying flow
Bull case — AI-photonics premium narrative driving recent flow signatures