Enfabrica is a Networking & Custom Silicon company founded in 2019 and based in Sunnyvale, USA.
No dated funding rounds on file. This company is privately held and may be raising or operating on undisclosed capital.
No articles ingested yet for Enfabrica. Once the hourly news pipeline is live, every article the classifier tags as mentioning this company appears here with its one-line AI summary and sentiment.
A 3.2 Terabit/second GPU networking chip delivering multi-port 800-Gigabit-Ethernet connectivity to GPU servers with four times the bandwidth and multipath resiliency of competing GPU-attached network interface controllers. Designed for rack-scale GPU compute clusters, it features Resilient Message Multipathing and Collective Memory Zoning to optimize AI training, inference, and retrieval-augmented generation workloads in data centers. Available for sampling and pilot deployments starting Q1 2025.
Enfabrica's elastic AI memory fabric system powered by the ACF-S chip, enabling up to 18 terabytes of CXL DDR5 memory pooling per node over 3.2 Tbps RDMA Ethernet connections. Solves memory bandwidth and stacking challenges for large language model inference by elastically connecting up to 144 CXL memory lanes to 400/800 Gigabit Ethernet ports, reducing GPU and HBM consumption while improving token generation economics and query throughput.
15 patents on file, but none with both an extractable figure and an abstract on Google Patents yet.
We don't have a live feed for this company's ATS. Their careers page has every open role.
View all careers ↗