#networking

LocalLLaMA readers noticed the infrastructure lesson: Zai claimed 15% more GPU inference throughput and 40.6% lower first-token P99 latency with the same GPUs, model, and software stack.

Cisco posted Q3 FY2026 revenue of $15.841 billion (+12% YoY) and GAAP EPS of $0.85 (+37%), beating expectations and pushing shares more than 15% higher toward an all-time high on May 14. CEO Chuck Robbins announced the elimination of approximately 4,000 positions to fund an accelerated pivot toward AI infrastructure.

Cloudflare said on April 10, 2026 that its global network passed 500 Tbps of external capacity across 330+ cities and now protects more than 20% of the web. The company frames the milestone as both DDoS headroom and a response to the changing traffic mix of AI crawlers and autonomous agents.

Microsoft detailed a new MicroLED-based datacenter networking system on March 17, 2026. The project matters because it tackles one of the less visible constraints on AI scaling: the energy, distance and reliability limits of the links connecting servers and GPUs.