The Voleon Group
| Company type | Private |
|---|---|
| Industry | Investment management |
| Founded | 2007 |
| Founder | Michael Kharitonov, Jon McAuliffe |
| Headquarters | Berkeley, California , United States |
| Products | Quantitative investment management, hedge funds |
Voleon Group (commonly known as Voleon) is a US-based quantitative investment management company headquartered in Berkeley, California. The firm integrates artificial intelligence and machine learning technologies into its trading methodologies.
History
Established in 2007 by Michael Kharitonov and Jon McAuliffe, both PhD graduates and former staff members at D. E. Shaw & Co., Voleon leveraged advancements in computational power and data accessibility to pioneer the use of machine learning in investment strategies.
Although initially met with investor skepticism, the firm commenced live trading amid the 2008 financial crisis. Despite the market’s eventual rebound, Voleon experienced losses for two consecutive years while fine-tuning its sophisticated machine learning frameworks.
By late 2011, the company transitioned from conventional machine learning techniques to proprietary models specifically designed for financial markets. Overcoming substantial computational hurdles, Voleon launched its second-generation platform in July 2012.
Starting in 2011, Voleon began generating positive returns, achieving 34.9% gains in 2012 and 46.3% in 2013. However, performance waned between 2014 and 2015, including a downturn exceeding 9% in 2016 that triggered investor apprehension. The firm's results remained below industry benchmarks in 2017.
As of 2017, the firm’s annualized returns since inception stood at 10.5%, marginally trailing the S&P 500’s 10.7%. Challenges such as market volatility and the transient nature of exploitable patterns tested the effectiveness of its machine learning-driven approach.
In 2018, during a period when the S&P 500 declined by 6.2%, Voleon posted a notable 14% return. The subsequent year saw returns fall to 7%, compared to hedge fund peers averaging 9.2%, while the flagship fund recorded a loss of 9% in 2020. Since inception, its longest-running fund has delivered approximately 9.5% annualized returns.
Voleon’s Institutional equity market neutral fund reported gains of 10.4% in 2022, increased to 13.6% in 2024, and rose by 14% through April of 2025. Meanwhile, the multi-strategy Voleon Composition fund achieved returns of 9% in 2022, improved to 13.5% in 2024, and was up by 12% year-to-date through April 2025. By May 2025, assets under management reached $16 billion.
In October 2023, Business Insider revealed that Voleon imposed strict non-compete clauses requiring departing employees to refrain from working within the industry for two years without compensation or face litigation. Following public backlash, the company declared it would nullify these agreements with former staff by mid-October.
Investment strategy
Voleon concentrates on statistical arbitrage, utilizing extensive data analysis to uncover trading signals and patterns across securities. The firm employs a machine learning methodology where algorithms are developed independently without direct human input to forecast trades. Founders characterize these algorithms as black boxes, as their internal mechanisms remain opaque even to them. For example, Voleon processes terabytes of data encompassing every stock price fluctuation over a span of fifteen years. Unlike competitors who combine machine learning with other approaches, Voleon relies solely on this technology for its trading decisions.