Chalk, a U.S.based startup building real-time infrastructure for AI applications, has closed a $50 million Series A funding round, pushing its valuation to an impressive $500 million.
The round was led by Felicis, with continued support from General Catalyst, Unusual Ventures, and Xfund. Founded by Marc Freed Finnegan, Elliot Marx, and Andy Moreland, Chalk is redefining how companies integrate proprietary data into machine learning pipelines.
Unlike legacy platforms such as Databricks or Snowflake, which focus heavily on batch processing and governance, Chalk is designed for live data and real-time AI decision-making.
“AI that acts in real time requires infrastructure that can keep up. We started Chalk to solve that exact challenge,” said CEO Marc Freed-Finnegan.
Trusted by Forward-Thinking Companies
Chalk’s platform is already powering mission-critical use cases at innovative companies:
- MoneyLion, a fintech firm, uses Chalk for instant fraud detection and automated loan approvals.
- Sunrun, a leading solar energy provider, leverages Chalk to optimize solar panel placements based on real-time geospatial and environmental data.
The latest funding follows a $10 million seed round in late 2023 and will be used to expand engineering and product teams, accelerate platform capabilities, and scale customer adoption.
Why Chalk Stands Out
- Chalk’s infrastructure stack is purpose-built for modern AI workloads and includes:
- A real-time feature store
- Native support for APIs and transactional data sources
- Integrated LLM (large language model) toolchain
- End-to-end monitoring and observability
- Branching and experimentation for data science teams
In contrast to traditional frameworks like Apache Spark, which were optimized for heavy, long running jobs, Chalk is tailored for low latency, high-throughput environments where speed is not a luxury, but a necessity.
As businesses race to operationalize AI in real-world settings, Chalk’s approach to streaming-first infrastructure positions it as a game-changer in the space.
