Skip to content
primarysourced Photonics sector Tower
TSEM
~10 min read · 2,222 words ·updated 2026-04-29 · confidence 19%

Confidence legend: ✓ verified-primary (Tower 20-F, Tower earnings releases, WSTS / SIA market data, Yole / IC Insights / IBS Group reports) · ◐ partial / aggregator · ⚠ inferred / estimate.

This page provides bottom-up addressable-market sizing across Tower Semiconductor’s five end-market segments. The framing distinguishes:

  • TAM (Total Addressable Market) — the total revenue pool served by ALL foundries (and IDMs) for the relevant end market.
  • SAM (Serviceable Addressable Market) — the subset of TAM that’s accessible to Tower given (a) Tower’s process portfolio, (b) Tower’s qualified-and-released technology, and (c) Tower’s geographic + capacity footprint.
  • SOM (Serviceable Obtainable Market) — Tower’s actual current revenue + reasonable forward share-capture, reflecting capacity, design-wins, and competitive dynamics.

The single most important framing fact: Tower’s FY 2025 revenue of $1.57B sits against a specialty-foundry SAM of approximately $25-35B globally, implying Tower captures 4-6% of the specialty-foundry pool today. The FY 2028 $2.84B target implies SOM rising to 8-10% of a ~30-40B specialty-foundry pool as the SiPh + SiGe + BCD + automotive segments compound at differentiated growth rates.

1. Aggregate Tower Revenue vs Specialty-Foundry Pool

The specialty-foundry pool

The total specialty-foundry-tier revenue pool aggregates:

  • TSMC specialty fabs — Fab 2 / 3 / 5 / 8 specialty + OIN node + RF / BCD / CIS volumes. Approximately $8-12B of TSMC’s ~$90B 2025 revenue. ⚠ analyst-class estimate.
  • GlobalFoundries — entire $6.79B 2025 revenue base (no leading-edge competition). ✓ GFS Q4’25 results.
  • UMC specialty — approximately $4-5B of UMC’s ~$7B 2025 revenue (some 28nm leading-edge competition). ⚠
  • SMIC specialty — approximately $3-4B of SMIC’s ~$8B 2025 revenue (gated by US BIS controls). ⚠
  • Tower Semiconductor — $1.57B FY 2025. ✓ Tower Q4 2025 release.
  • Vanguard International Semiconductor (VIS) — ~$1.3B 2025 revenue. ⚠
  • DBHiTek (Korea) — ~$0.8B 2025 revenue. ⚠
  • SilTerra (Malaysia) — ~$0.3B 2025 revenue. ⚠
  • IDM-foundry merchant volumes (Samsung Foundry specialty, Intel-Tower-deal-failure specialty residuals, others) — ~$2-4B aggregate. ⚠

Aggregate specialty-foundry pool: ~$28-40B in 2025. The mid-point estimate is ~$33B, of which Tower captures ~5% at $1.57B.

Tower’s pool capture trajectory

YearTower revenue ($B)Specialty-foundry pool ($B)Tower share
FY 2022~1.68~$30~5.6%
FY 2023~1.42~$28~5.1%
FY 20241.44~$30~4.8%
FY 20251.57~$33~4.8%
FY 2026E~1.85~$36~5.1%
FY 2028E (Tower target)2.84~$42~6.8%

⚠ The 2026+ specialty-foundry pool figures are analyst-class estimates assuming the broader specialty-tier grows at 5-7% CAGR (slower than aggregate semiconductor industry due to commoditization at the low end and TSMC specialty-share expansion at the high end).

Tower’s path to the FY 2028 $2.84B target requires roughly 2.0-percentage-point share expansion from 4.8% (FY 2025) to 6.8% (FY 2028). The share expansion is disproportionately concentrated in the silicon-photonics sub-segment, where Tower’s positioning is structurally favorable.

2. Per-Segment TAM / SAM / SOM Decomposition

Segment 1 — RF Infrastructure (incl. SiPh + SiGe + RF Infra-CMOS)

Tower FY 2025 revenue: $423M (27% of $1.57B; +73% YoY)

TAM sizing:

  • SiPh foundry-wafer pool: ~$1.1B in 2026 → ~$2-3B by 2028 → ~$3-4B by 2030 (per silicon photonics market analysis).
  • SiGe BiCMOS driver-IC pool (for transceivers): ~$0.5-0.8B in 2026 → ~$1.0-1.5B by 2028.
  • 5G + 6G base-station + LEO satellite RF infra-CMOS pool: ~$1.5-2.0B in 2026 → ~$2.0-2.5B by 2028.
  • Total RF Infrastructure TAM: ~$3.0-3.9B in 2026 → ~$5.0-7.0B by 2028.

SAM (Tower-accessible):

  • ~80% of SiPh pool (Tower has production-ready PH18 + 300mm SiPho + CPO Foundry).
  • ~70% of SiGe driver-IC pool (Tower’s #1 position per Q3 2025 release).
  • ~30% of 5G/6G/satellite RF infra-CMOS pool (Tower competes with TSMC, GFS specialty).
  • Total RF Infrastructure SAM: ~$1.8-2.4B in 2026 → ~$3.0-4.5B by 2028.

