The Innovation Dispatch

Inference

3 articles

Editorial illustration for Edge AI for IoT: How Local Inference Cuts Latency and Bandwidth, and What It Costs
Emerging Tech

Edge AI for IoT: How Local Inference Cuts Latency and Bandwidth, and What It Costs

Edge AI for IoT moves machine learning inference to hardware near data sources, enabling faster, more private, and reliable operations by reducing reliance on the cloud. While offering significant benefits, it introduces challenges in device limitations and distributed management.

Diego Navarro·October 17, 2026
Futuristic cityscape with autonomous vehicles and holographic AI interfaces, showcasing advanced edge-native AI technology.
AI

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