Embedded Optical Transmission Tech, Arrival of CPO Set to Deliver AI Supercomputing at Scale

0
Jul 15, 2025
  • The implementation of semiconductor architectures with embedded optical interconnect (I/O) technologies is gaining traction this year.
  • The shift from copper to optical technologies will bring more bandwidth with reduced power needs.
  • Shipments of embedded optical modules are set to grow at a CAGR of 50% through 2033.
  • Of the various types of I/O solutions — OBO, NPO and CPO — the latter will bring the biggest generational shift in AI computing by enabling huge bandwidth scaling, lower power consumption and AI supercomputing scaling.
  • By 2027, broad adoption of NPO and CPO will help push integrated revenue growth to triple digits YoY, while the share of capacity shipped in the total will hit double digits.
  • By 2033, over half of all revenues and capacity shipped will come from integrated semiconductor optical I/O solutions.
  • NVIDIA, Intel, Marvell and Broadcom are currently leading the way towards CPO, but the evolutionary path will be gradual.


Seoul, Beijing, Berlin, Buenos Aires, Fort Collins, Hong Kong, London, New Delhi, Taipei, Tokyo – July 15, 2025

Embedded or integrated semiconductor optical modules are starting to gain traction, with the shipments of On-Board Optics (OBO), Near-Packaged Optics (NPO) and Co-Packaged Optics (CPO) solutions projected to grow at a CAGR of 50% through 2033, according to Counterpoint Research’s Silicon Photonics (SiPh) and Co-Packaged Optics (CPO) Report.

Pluggable optical solutions have been around since 2016, but integrated solutions — namely OBO, NPO and CPO — are now bringing massive improvements in transport capacity and especially processing for AI systems, delivering much higher bandwidth against far less power consumption and enabling the dense, high-bandwidth fabrics AI clustering requires in order to scale.

Semiconductor Optical Interconnect Shipments by Tech and Share of Total Capacity Shipped

Semiconductor Optical Interconnect Shipments by Tech and Share of Total Capacity Shipped, Counterpoint Research
Source: Counterpoint Research Silicon Photonics (SiPh) and Co-Packaged Optics (CPO) Report

“It’s like going from ADSL to FTTH broadband but at the chip level. All this additional speed and efficiency is going to usher in the next stage of AI computing,” said Associate Director Leo Liu, adding, “OBO was the first iteration, with products from companies like Applied Optoelectronics seeing wider adoption in 2023. But CPO is the big game-changer with virtually the entire transport layer being optical.”

CPO will bring the biggest generational shift in AI computing by enabling:

  • Huge bandwidth scaling through ultra-high speeds, low latency and ‘speed-of-light’ throughput between critical components like GPUs and accelerators.

  • Lower power consumption at scale.

  • AI supercomputing scaling by delivering the dense, high-bandwidth fabrics AI clustering requires to scale.

NVIDIA, Intel, Marvell and Broadcom are currently leading the way towards CPO, but the evolutionary path will be gradual.

By 2027, broad adoption of NPO and CPO will help push integrated revenue growth to triple digits YoY, while the share of capacity shipped in the total will hit double digits, according to the Counterpoint report. By 2033, over half of all revenues and capacity shipped will come from integrated semiconductor optical I/O solutions.

Less Copper, More Optics: Evolution of Optical Communication

Less Copper, More Optics: Evolution of Optical Communication, Counterpoint Research
Source: Counterpoint Research Silicon Photonics (SiPh) and Co-Packaged Optics (CPO) Report

“The shift towards ‘less copper, more optics’ means that as we move from OBO to NPO to CPO, far less copper is being used at each stage, resulting in non-linear performance improvements,” said Research Associate David Wu. Ultimately, we can expect to see an 80x generational leap in performance, with the 3D CPO likely to perform 80x better than current solutions.

Gradual replacement of copper with optical brings benefits beyond better chip performance. “What this really means is the delivery of high-speed, high-bandwidth data movement needed for scalable AI clustering, a key step towards democratizing and scaling AI supercomputing capabilities,” noted Liu.

About Counterpoint Research

Counterpoint Research is a global market research firm specializing in products across the technology ecosystem. We advise a diverse range of clients – from smartphone OEMs to chipmakers and channel players to Big Tech – through our offices located in the world's major innovation hubs, manufacturing clusters and commercial centers. Our analyst team, led by seasoned experts, engages with stakeholders across the enterprise – from the C-suite to professionals in strategy, analyst relations (AR), market intelligence (MI), business intelligence (BI), product and marketing – to deliver services spanning market data, industry thought leadership and consulting. Our core areas of coverage include AI, Automotive, Consumer Electronics, Displays, eSIM, IoT, Location Platforms, Macroeconomics, Manufacturing, Networks and Infrastructure, Semiconductors, Smartphones and Wearables. Visit our Insights page to explore our publicly available market data, insights and thought leadership, and to understand our focus, meet our analysts and start a conversation.

Follow Counterpoint Research

press@counterpointresearch.com

Summary

Published

Jul 15, 2025

Author

Team Counterpoint

Counterpoint research is a young and fast growing research firm covering analysis of the tech industry. Coverage areas are connected devices, digital consumer goods, software & applications and other adjacent topics. We provide syndicated research reports as well as tailored. Our seminars and workshops for companies and institutions are popular and available on demand. Consulting and customer

Related Research