Barcelona's Cala raises €7M pre-seed to build a knowledge layer for AI agents
What's the deal? Cala, a Barcelona-based AI infrastructure startup, has emerged from stealth with a €7 million pre-seed round led by Lightspeed Venture Partners. Kibo Ventures, Kfund, and Masia also participated.
Founded by Elisenda Bou-Balust and Issey Masuda Mora, Cala is building a verified knowledge layer for AI systems. Its product, Matter of Facts, structures public web information into machine-readable datasets covering companies, people, pharmaceuticals, legal cases, financial instruments, and products.
The platform targets developers building AI agents that need factual, current, and structured data. It offers access through a REST API, CLI, MCP server, and SDKs for Python and TypeScript.
Why now? AI agents are rapidly moving from demos to production, and their usefulness depends on the quality of data they can access. As more companies deploy agentic systems, the demand for structured, verified knowledge — rather than raw web scraping — is growing fast.
Lightspeed's involvement signals that top-tier US investors see real opportunity in the European AI infrastructure stack, particularly in the tooling layer that sits between foundation models and end-user applications.
What could go wrong? Cala enters a crowded space. Several startups and incumbents are racing to provide structured data for AI systems, from knowledge graph companies to search API providers. Differentiating on accuracy and freshness will be essential — but hard to sustain as competitors scale.
There's also the question of sourcing. Structuring public web information raises potential legal and ethical issues around data provenance, especially as publishers increasingly push back against AI companies using their content.
The signal: A €7 million pre-seed is large by European standards and reflects growing investor appetite for AI infrastructure plays outside the US. Barcelona continues to punch above its weight as a tech hub, attracting global capital for deep-tech startups.
More broadly, Cala's bet illustrates a shift in AI investment: away from model training and toward the data and tooling layers that make AI systems reliable enough for real-world use. The next wave of AI value may be built not by the model makers, but by the companies that feed them trustworthy information.
