ams OSRAM Joined Open Compute Project Workstream
The company will contribute micro-emitter technology to develop standardized optical interconnects for AI data centers.
Updated on Oct. 13, 2026 in Data Centers
Live Poll
Do you believe new optical technologies will make digital infrastructure more sustainable and affordable?
ams OSRAM has joined the Open Compute Project Short-Reach Micro-Emitter Optics workstream to help develop interoperable optical standards. The project focuses on improving connections for AI data center infrastructure.
Why it matters
Data center operators face increasing pressure to connect massive arrays of GPUs efficiently. Standardized optical modules are essential to managing latency, power consumption, and costs at scale.
The new 850 nm Thin-Film VCSEL platform utilizes gallium arsenide technology to enable high-speed data transmission. This approach supports a wide-and-slow optical architecture that distributes data across numerous parallel channels.
The players
ams OSRAM
A global leader in optical solutions that develops high-performance sensors, emitters, and photonics technology.
Open Compute Project
A collaborative community focused on redesigning hardware technology to efficiently support the growing demands on compute infrastructure.
Credo
A high-performance semiconductor company that provides secure, high-speed connectivity solutions for the data infrastructure market.
The details
The workstream, led by Credo, brings together hyperscalers and technology suppliers to create common specifications for signaling, power efficiency, and connector interfaces. ams OSRAM will integrate its 2D addressable microLED and microVCSEL emitter arrays into these standardized designs.
Timeline
In 2026, ams OSRAM introduced its Thin-Film VCSEL platform at ECOC.
A common specification for the project is targeted for OCP approval in 2027.
The Tech Race
This partnership reflects the industry-wide shift toward optical interconnects as copper wiring struggles to keep pace with the massive throughput required for modern AI clusters. It positions optical suppliers to replace legacy copper systems with standardized, high-efficiency modules.
For data center operators and tech developers, this standardization promises to lower costs and simplify the integration of high-bandwidth AI hardware. Users may see improvements in system latency and energy efficiency as these common specifications reach widespread adoption.
The takeaway
As AI clusters grow larger, the reliance on proprietary, non-interoperable hardware becomes a major barrier to scalability. Adopting open standards will be critical for businesses looking to build cost-effective and energy-efficient AI infrastructure in the coming years.
Further reading
Learn more about the latest innovations in Data Centers.
Live Poll
Do you believe new optical technologies will make digital infrastructure more sustainable and affordable?
