Hedge Fund Deploys GPT-5.4 for M&A
Balyasny is now using an AI agent called "Merger Arbitrage Superforecaster," reportedly powered by GPT-5.4, to enhance its investment strategies. The move is a prime example of how quantitative models and advanced AI are being deployed to make M&A-related investment decisions, transforming traditional arbitrage approaches.
Balyasny's foray into AI-driven M&A is spearheaded by its Applied AI team, a dedicated unit of 20 researchers and engineers established in late 2022. This team is tasked with creating proprietary AI-native tools that are integrated directly into the workflows of the firm's approximately 180 investment teams. The initiative is led by Chief AI Officer Charlie Flanagan, a former Google data scientist. The "Merger Arbitrage Superforecaster" is a specific agent within Balyasny's broader AI investment research system. This system is designed to emulate a skilled analyst by reasoning, retrieving information, and taking action. The AI continuously monitors and updates deal probabilities, a significant upgrade from the previous method of using manual spreadsheets and alerts. The firm utilizes GPT-5.4 as a core reasoning engine for its AI platform. This model was selected after a rigorous evaluation process that tested over a dozen dimensions, including forecasting accuracy, numerical reasoning, and a reduction in hallucinations. This centralized approach to AI development, with local customization for different investment teams, has led to adoption by roughly 95% of Balyasny's investment teams. This move into advanced AI for M&A analysis is part of a broader trend where quantitative models are becoming central to merger arbitrage. These strategies rely on systematic analysis of factors like historical deal completion rates, regulatory environments, and market sentiment to assess the probability of a deal's success. The goal is to capture the "spread," or the difference between the target company's stock price and the acquisition price. The application of large language models in M&A extends beyond probability analysis. These models can rapidly synthesize vast amounts of documents, such as regulatory filings and broker research, a process that once took days can now be completed in hours. They can also be used for sentiment analysis around a deal, gauging public and market perception. Balyasny's embrace of AI follows a period of significant internal change after the firm's assets under management dropped by half in 2018. In an email to staff at the time with the subject "Adapt or Die," founder Dmitry Balyasny emphasized the need for change. As of January 2026, the firm's assets under management have rebounded to approximately $31 billion.