Balyasny deploys Gemini across internal research platforms

Balyasny Asset Management has deployed Google Cloud's Gemini models across its internal artificial intelligence research platforms for investment teams worldwide.

The hedge fund manager is using the models in proprietary tools built for analysts who review large volumes of financial information, including earnings transcripts, central bank announcements, regulatory filings and market data.

Balyasny manages more than USD $37 billion in assets, and its investment staff work across more than 200 teams globally. Those teams assess thousands of documents and market updates each week, leading the firm to build internal applications to support research workflows.

One of those tools is BAMAgent, an internal system designed to coordinate specialised AI models and software agents across more than 80 internal financial databases and enterprise tools. Gemini has been added to that broader multi-model setup for work involving large financial datasets, document retrieval, and analysis of visual and tabular material.

The move offers a window into how large investment firms are adopting generative AI in tightly controlled internal settings rather than through public consumer tools. In financial markets, firms have weighed potential gains in speed and coverage against concerns about accuracy, confidentiality, and regulatory compliance.

Research workflows

Gemini is being used for high-volume research tasks tied to Balyasny's internal data sources and market feeds. The models help synthesise research material, earnings transcripts and live market information for analysts seeking faster access to market developments.

The models are also being applied to multimodal analysis inside BAMAgent and related platforms. That includes workflows that require interpreting charts, balance sheets and regulatory filings alongside text-based materials.

Security was a central part of the deployment. The system has been integrated into Google Cloud infrastructure using VPC Service Controls, which, according to the companies, keeps Balyasny's proprietary trading data and strategies isolated within its private cloud environment.

Charlie Flanagan, Chief AI Officer at Balyasny Asset Management, said the firm uses different models for different research tasks.

"Across our investment practice, different research tasks demand specialized model capabilities," said Charlie Flanagan, Chief AI Officer at Balyasny Asset Management.

"When our research agents evaluate thousands of market feeds simultaneously, model speed, retrieval accuracy, and cost efficiency are critical. Gemini delivers the throughput our analysts need for high-volume document search, as well as the multimodal reasoning required to analyze complex financial charts and corporate disclosures," Flanagan said.

Sector pressure

The deployment reflects broader pressure on buy-side firms to process growing volumes of structured and unstructured information without adding proportionate numbers of analysts. Portfolio managers and researchers increasingly face a daily flow of company updates, policy signals, broker research, industry data and alternative datasets that can be difficult to search and compare at scale.

Firms across the sector have been experimenting with AI to summarise documents, identify anomalies, retrieve past disclosures and surface links across datasets. Yet financial institutions have generally taken a cautious approach, preferring systems that can be linked to internal repositories and governed under existing security controls.

Google Cloud framed the project as an example of AI use in capital markets, where speed, precision and data governance are closely watched.

"Capital markets require speed, precision, and strict data governance, making them a natural proving ground for advanced artificial intelligence," said Rohit Bhat, General Manager and Managing Director of Financial Services at Google Cloud.

"By integrating Gemini into its custom research platforms, BAM is showing how global investment firms can securely deploy specialized models to analyze complex market signals and give their analysts an edge," Bhat said.

Balyasny was founded in 2001 and employs more than 2,700 people across 24 offices in North America, Europe, the Middle East and Asia. Its investment activities span equities (long/short), fixed income and macro, commodities, multi-asset arbitrage, and systematic strategies.

The firm's Applied AI team is also testing upcoming Gemini models through early-access programmes as it seeks to improve internal research systems.