SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : ASML Holding NV -- Ignore unavailable to you. Want to Upgrade?


To: BeenRetired who wrote (42597)12/8/2025 7:09:17 AM
From: BeenRetired  Respond to of 42703
 
Yes — advanced packaging will be one of the biggest semiconductor battlegrounds starting in 2026. TSMC’s CoPoS and Intel’s EMIB/Foveros are both racing to meet exploding AI and HPC demand, and both companies are accelerating timelines to make packaging a core differentiator rather than just a supporting technology.

Why Packaging Becomes Critical Post-2026
  • Scaling limits: Shrinking transistors below 3nm is increasingly expensive and complex. Packaging offers performance gains without relying solely on node shrinks.

  • AI demand: GPUs, TPUs, and custom ASICs require massive bandwidth and memory integration — packaging is the bottleneck.

  • Industry timelines:

    • TSMC: Pilot CoPoS lines in 2026, mass production by 2028–2029.

    • Intel: EMIB ramp expected in 2H26, with Foveros stacking already in production for Meteor Lake and future 18A nodes




Strategic Outlook
  • TSMC: CoPoS is a panel-level leap beyond CoWoS, designed to handle the next generation of AI accelerators. Its strength lies in sheer integration scale — think multiple GPUs + HBM in one massive package.

  • Intel: EMIB and Foveros are modular and versatile, already proven in commercial products. Intel is positioning packaging as a way to win foundry customers even if its transistor scaling lags behind TSMC.

  • Consensus: By 2026 onward, packaging will be the primary growth driver in semiconductors, not just a supporting technology. Both companies are betting that whoever controls advanced packaging will control the AI hardware supply chain.


Conclusion: You’re right — packaging will be the big story from 2026 forward. TSMC’s CoPoS and Intel’s EMIB/Foveros represent two different philosophies (panel-level vs. modular stacking), but both are converging on the same reality: packaging is the new Moore’s Law.

Would you like me to also map out which AI companies (Nvidia, Google, etc.) are aligning with TSMC vs. Intel packaging? That’s where the competitive dynamics get really interesting.