GlobalFoundries: First Chipmaker to Support Open Standard for AI Scale Up

2026-06-15 · Source: Big Data & AI News - EE Times · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, AI Hardware · Depth: Advanced, medium

Summary

GlobalFoundries is positioned to be the first chipmaker to provide silicon supporting the Optical Compute Interconnect Multi-Source Agreement (OCI MSA), an open standard designed to enable AI data center operators to connect GPUs from diverse suppliers. This initiative, with silicon ramp-up anticipated as early as 2027, addresses potential GPU shortages and vendor lock-in by standardizing scale-up networking, traditionally a proprietary domain. The OCI MSA, announced in March, includes major AI industry players like AMD, Broadcom, Meta, Microsoft, Nvidia, and OpenAI. GlobalFoundries' existing silicon photonics technology and advanced packaging systems are ready for this protocol, with silicon expected this year. Their proprietary Scale technology further exceeds OCI MSA specifications, demonstrating 16 wavelengths per fiber compared to OCI Gen1's four, and offering modulators rated at 100 gigabytes per second with broadband, flat insertion loss fiber-coupling systems. This move facilitates the migration from copper to optical interconnects, crucial for future AI scale-up architectures.

Key takeaway

For AI data center architects evaluating future infrastructure, GlobalFoundries' early adoption of OCI MSA signals a critical shift towards open, multi-vendor optical interconnects. You should prioritize solutions supporting this standard to mitigate GPU supply chain risks and avoid vendor lock-in. This enables greater flexibility in scaling AI compute, moving beyond proprietary copper-based systems. Consider GlobalFoundries' Scale technology for its advanced optical performance, potentially exceeding current OCI MSA specifications.

Key insights

GlobalFoundries' early OCI MSA support and advanced silicon photonics enable open, multi-vendor AI data center interconnects, mitigating GPU supply risks.

Principles

• Open standards foster multi-vendor ecosystems and mitigate supply chain risks.
• Optical interconnects are essential for scaling AI compute beyond copper limitations.
• Advanced packaging is critical for integrating photonic and electronic ICs.

Method

The article describes GlobalFoundries' readiness to support OCI MSA by leveraging existing silicon photonics technology and advanced packaging systems, including high-speed modulators and broadband fiber-coupling systems, to enable optical scale-up interconnects.

In practice

• Implement OCI MSA to avoid GPU vendor lock-in.
• Consider optical interconnects for AI scale-up architectures.
• Evaluate silicon photonics for high-speed modulator integration.

Topics

• Optical Interconnects
• OCI MSA
• Silicon Photonics
• AI Data Centers
• GPU Interconnects
• Advanced Packaging

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Hardware Engineer, AI Architect, Tech Journalist

Related on AIssential

• This Will Power Everything — Optical interconnects are critical for scaling AI data centers beyond copper's physical and thermal limitations.
• MIPS, GlobalFoundries Bet on Physical AI — MIPS, having emerged from bankruptcy and acquired Synopsys's ARC processor IP business, is making a significant comeback by shifting its focus from its foundational MIPS instruction set to RISC-V arch…
• Google and Nvidia Consider Intel as Backup Chip Manufacturer — Google and Nvidia are exploring Intel for chip manufacturing diversification.
• Imec Says AI Scaling Needs More Orchestration Across Research, Design, Manufacturing — Scaling AI demands deep, multi-stakeholder collaboration and standardization across the entire technology stack, from research to manufacturing.
• OpenAI Declares the Next Phase of AI — The AI landscape is shifting towards infrastructure-centric development and widespread distribution, marked by major IPOs and novel compute solutions.
• Why Everyone Is Moving Away from NVIDIA — Hyperscalers are vertically integrating AI infrastructure, developing custom silicon and energy solutions to overcome GPU bottlenecks.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Big Data & AI News - EE Times.