A Cambridge-based AI company founded by a former Amazon Alexa researcher is building a self-learning economic map that models companies as multidimensional entities rather than forcing them into a bucket. MiddleGame Ventures led; Uncanny Ventures is back.
Amazon is not a retailer. It is also a cloud computing provider, logistics operator, advertising network, media company and hardware manufacturer. The most way financial markets classify it as a single entry in Consumer Discretionary, it captures none of these.
The GICS and ICB classification systems that underpin how portfolios are constructed, indices constructed and risk measured were developed during a different era of corporate structure and have not changed significantly since then.
Theia InsightsOne Cambridge-based deep-tech company spent four years building an alternative and raised $8 million in Series A to accelerate it.
The round was led by MiddleGame Ventures, with participation from Further Ventures and Unusual Ventures, which led Theia’s previous round. Total funding is currently $14.5 million.
The company was founded in 2022 by Dr Ye Tian, a former PhD researcher with an academic background in NLP and artificial intelligence at Amazon Alexa, who assembled a team drawing on backgrounds from Nasdaq, Morgan Stanley, Meta, UC Berkeley’s economics department and Cambridge University’s computer science department.
Theia’s core technology takes regulatory filings, earnings transcripts, press releases and financial data from public companies, applies proprietary NLP and quantitative modeling to extract business activities and map them to what the company calls a dynamic, self-learning ontology of the global economy.
Unlike static classification systems, it models companies as multidimensional entities, tracking “who does what and how much” across all of a company’s revenue lines as those lines change over time.
The resulting map features four products: a Dynamic Industry Classification system (TIIC), a Concept2Universe tool (C2U) that transforms investment themes into evidence-based company universes, a Thematic Factor Model (TFM) to identify structural trend drivers behind stock price movements, and Theme Tracking Indices (TWI) to track the global industry in real-time.
The company’s current clients are institutional: leading index providers, large banks, large asset managers and hedge funds. The strategic rationale for new capital is expansion into private markets where no comparable dynamic classification system currently exists, a gap that is becoming increasingly significant as more institutional capital is allocated to unlisted assets.
The round also funds engineering and research depth and commercial scale.
Patrick Pinschmidt, co-managing partner of MiddleGame Ventures, explained the challenge in terms of both structure and timing: “Financial markets still rely on static classification systems that have changed little over the past few decades. Theia’s approach builds a dynamic, AI-driven map of a company, sector or investment topic, providing game-changing tools to justify investors and AI systems.
The AI angle is intentional: as financial institutions establish themselves AI workflows they need structured economic data that machines can reason with, not just analysts, for research, portfolio construction, and capital allocation. Theia deploys its ontology as an infrastructure for both.
