🏭 Module 7 · Fabrication and Ecosystem · Chapter 7.5 · 8 min read

METU CMP Collaboration (Y10)

Türkiye's mini-fab project — METU + ASELSAN + TÜBİTAK.

Prerequisites

What you'll learn here

  • Identify the Y10 mini-fab stakeholders and roles
  • State the Ankara location and facility plan
  • View the 5-year investment + schedule
  • Explain the public-private-academic collaboration model
  • Compare to peers (TSMC, Intel, ASML)

Hook: Y10 = Domestic Mini-Fab

Y1 starts with the UNAM workshop. Y10 grows it 20×. Can’t be done in a university lab. Mini-fab needed.

Project name: METU CMP (METU Chip Manufacturing Project). Ankara-based, ASELSAN + TÜBİTAK + METU partnership.

This chapter introduces the project.

Intuition: 4 Stakeholders, 1 Facility

StakeholderRoleContribution
METUAcademic + designEngineer training, R&D
ASELSANIndustry + salesDesign transfer, customer reach
TÜBİTAKFunding + oversightState support, strategic approval
SIDRAKnow-how + operationsTechnology, product roadmap

Location: Ankara OSTİM or Teknokent. Area: 10,000 m² (building) + 1000 m² cleanroom. Investment: $70M (building + equipment). TÜBİTAK grant + loan.

Formalism: Project Details

L1 · Başlangıç

Schedule:

  • 2026: project approval, site selection.
  • 2027: building construction begins.
  • 2028: building done, equipment installed.
  • 2029: pilot production (SIDRA Y10 prototype).
  • 2030: serial production, 1M chips/year.

Facility plan:

  • Class 4 cleanroom 1000 m².
  • Class 3 (special steps) 100 m².
  • Office + management 500 m².
  • Test lab 500 m².
  • Materials + storage 1000 m².
  • Power + infrastructure (chiller, UPS, gas) 2000 m².

Equipment list:

  • 8 ALD reactors.
  • 2 DUV 193 nm steppers (new).
  • 1 DUV immersion (more precise, 45 nm capability).
  • 4 ICP-RIE chambers.
  • 4 CMP stations.
  • 4 PVD sputter.
  • 2 CVD.
  • Dicing saw, probe stations, inspection.
  • Total equipment: $50M.

Capacity:

100 wafers/day × 250 days = 25K wafers/year. Y10 die yields ~20 chips (larger die) × 70% yield = 14 chips/wafer. Total: 350K chips/year start.

Scale up (2031-2032): 1M chips/year target.

L2 · Tam

Funding model:

  • $40M TÜBİTAK grant (5 years).
  • $20M ASELSAN private capital.
  • $10M Türkiye Development Bank loan.
  • Total $70M startup investment.

Operations (annual):

  • Personnel 95 × 50K=50K = 4.75M.
  • Consumables + electricity: $3M.
  • Maintenance + amortization: $5M.
  • Total $13M/year opex.

Revenue:

  • Y10 chip price 500×350K=∗∗500 × 350K = **175M/year.**

ROI: (175M−175M − 13M) / $70M = 230% annual. Strong.

Advantages vs external foundry:

  • Supply security: no TSMC slot pressure.
  • IP security: design stays in Türkiye.
  • Domestic employment: 100+ engineers.
  • Strategic: defense + space applications.

Challenges:

  • Equipment imports: ASML, Applied Materials, etc. Vendor approvals required.
  • Know-how: experienced process engineers scarce.
  • Bureaucracy: investment + setup takes time.
  • Suppliers: chemicals, gases (local alternatives?).

Risk management:

  • 2026 project approval delays → lose 1 year.
  • Equipment delivery delays (geopolitics) → lose 6 months.
  • Yield runs low → year-1 loss, then improvement.
L3 · Derin

Comparable projects worldwide:

Intel Ohio fab ($100B): Started 2022. 3 nm capacity. Budget 1000× SIDRA’s.

TSMC Arizona fab ($40B): 5 nm, 2025 production target.

Samsung Taylor Texas ($17B): 2 nm logic.

UAE G42 + OpenAI: datacenter + AI chip fab planned.

SIDRA Y10 mini-fab 70M→∗∗modest∗∗butenoughtostart.Tu¨rkiye′s2030scouldsee70M → **modest** but enough to start. Türkiye's 2030s could see 1B+ fab projects.

METU’s role:

  • Process engineer training program (50% staff METU alumni).
  • R&D partnership (SIDRA roadmap + METU publications).
  • PhD intern positions.

ASELSAN’s role:

  • Customer network (defense, satellite, automotive).
  • Design consulting (radiation-hardened variants).
  • Sales + marketing (international).

TÜBİTAK’s role:

  • Strategic approval + grant.
  • Audit + reporting.
  • Semiconductor policy interface (Ministry of Industry).

Y100 growth:

Y10 mini-fab (70M)→Y100fullfab(70M) → Y100 full fab (5B):

  • Ankara mini-fab → 10× scale-up (Y100 version).
  • EUV lithography (if needed, $200M equipment).
  • 2035 target.

International benchmark:

Taiwan: TSMC founded 1987. 40 years later, world’s leading foundry (~55% market share). Could Türkiye’s Y10 mini-fab follow a similar arc 40 years on?

Probability low but nonzero. SIDRA is the start.

Experiment: Y10 Mini-Fab Annual Economics

Revenue:

  • 350K chips × 500=500 = 175M.

Expenses:

  • Personnel: $5M.
  • Consumables + electricity: $3M.
  • Maintenance: $5M.
  • TSMC wafer purchase still required (substrate) → $10M.
  • Total opex: $23M.

Net annual: $152M margin.

Investment payback: 70M/70M / 152M = 6 months. Excellent.

Realistic adjustment: first 2 years low volume + low yield + sales ramp. $50M/year net. 2-year payback.

Strategic bonus: a leap in the Türkiye semiconductor roadmap. State support continues.

Quick Quiz

1/6What is the METU CMP project?

Lab Exercise

Y10 mini-fab risk-benefit analysis.

Benefits:

  • Domestic production (strategic).
  • $150M/year net.
  • 100+ jobs.
  • Semiconductor ecosystem creation.

Risks:

  • Equipment-import barriers (US sanctions).
  • Yield low at start (50% year 1).
  • Market competition (NVIDIA, Intel).
  • Budget overrun risk.

Mitigations:

  • Parallel equipment suppliers (Europe + Asia).
  • SIDRA digital twin to simulate process (pre-optimize yield).
  • Niche market focus (defense, space, automotive).
  • TÜBİTAK buffer fund.

Decision: Y10 mini-fab is strategically + economically smart for Türkiye. Start in 2026.

Cheat Sheet

  • METU CMP: Y10 mini-fab, Ankara.
  • Stakeholders: METU + ASELSAN + TÜBİTAK + SIDRA.
  • Investment: $70M.
  • Schedule: 2026 approval → 2029 pilot → 2030 serial.
  • Capacity: 350K-1M chips/year.
  • ROI: 2 years (realistic).

Vision: From Mini-Fab to Full Fab

  • Y10 (2030): Ankara mini-fab, 1M/year.
  • Y100 (2035): Same site × 10 = 10M/year. EUV added.
  • Y1000 (2040+): Türkiye multi-fab network. 100M/year global export.

The four-stakeholder model is sustainable. Backbone of Türkiye’s semiconductor industry.

Further Reading