GlobalFoundries: Silicon Photonics and Geographically Resilient Capacity Drive a New Growth Narrative

Core Conclusion

GlobalFoundries' investor day laid out a clear pivot: silicon photonics, AI power delivery, and custom silicon/IP/software are the growth engines, while smartphone reliance is deliberately reduced. The company’s geographic diversity (U.S., Europe, Singapore) and >500 design wins (~95% sole-sourced) validate customer stickiness. Yet the 2028 financial model shows gross margin and opex targets diverging from Street estimates, and the current stock price ($70.93) trades above the $65 target, implying limited near-term upside. Strategic value provides a floor but the growth premium is already priced.

What the Market May Be Underestimating

First, GFS’s silicon photonics potential in co-packaged optics (CPO) is underappreciated. The SCALE platform—industry’s first MSA-capable CPO platform—already exceeds the OCI MSA standard backed by AMD, Broadcom, NVIDIA, Meta, Microsoft, and OpenAI, with tape-outs underway. All top-4 U.S. hyperscalers and four of the top-5 pluggable transceiver players are engaged. Second, the strategic premium for geographically resilient specialty capacity is growing: more than 50% of global nodes above 10nm are in China/Taiwan. GFS offers fungible, cross-qualified capacity across three regions plus China-for-China optionality. This could become a pricing tailwind as supply-chain diversification accelerates.

Evidence Chain

1. Silicon Photonics: The Clearest Growth Pillar

• Revenue trajectory: >$400M in 2026 → >$1B run-rate by 2028 → ~$2B annual by 2030.
• SCALE platform exceeds OCI MSA requirements; all top-4 hyperscalers engaged.
• SiGe capacity sold out through H1 2027; broader optical networking solution includes TIAs, drivers, low-loss fiber, CWDM filters, germanium photodiodes, and electro-optical testing.
• Investment implication: silicon photonics alone could justify a significant portion of GFS’s current enterprise value if execution holds. But CPO is early-stage; top-two positioning is aspirational.

2. Customer Traction and Geographic Moat

• Design wins: >500 in 2025 vs ~325 in 2024; ~95% sole-sourced.
• Customer breadth: all top-6 mobile fabless/OEM, 7 of top-8 industrial IDMs, all top-5 auto OEMs, 4 of top-5 U.S. aerospace/defense primes, all top-4 hyperscalers.
• Capacity fungibility across U.S., Europe, Singapore provides unique supply-chain insurance.
• Investment implication: sole-sourced revenue creates switching-cost moat. Any shift in sourcing away from China/Taiwan disproportionately benefits GFS.

3. AI Power and Custom Silicon Each Add ~$1B Incremental Revenue by 2030

• AI power: BCD, GaN, advanced analog—data center power delivery is the largest sub-segment. Total incremental ~$1B by 2030.
• Custom silicon, IP, software: combined MIPS+ARC footprint >300 customers, 41 RISC-V and AI cores. Expected ~$1B run-rate by 2030.
• Segment growth: Comms Infrastructure & Data Center SAM growing at 28% CAGR (2025-30); Home & Industrial IoT mid-teens; Auto low-double-digits; Smart Mobile flat.
• Investment implication: IP/software drives higher margins but requires upfront engineering opex—a trade-off that dilutes near-term operating leverage.

4. Financial Model: Gross Margin Gap and Opex Growth

• 2028 Q4 targets: Revenue $2.5B, Gross Margin 40% (MS est. 36%), Opex $369M (MS est. $267M), Operating Margin 25%, EPS $1.00.
• GFS targets opex growing faster than revenue, which adds risk if high-margin revenue streams (IP, custom silicon) under-deliver.
• Capital spending guidance remains ~20% of revenue; depreciation and amortization should boost free cash flow more than today’s model suggests.
• Investment implication: the 40% gross margin target is 6 years after 2022’s similar target; path to it has shifted to IP/software rather than mix alone. Credibility requires demonstrated execution.

Key Divergences and Risks

• Opex growth vs revenue: IP/custom silicon requires heavy engineering hiring. If revenue mix doesn’t shift as projected, operating margins will compress.
• Silicon photonics execution risk: CPO market is nascent; competition from TSMC, Intel, and others; GFS’s top-two claim is unproven.
• Smartphone dependency persists: Mobile is flat revenue through 2030, but SAM grows at only 2% CAGR. Any demand weakness in mature phones could hurt base business.
• Long-term target credibility: 2028 targets mirror 2022’s 40% gross margin timeline; drivers are different. Tangible near-term milestones are more relevant than 2030 aspirational numbers.

Valuation or Trading Implication

At $70.93, GFS trades at ~27x MS 2027 EPS of $2.39, a premium to Asian foundry peers but in line with U.S. semi customers. The $65 price target implies ~8% downside. Equal-weight rating reflects balanced risk/reward: upside from silicon photonics and strategic capacity is offset by execution risk and limited near-term catalysts. Patient investors could wait for more concrete silicon photonics revenue inflection or a pullback driven by industry cyclical fears. The new dividend ($0.12/quarter, up to 50% FCF return) provides a modest yield floor.

Appendix: 2028 Q4 Financial Model Comparison