Keenable Secures $26M to Build AI-First Web Search Index With 100 Billion Documents

Former Search Giants Build AI-Native Infrastructure

Keenable has emerged from stealth with $26 million in seed funding to fundamentally revamp web search infrastructure for the age of AI agents1

1

2

2

. Co-founded by Andrey Styskin, who spent 20 years leading search, AI and cloud divisions at Yandex and Amazon, and Matthias Petri, a German AI scientist from Amazon AGI, the startup addresses a critical gap in how autonomous systems access web information. Accel led the funding round with participation from Conviction Partners, Brightwing Capital, scOp Venture Capital, and unnamed executives from Google and Amazon as angel investors2

2

.

Building a 100 Billion Document Index for AI Consumption

Keenable has constructed a web search index spanning over 100 billion documents, specifically optimized for the low-latency, high-frequency queries that AI agents conduct2

2

. Unlike traditional search engines designed for human users who scan snippets, this agentic web search infrastructure enables AI systems to process and ground their responses in complete source documents. The startup's API is already deployed in production at several AI labs and inference providers during both training and runtime phases, though customer names remain undisclosed1

1

. Keenable recently partnered with voice AI company Gradium to support live information retrieval capabilities.

Cost-Efficient Alternative to Enterprise Search Solutions

Styskin explained that Keenable's approach differs fundamentally from enterprise search solutions that become prohibitively expensive at web scale. "If you do not fine-tune your index structures for a specific task, the cost of serving and scanning the whole internet is enormous because of the volume," he told TechCrunch1

1

2

2

. The startup's APIs support straightforward natural language searches and deliver model-ready inputs with minimal delay, priced at $1 per 1,000 API requests for frontier-scale users2

2

. This AI-specific search infrastructure serves use cases including market mapping, pricing monitoring, and lead enrichment.

Timing the Market as Tech Giants Retreat

According to Accel partner Zhenya Loginov, who led the investment, AI players face limited options for web-scale search infrastructure as Google and Microsoft shut down existing search APIs to prevent cannibalization1

1

. These tech giants now favor bundled approaches and selective partnerships. Styskin recognized this opportunity while at Amazon working with Petri on web search infrastructure for applications like Alexa, noting Cloudflare data showing AI crawlers commanding growing search volume shares. While building a giant search index remains "painfully expensive," Styskin believes the innovators' dilemma makes Google "beatable" on agentic queries, allowing smaller companies to offer more cost-efficient solutions1

1

.

Proprietary Retrieval and Historical Search Capabilities

Keenable is developing proprietary retrieval capabilities including WebQueryLanguage, an upcoming product designed to help AI systems answer questions by synthesizing information from multiple web sources when no single source contains the complete answer1

1

2

2

. The platform also supports point-in-time historical queries, enabling models to search the internet as it existed at specific moments rather than only accessing current content2

2

. The startup claims its search quality already surpasses rivals like Tavily, Exa Labs, and Perplexity AI in retrieving relevant, cleaned information optimized for AI agent consumption.

Scaling Operations to Capture Agent Economy

With a current team of 15 engineering staff distributed across the U.S. and Europe, Keenable plans to double its headcount by year-end to accelerate its go-to-market strategy1

1

2

2

. Styskin tapped his network to recruit former colleagues with deep search expertise. The opportunity stems from AI agents currently relying on traditional search APIs or building custom web crawlers from scratch—an inefficient approach when dedicated indexes can handle retrieval while developers focus exclusively on agent development. As the era of "ten blue links" potentially closes for both humans and agents, Keenable positions itself to become what Styskin calls "the next Google for AI agents," though competition from players like Brave, Exa, and Google's own AI search overhaul signals the market's recognition of this infrastructure shift1

1

.