What's Happening?
Credo Technology Group Holding Ltd. and Oracle have announced that the Open Compute Project (OCP) Community has approved two new open optical telemetry specifications. These specifications define how transceivers and host systems will share link health
data and coordinate out-of-band communication within AI data center networks. The OCP-backed framework is built upon Credo's ZeroFlap host specification, providing a standardized software interface. This development is expected to support the broader adoption of Credo's optical Digital Signal Processors (DSPs) and its reliability-focused connectivity roadmap across a wider range of AI fabric deployments. The workstream, co-led by Credo and Oracle, was established in January 2026 with the goal of creating a common software interface for exchanging detailed connection health information between optical transceivers and host systems. The approved standards will standardize optical link-health monitoring, telemetry sharing, event tracking, and automated fault responses, enabling optical DSPs to exchange information without disrupting network traffic.
Why It's Important?
This OCP approval is significant for the AI data center industry as it addresses a critical challenge in maintaining reliable, high-speed connections within increasingly massive computing clusters. The standardization of optical telemetry is crucial for the next generation of AI infrastructure, where data integrity and network reliability are paramount. For Credo Technology Group, this strengthens its technical credibility in AI fabric reliability and its position within the optical connectivity ecosystem. The adoption of these standards could lead to wider acceptance of Credo's ZeroFlap optics and optical DSPs, potentially translating into increased demand and diversified orders. The partnership with Oracle further solidifies Credo's role in shaping future AI data center connectivity. The move towards standardized optical connectivity is essential for the scalability and efficiency of AI infrastructure, impacting hardware manufacturers, data center operators, and technology providers by offering a more robust and interoperable framework.
What's Next?
The immediate focus for Credo Technology Group will be on the execution of its fiscal 2027 optical roadmap, with expectations for a significant increase in quarterly revenue from optical products. Investors will be closely monitoring how quickly these new designs translate into large, diversified orders and whether rising competition in active cables and 1.6T solutions impacts pricing or margins. Credo is also actively involved in an OCP workstream targeting 2027 approval for micro-emitter optics specifications, further shaping next-generation AI data center connectivity. This initiative, which includes optical technology providers, connectivity specialists, and a leading hyperscaler, aims to set specifications for a new class of high-speed data center interconnects. The company recently launched 224G ZeroFlap optical transceivers supporting 1.6T port speeds and AI architectures, indicating a continued push into advanced optical solutions.
Beyond the Headlines
The approval of these optical telemetry standards represents a broader industry shift towards open standards and interoperability in high-speed data center connectivity. This collaborative approach, exemplified by the OCP Community, is vital for fostering innovation and preventing vendor lock-in in the rapidly evolving AI infrastructure landscape. The emphasis on 'ZeroFlap' technology and out-of-band communication highlights the increasing complexity and performance demands of AI workloads, where even minor disruptions can have significant consequences. This move could also set a precedent for future standardization efforts in other critical areas of AI hardware and networking, promoting a more resilient and efficient global digital infrastructure. The long-term implications include enhanced reliability, reduced operational costs for data centers, and accelerated development of more powerful AI systems due to improved underlying connectivity.
