Inference

2 articles

Futuristic cityscape with autonomous vehicles and holographic AI interfaces, showcasing advanced edge-native AI technology.

What Are Edge-Native MoE Serving Engines for AI?

A new generation of edge-native serving engines, exemplified by MoEless, slashes AI inference latency by 43% and operational costs by 84% compared to existing solutions.

Arjun Mehta·August 23, 2026

Abstract AI core with holographic neural networks and data streams, symbolizing Infinity AI's challenge to Nvidia's AI dominance in a futuristic setting.

Startups

Infinity AI raises $15M to challenge Nvidia's AI dominance

In a single day, Infinity AI's research agent, Ignition, boosted the inference throughput for the Qwen3-8B model from approximately 1,400 tokens per second to over 20,000 tokens per second.

Diego Navarro·July 20, 2026