SOM (Tower current + projected):

  • FY 2025: $423M = ~22-23% of current SAM.
  • FY 2028 implied (per $2.84B model): ~$1.0-1.2B = ~25-30% of projected 2028 SAM.

Growth trajectory: 25-35% CAGR through 2028 with continuation of capacity-ramp + capacity-prepayment dynamics. The single most-load-bearing segment in the FY 2028 target model.

Segment 2 — RF Mobile (handset RF front-end)

Tower FY 2025 revenue: $360M (23% of $1.57B; -14% YoY)

TAM sizing:

  • Global handset RFFE foundry-wafer pool: ~$5-6B in 2025 (handset PA / FEM / switch / LNA wafer revenue across all foundries).
  • Driven by: 1.2B-unit annual handset volume × increasing RF content per phone (5G mid/high-band + integrated SiGe modules).

SAM (Tower-accessible):

  • ~40% of total handset RFFE foundry-wafer pool (Tower competes with TSMC, GFS, UMC specialty for premium SiGe BiCMOS + RF SOI volumes).
  • Total RF Mobile SAM: ~$2.0-2.4B.

SOM (Tower current + projected):

  • FY 2025: $360M = ~15-18% of current SAM.
  • FY 2028 implied: ~$500-600M = ~20-25% of projected SAM (with envelope-tracker design-win contribution).

Growth trajectory: Range-bound near-term; recovery to $500-600M by 2028 driven by envelope-tracker ramp + integrated-SiGe-module penetration. Cyclical stabilizer rather than growth contributor.

Segment 3 — Sensors and Displays (CIS, MEMS, OLED-on-silicon)

Tower FY 2025 revenue: $251M (16% of $1.57B; +9% YoY)

TAM sizing:

  • Specialty CIS foundry pool (excluding mobile-CIS / OmniVision-at-TSMC volume): ~$2-3B in 2025 (industrial machine-vision + automotive ADAS + medical + scientific imaging).
  • MEMS-on-CMOS foundry pool: ~$1.0-1.5B in 2025.
  • OLED-on-silicon AR display backplane pool: ~$0.3-0.5B in 2025 (early-stage); projected to ~$1-2B by 2028 if AR cycle accelerates.
  • Total Sensors and Displays TAM: ~$3.3-5.0B in 2026 → ~$5.0-7.5B by 2028.

SAM (Tower-accessible):

  • ~25-30% of specialty CIS pool (Tower competes with Sony, Samsung LSI internal, ON Semi internal).
  • ~30% of MEMS-on-CMOS pool.
  • ~15-25% of OLED-on-silicon pool (early-stage; rapid evolution of competitive positioning).
  • Total Sensors and Displays SAM: ~$1.0-1.5B.

SOM (Tower current + projected):

  • FY 2025: $251M = ~17-25% of current SAM.
  • FY 2028 implied: ~$450M (+OLED ramp) = ~30% of projected SAM.

Growth trajectory: 10-20% CAGR depending on AR cycle. Asymmetric upside if OLED-on-silicon for AR HMDs ramps as projected.

Segment 4 — Power Management (BCD)

Tower FY 2025 revenue: $251M (16% of $1.57B; +20% YoY)

TAM sizing:

  • BCD foundry-wafer pool (excluding IDM-internal volumes): ~$3-4B in 2025.
  • Split by end-application: 40% smartphone PMIC + envelope tracker ($1.4B), 35% automotive power conversion + battery management ($1.2B), 20% industrial / consumer ($0.7B), ~5% other.
  • Total BCD TAM: ~$3-4B in 2025 → ~$4.5-6B by 2028 (driven by EV cycle + premium-handset PMIC content uplift).

SAM (Tower-accessible):

  • ~25-30% of BCD foundry pool (Tower’s 0.18 / 0.13 µm BCD + Agrate 300mm BCD competes with TSMC specialty BCD, GFS BCD, UMC BCD, IDM-internal).
  • Total Power Management SAM: ~$1.0-1.3B.

SOM (Tower current + projected):

  • FY 2025: $251M = ~20-25% of current SAM.
  • FY 2028 implied: ~$500-600M = ~30-35% of projected SAM (with Agrate ramp + envelope-tracker design-win contribution).

Growth trajectory: 15-25% CAGR through 2028 driven by Agrate 300mm BCD utilization + envelope-tracker ramp + EV cycle compounding.

Segment 5 — Discretes / Other

Tower FY 2025 revenue: ~$282M (~18% of $1.57B; -11% YoY)

TAM sizing:

  • Catch-all for legacy 150mm flows + miscellaneous foundry runs.
  • TAM definition is ambiguous — partially overlaps with the four primary segments, partially distinct.

SAM (Tower-accessible): ~$1-2B catch-all for legacy + miscellaneous. Tower’s strategic posture is to run this segment down as legacy 150mm Fab 1 capacity is retired.

