Cast Insights raises $4.5M to turn the world's TV and radio chatter into market signals
What's the deal? Cast Insights, a San Francisco startup building a real-time intelligence layer for spoken audio, launched its platform after closing a $4.5 million pre-seed round. Abstract Ventures led, with HF0Dealroom has a profile for this one. Try Dealroom →, Village Global, Max VenturesDealroom has a profile for this one. Try Dealroom →, Embassy Ventures, and Stratus VenturesDealroom has a profile for this one. Try Dealroom → participating.
What's the endgame? The startup uses AI to index what it calls "ephemeral speech data" — the millions of statements broadcast daily across TV, radio, and podcasts that vanish the moment they are spoken. It transcribes spoken audio, identifies speakers, and maps how narratives and sentiment shift in real time.
Why now? Co-founder and chief executive officer Otávio Costa Miranda argues that traditional forecasting relies on lagging indicators because tools to capture public speech have not existed. "Most of what gets said in the world every day across radio stations, legislatures, podcasts and local broadcasts is ephemeral by nature," he said. "It surfaces publicly, it influences people and then it vanishes."
What's the product? Emerging from the HF0 Residency, a program for repeat founders, Cast monitors thousands of sources to help institutional decision-makers spot market, policy, and geopolitical shifts before consensus forms. In one case study on the Strait of Hormuz, it tracked mentions rising from near-zero to a peak of 2,946 a day as a US-Iran ceasefire collapsed, with positive sentiment falling from 7.7% to 3.3%.
By the numbers: Cast has processed more than 2.3 million hours of speech across 20 countries and 10 languages while working with early adopters. The new funding will help it scale to 20 million hours and monitor 500,000 hours of live content daily.
The signal: As markets increasingly hunt for alternative data, Cast is betting that the spoken word — long too fleeting to track — becomes a tradeable input. "The signal is always there," Miranda said. "It just slips through everyone's fingers."
Read more: SiliconANGLE
