Agentic AI Startup Stacks Raises $23M
Stacks, a startup that develops agentic AI for enterprise finance, has raised $23 million in a Series A funding round. The company plans to use the capital to scale its workflow automation platform. Stacks focuses on automating corporate finance tasks such as month-end close, financial reporting, and compliance.
- The Series A round was led by Lightspeed, with existing investors EQT Ventures, General Catalyst, and S16VC also participating. This follows a $12 million seed round less than a year prior, led by General Catalyst. Stacks has onboarded over 30 enterprise clients, saving them a combined 100,000 hours annually. - Agentic AI in quantitative finance allows autonomous systems powered by LLMs to handle multi-step workflows like ETL, model training, and reporting, accelerating research and validation. These systems can autonomously research hypotheses, fetch data, run backtests, and evaluate performance without continuous human intervention. For trading, LLMs are being used to generate alpha factors from unstructured data like news sentiment and to make direct trading decisions. - To achieve low-latency in trading systems, which is critical for high-frequency trading, key architectural considerations include co-locating servers with exchanges, using kernel bypass technologies, and employing event-driven architectures. Programming languages like C++ and Rust are often chosen for their direct hardware access and control over memory, which helps in minimizing delays. - For quantitative analysis and backtesting in Python, popular open-source libraries include Backtrader for detailed simulations, Zipline for its integration with pandas, and PyAlgoTrade for its event-driven architecture. QuantLib is a comprehensive library for derivatives pricing and risk management. - Real-time payment infrastructure relies on APIs (often RESTful or ISO 20022 compliant) to connect to payment rails, enabling instant settlement, which reduces counterparty risk and the need for credit. This "always-on" infrastructure is becoming essential for everything from B2B transactions to peer-to-peer payments. - Alternative data sources are increasingly used in quantitative strategies to gain an edge. These include unstructured data from social media, product reviews, satellite imagery, and web traffic to gauge consumer sentiment and economic activity. - Quantum computing is poised to revolutionize financial modeling by significantly speeding up complex calculations like Monte Carlo simulations for risk assessment and portfolio optimization. Quantum machine learning also shows promise in enhancing fraud detection by identifying anomalies in massive datasets. - Key fintech regulations to be aware of in 2025 and 2026 include the EU's Digital Operational Resilience Act (DORA) and the Markets in Crypto-Assets Regulation (MiCA). In the UK, the Financial Services and Markets Act 2000 (Cryptoassets) Regulations 2026 will establish a new regulatory regime. There is also a growing focus on AI accountability and explainability in financial services.