SOM (Tower current + projected):

  • FY 2025: $282M.
  • FY 2028 implied: $200-300M (continued runoff).

Growth trajectory: Declining at 5-10% CAGR by management design (capacity-shift to higher-margin 200mm / 300mm work).

3. Aggregate FY 2028 SOM Build

Combining the per-segment SOM trajectories to triangulate the FY 2028 $2.84B target:

SegmentFY 2025 ($M)FY 2028E ($M)YoY CAGR% of FY 2028 Mix
RF Infrastructure4231,000-1,20033-42%35-42%
RF Mobile360500-60012-19%18-21%
Sensors and Displays251400-50017-25%14-18%
Power Management251500-60026-34%18-21%
Discretes / Other282200-300-5 to -11%7-11%
Total1,5672,600-3,20018-27%100%

⚠ The mid-point of $2,900M aligns closely with Tower’s $2.84B FY 2028 target. The model implies:

  • RF Infrastructure compounds at ~37% CAGR (the load-bearing assumption — see silicon photonics market sensitivity analysis).
  • RF Mobile recovers at low-double-digit CAGR (envelope-tracker design-win-anchored).
  • Sensors and Displays compounds at ~21% CAGR (OLED-on-silicon ramp dependent).
  • Power Management compounds at ~30% CAGR (Agrate utilization + EV cycle dependent).
  • Discretes runs off at -8% CAGR.

The single most-load-bearing assumption is the RF Infrastructure compound rate. Sensitivity:

  • If RF Infrastructure compounds at 25% CAGR (rather than 37%), FY 2028 RF Infrastructure ≈ $830M, total revenue ≈ $2.5B (-12% vs target).
  • If RF Infrastructure compounds at 45% CAGR, FY 2028 RF Infrastructure ≈ $1,300M, total revenue ≈ $3.0B (+5% vs target).

The capacity-prepayment-with-cash structure (>70% of SiPh capacity reserved through 2028) provides material cushion against the lower-end scenarios for at least the SiPh portion of RF Infrastructure.

4. The Larger TAM (Beyond Specialty-Foundry Pool)

If the analytical lens widens beyond the specialty-foundry pool to include Tower’s potential adjacencies:

  • Total semiconductor industry TAM (WSTS forecast ◐): ~$680B in 2025 → ~$880B by 2028 → ~$1.0T+ by 2030.
  • Foundry industry TAM (pure-play + IDM-foundry): ~$130-150B in 2025 → ~$190-220B by 2028.
  • Tower’s potential maximum SAM (if Tower could compete across the full specialty + adjacent merchant tier): ~$50-70B.
  • Tower’s actual SOM: $1.57B = 2-3% of potential maximum SAM.

The widening between “potential maximum SAM” and “actual SOM” reflects:

  • Process portfolio constraints — Tower does not compete in leading-edge digital, leading-edge advanced HBM, advanced GPU/CPU.
  • Capacity constraints — Tower’s wafer-starts capacity is limited; would need multi-billion-dollar incremental capex to expand beyond current footprint.
  • Customer-relationship depth — Tower’s incumbency at Skyworks RFFE, Coherent SiPh, etc. is multi-decade; new-customer-acquisition is slow.
  • Geographic constraints — Tower’s Israel + US + Italy + Japan footprint is strategic but doesn’t include China-domestic capacity that some Chinese fabless customers prefer.

The structural framing: Tower’s TAM is large in absolute terms but its addressable share is constrained by process portfolio. The thesis-relevant question is whether Tower can grow SOM faster than the specialty-foundry pool grows — which it has done through 2025 (FY 2024 share ~4.8% → FY 2025 share ~4.8% reflects the AI-photonics-driven outperformance offsetting RF Mobile compression).

5. Key Cross-References to Independent TAM / SAM Sources

6. Open Items / Backfill Queue

  1. Specialty-foundry pool sizing for 2025 from a single primary source — currently triangulated; would benefit from a Yole or IBS Group annual report citation. ⚠
  2. Tower-disclosed FY 2028 segment-mix % for the $2.84B target — Tower has not publicly disclosed exact category percentages for 2028; the per-segment trajectories above are analyst-class triangulation. ⚠
  3. SiGe driver-IC standalone TAM — the SiGe-only revenue is implied at ~$193M FY 2025 (from $421M SiPh+SiGe combined - $228M SiPh-only); but Yole / Dell’Oro do not separately report SiGe-driver-IC TAM. ⚠
  4. Automotive-specific Tower revenue — the automotive contribution to BCD + CIS + SiGe radar is approximately 20-25% of total Tower revenue per analyst-class estimate; Tower does not disclose this directly. ⚠
  5. OLED-on-silicon TAM — the AR / micro-OLED foundry pool sizing is wide-range ($0.3-0.5B today to $1-2B+ by 2028); Yole and IDTechEx publish forecasts but with material methodological differences. ⚠
  6. OFC 2026 / OFC 2027 Coherent + Tower demo cadence — translates to incremental design-win-class revenue but specific revenue contribution is not publicly disclosed. ⚠

Cross-section pointers