Moonshot Research

Moonshot Research employs systematic, machine-learning-driven strategies in public equity markets. The firm builds proprietary models that process large alternative datasets — including satellite imagery, transaction logs, and supply-chain data — to forecast stock price movements. Its approach is comparable to quantitative hedge funds like Two Sigma and D.E. Shaw, emphasizing automation over fundamental analyst discretion.

The firm acquires raw alternative data from third-party vendors and public sources, then cleans, normalizes, and ingests it into training pipelines built in-house. Signal generation relies on deep learning architectures and ensemble tree models backtested over multi-decade market histories. This end-to-end control of the data pipeline is central to the firm's intellectual property.

Moonshot Research does not centralize investment authority in a single Chief Investment Officer. The firm's multi-office structure supports independent quantitative research teams that each develop, test, and deploy alpha models. A centralized risk framework allocates capital across competing internal strategies based on realized performance and diversification benefit.

As an asset manager, Moonshot Research likely operates as a hedge fund managing external institutional capital, though specific fund structures and investor relationships are not publicly disclosed. Quantitative firms of this profile typically offer commingled and separately managed accounts to pensions, endowments, and sovereign wealth funds.

As a systematic equity manager, the firm diversifies across hundreds to thousands of stocks, with position sizes constrained by volatility-targeting algorithms. Quantitative firms typically apply portfolio-level leverage between 2x and 6x gross notional, though Moonshot Research's precise limits are undisclosed. Risk management is automated and code-enforced, with drawdown triggers that reduce exposure during tail events.

Moonshot Research eliminates the discretionary stock-picker from the investment process. Human researchers design and maintain the machine-learning infrastructure — feature engineering, model architecture, execution algorithms — but do not make individual buy or sell decisions. This separates the firm from hybrid platforms like Point72 or Citadel, which blend quant and fundamental pods.

The firm lists offices in New York, Hingham, San Francisco, Brooklyn, Los Angeles, and Oakland. This distributed footprint mirrors the pod-shop model common in quantitative firms, where small, autonomous teams in separate locations run competing alpha strategies. The structure is designed to limit groupthink and promote uncorrelated signal generation through operational separation.