Octen Secures $10M Seed to Launch World’s Fastest AI Search Infrastructure

Search infrastructure company Octen has emerged from stealth with $10 million in seed funding led by Square Peg. The company launched what it calls the fastest web search API in existence, specifically engineered for the high-concurrency needs of autonomous AI agents. This new stack achieves an average response time of 60 milliseconds while handling over one million queries per second.

Traditional search engines built for human eyes focus on blue links and advertisement rankings. In contrast, Octen provides a structured data layer that allows AI models to reason over the live web with the speed of internal memory. The system manages a trillion-scale index with data updates occurring every few minutes to ensure accuracy.

Kuan Zou, the former product lead for Alibaba Cloud AI search, founded the company to solve the latency bottlenecks inherent in legacy systems. He has assembled a specialized team from Meta, Google, and TikTok to build this retrieval layer. The new capital will be used to scale distributed server architecture and grow engineering teams in San Francisco and Singapore.

Octen faces competition from AI-native search providers such as Tavily and Exa.ai. While these platforms offer citation-backed results, their latencies often range from 180ms to over 300ms. Octen aims to differentiate itself by offering significantly higher throughput and granular concurrency handling for complex tasks. Along with the search API, the company released the Octen-Embedding-8B model which recently topped the Retrieval Embedding Benchmark. This open-source tool is optimized for industry verticals like legal and healthcare. Currently, the search API is available in an invitation-only beta for developers building agentic workflows.