- Oriole Networks has landed its first commercial deployment by teaming with AMD to test its PRISM pure photonic networking system.
- The startup states that PRISM replaces electrical switches with an all-photonic architecture.
- Oriole is entering a growing race to fix AI data center bottlenecks.
- The company claims that its technology will achieve an equivalent latency of 10 nanoseconds, which is a more than order of magnitude level latency improvement.
- Oriole is targeting a substantial opportunity, as AI back-end networking is expected to represent a close to $200 billion market by 2030.
Oriole Networks has introduced its PRISM pure photonic networking system, which promises significant latency improvements by eliminating electrical switches. Co-founder claimed, "What we do is we replace all the electrical switches." The technology is expected to achieve less than 10 nanoseconds in latency, representing an order of magnitude enhancement over current solutions.123456
The move comes at a crucial time, as AI data centers grapple with increasing bottlenecks that hinder performance. Oriole is entering a growing race to address these issues, targeting a market projected to hit $200 billion by 2030. This expansive opportunity positions Oriole at the forefront of a technological shift aimed at optimizing network efficiency in an era driven by artificial intelligence.
With a strong partnership with AMD, Oriole’s strategy emphasizes the urgent need for innovative solutions in network infrastructure. The startup aims to redefine the way data is handled in AI contexts, providing necessary bandwidth and speed to meet future demands. “There’s a lot of stuff in the pipe,” hinted one insider, indicating further developments are on the horizon. Oriole’s PRISM system not only promises to enhance performance but also to fundamentally change the architecture of networking in the AI landscape.
“Oriole Networks is launching its PRISM pure photonic networking system in collaboration with AMD. The startup aims to address significant AI data center bottlenecks as the market is projected to reach nearly $200 billion by 2030.”
