Copilot: Plausible 7nm & below could exceed 80% of TSMC’s revenue 2Q26.
  Yes, it’s very plausible that 7nm and below could exceed 80% of TSMC’s revenue by Q2 2026—driven by accelerating 2nm adoption, continued strength in 3nm and 5nm, and tapering demand for older nodes.
  Why 80%+ Is Likely by Mid-2026
 - Current Share: In Q3 2025, nodes 7nm and below already accounted for 74% of TSMC’s revenue, with 3nm alone contributing 24%.
   - 2nm Ramp:
   - TSMC plans to produce 60,000 wafers/month across four fabs by 2026.
   - Wafer pricing is ~50% higher than 3nm, boosting revenue share even with modest volume.
   - Early yields (~60%) and strong tapeout momentum suggest faster adoption than 3nm or 5nm.
      - Customer Demand:
      - TSMC has secured 15 customers for 2nm, with 10 focused on HPC—a shift from mobile-centric adoption at earlier nodes.
   - Apple, AMD, Qualcomm, and MediaTek are expected to launch 2nm designs in 2026.
   
 
     
 Older nodes (16nm and above) will continue to decline as mobile, AI, and HPC shift to advanced packaging and chiplets.  - High wafer pricing at 2nm amplifies revenue impact even before full volume ramps.
      Strategic Implications
 - TSMC’s SPR and GAAFET integration at 2nm gives it a competitive edge over Samsung SF2 and Intel 18A.
   - Packaging and chiplet ecosystems (CoWoS, InFO, SoIC) will further concentrate revenue in advanced nodes.
   - AI and mobile convergence—with edge inference and connectivity—will push even mid-tier designs toward 5nm and below.
         Bottom line: With 2nm ramping faster than expected and 3nm/5nm still growing, TSMC’s sub-7nm revenue share could easily cross 80% by Q2 2026, marking a major inflection point in node consolidation.
    Would you like a breakdown of how wafer pricing and yield at 2nm compare to 3nm and 5nm—and how that affects margin mix? I can chart that out next.
    PS  I'm here for the EXE/NXE/NXTi Age. What proceeded? Warmups.
  ASML Village |