B. Riley has reported that Amazon’s RNG and OpenAI’s MRC technologies are becoming significant challenges for the transceiver total addressable market (TAM). The firm indicates that previous assumptions about increasing transceiver demand due to deeper networking tiers are now facing obstacles as hyperscalers shift toward flatter network architectures to mitigate complexity. This transition is expected to lead to a reduction in active networking equipment and a potential decrease in structural transceiver counts by 40% to 50%. Companies like $AAOI could be particularly affected, as Amazon and Oracle are anticipated to be key customers driving growth in the 800G/1.6T transceiver market.

Amazon: Amazon operates extensive cloud and AI infrastructure through its AWS division, including custom networking technologies. The company is advancing RNG as a solution to simplify and flatten its scale-out networks for inference, data ingestion, and retrieval-augmented generation workloads.
OpenAI: OpenAI develops advanced artificial intelligence models and supporting infrastructure for large-scale training. Its MRC technology is designed to help training clusters overcome physical network limitations in multi-tier architectures.
B. Riley: B. Riley Financial is a financial services firm known for equity research and analysis across technology and growth sectors. In this news, the firm issued a report highlighting structural shifts in AI networking that challenge prior assumptions about transceiver demand.

AI Networking Trends: Hyperscalers are increasingly adopting flatter network designs to reduce complexity in AI clusters.
Transceiver Market Impact: Changes in network architecture are altering demand patterns for high-speed optical components in data centers.