Exclusive: Seltz, a startup aiming to transform web search for AI agents, secures $12.5 million in seed funding.

Exclusive: Seltz, a startup aiming to transform web search for AI agents, secures $12.5 million in seed funding.
Summary
AI startup Seltz raised $12.5 million in seed funding to develop a new search engine.
Seltz's system crawls millions of pages daily, returning results under 200 milliseconds.
The company aims to replace traditional search methods with AI-specific information retrieval solutions.

Share

Bookmark

Newsletter

The surge in artificial intelligence has reignited competition in the search engine market, a landscape once dominated by traditional players. As the need for AI-driven chatbots and agents to access timely and relevant information across various sectors—news, products, and services—grows, AI startup Seltz is stepping up to challenge incumbents like Google. Seltz believes that these advanced AI tools necessitate a new type of search engine.

In a recent announcement, Seltz revealed it has secured $12.5 million in seed funding. This financing round was spearheaded by the European venture capital firm Speedinvest and Boston-based B Capital, with contributions from the Italian Founders Fund, United Ventures, and Future Back Ventures, the investment arm of Bain & Company.

Antonio Mallia, the founder and CEO of Seltz, observed that conventional search engines cater to users inputting brief keyword queries and subsequently browsing a list of ranked links. However, AI agents operate differently. They initiate longer, more specific queries—sometimes executing dozens or even hundreds simultaneously—requiring precise, machine-readable information rather than enticing snippets intended for human engagement.

“The previous search methodologies fall short because they were designed with human users in mind,” Mallia explained to Fortune. “What the AI agents require is not found in the snippet; it is embedded in the page’s body, including tables, images, and various content formats that can benefit large language models (LLMs) or agents.”

Mallia has been gearing up for this opportunity for much of his professional life. His doctoral research in computer science at New York University concentrated on information retrieval. He previously worked as an applied scientist with Amazon’s artificial general intelligence team and as a research scientist at Pinecone, a vector-database firm, before launching Seltz. He likened the current technological landscape to the early 2000s when Google’s PageRank transformed search functionalities. “We’re witnessing a revolution once more,” he affirmed, fueled this time by transformer models and AI workflows that autonomously conduct searches.

What distinguishes Seltz, according to Mallia, is its ownership of the entire search framework—encompassing web crawlers, search indexes, retrieval models, and ranking systems—rather than relying on existing search engines. A multitude of AI search tools currently depend on APIs from Google, Bing, or Brave. Companies like OpenAI and Perplexity have also been reportedly striving to create their own search indexes and web crawlers to acquire the latest information. Nevertheless, Google retains a competitive edge, its expansive user base allowing it to cover far more of the web than its competitors.

In December, Google initiated legal action against SerpApi, a service that scraped Google’s search results and allegedly serviced OpenAI, for purportedly bypassing anti-bot security measures. Meanwhile, rivals such as Anthropic and Mistral have been said to utilize Brave’s index for powering web searches in their chatbots. Mallia cites these arrangements as proof that even the major players have not yet developed fully independent web retrieval capabilities, thereby opening the door for a purpose-built solution.

Seltz’s system navigates hundreds of millions of web pages daily, delivering search results in under 200 milliseconds. Rather than always presenting an entire page or a summary, it identifies and extracts specific passages, tables, text, or images as needed by an agent—a process Mallia refers to as context engineering.

Entering a competitive arena, Seltz faces several well-funded rivals. Parallel, initiated by former Twitter CEO Parag Agrawal, recently amassed $100 million at a valuation of $2 billion. Another AI-centric search firm, Exa, secured $85 million, while Tavily was acquired by the AI-cloud entity Nebius for as much as $400 million earlier this year. Despite this, Mallia is optimistic that managing the complete search infrastructure and offering a superior product will set Seltz apart.

Founded in October and based in the U.S., Seltz is a streamlined operation, employing just 15 people, half of whom are full-time. The remote team is divided between the San Francisco Bay Area and European university hubs in Pisa, Italy, and Leipzig, Germany. Many team members hold PhDs in information retrieval, and Mallia has brought on several veterans from Amazon's AI initiatives. The advisory and angel investor roster includes leaders from Google, Ramp, Cohere, Synthesia, and Databricks, alongside academics from information-retrieval research labs at NYU and the University of Glasgow.

Seltz indicated that the funds raised will be directed toward further developing its search stack, expanding its workforce, and initiating an enterprise sales strategy.

Loading comments...