Research Notes

NVIDIA MRC Enables Spectrum-X Ethernet for AI-At-Scale, Now at OCP

Steve McDowell May 10, 2026

NVIDIA recently announced that Multipath Reliable Connection (MRC), an RDMA transport protocol developed for and in production on Spectrum-X Ethernet hardware, is now available to the broader industry via the OCP.

Research Note: Extreme Networks’ Platform Pivot Accelerates

Michael McDowell May 10, 2026

If the company continues to execute on this direction, it will be competing at the AI-networking architectural level where long‑term winners are defined.

Cisco Acquiring Astrix Security to Extend Zero Trust to the Agentic Workforce

Steve McDowell May 6, 2026

Cisco announced its intent to acquire Astrix Security, a Tel Aviv-based company specializing in Non-Human Identity (NHI) security. The acquisition targets the growing, largely underserved segment of enterprise security focused on governance, visibility, and threat detection for non-human identities, including API keys, service accounts, OAuth tokens, and AI agents.

Palo Alto Networks: Portkey Acquisition Anchors its Agentic Security Stack

Steve McDowell May 4, 2026

Palo Alto Networks recently announced its intent to acquire Portkey, a San Francisco-based AI gateway and control plane company. The deal, expected to close in Palo Alto Networks’ fiscal fourth quarter of 2026, will integrate Portkey’s technology into Prisma AIRS, the company’s agentic AI security platform.

HPE ProLiant: Edge Compute for AI and Mission-Critical Workloads

Steve McDowell May 4, 2026

HPE recently expanded its HPE ProLiant edge portfolio with three new platforms: the HPE ProLiant Compute EL2000 chassis, two new Gen12 servers built for the EL2000 (the EL220 and EL240), and an enhanced version of the HPE ProLiant DL145 Gen11 server, now powered by AMD EPYC 8005 series processors. The announcement also introduced an Environmental Ruggedization Option Kit applicable across the portfolio.

Google Cloud: 8th-Generation TPU Family Splits Training and Inference

Steve McDowell April 27, 2026

At Google Cloud Next 2026, Google unveiled its 8th-generation Tensor Processing Unit (TPU) family, splitting its custom silicon lineup into two purpose-built architectures for the first time.

The TPU 8t targets large-scale model training, emphasizing compute throughput and scale-up bandwidth, while the TPU 8i addresses inference and reasoning workloads, emphasizing memory bandwidth and low-latency communication.