TL;DR AI can’t trade, but portfolio managers can. Kimpton combines high-resolution public financial markets data, portfolio manager strategy/mandate, portfolio holdings, and transaction data to generate trade ideas for managers.

https://www.youtube.com/watch?v=3aED4nL3G50

Our Background

We are former Goldman Sachs and Vistra Energy engineers who started a quantitative hedge fund together in 2021 after raising $10M from a family office at the ages of 22 and 21 years old, scaling it to $35M at its peak.

We spent four years systematically trading using the tools that we built in-house to support our systems. Internally, we developed Kimpton to automate many of our own workflows.

The Problem

AI can’t trade. After all, LLMs are the most overfit backtest in all of human history. They are consensus machines and sycophantic by nature. As a result, AI goes largely unused by portfolio managers today.

Agents finally had depth in 2025. We tested LLMs on the markets and were unimpressed with the output. LLMs can’t trade because they are consensus machines and sycophantic by nature. They lack the edge that portfolio managers provide.

AI can’t trade like a portfolio manager, but they can research at 100X the efficiency of an investment analyst. For this reason, Kimpton revolves around the portfolio manager, solving increasingly all of the workflows that they need.

Analysts spend weeks aggregating data and synthesizing it to generate singular trade ideas that are passed up to portfolio managers for review. It’s time for a change.

Investment management is built on the idea that analysts and PMs need to spend their days staring at screens all day to find edge in the markets.

To refactor the investment management industry, we can’t simply provide better tools. We have to automate the entire research layer instead.

The Solution

Kimpton automates the analysis layer by securely ingesting firm-wide context, portfolio, and transaction data to synthesize fully-formed trade recommendations fitted to their portfolio.

Portfolio managers can challenge the results of the trade proposals through a process we call Adversarial Review, an agentic Socratic questioning to assist portfolio managers in coming to their own conclusions.

We developed a harness specifically for the financial markets to make aggregation efficient, fast, and accurate. It has access to high-quality institutional data sources through Factset, Massive, and Tiingo. Real-time data on tens of thousands of global assets across equities, ETFs, indices, insider trades, SEC filings, events, transcripts, company fundamentals, corporate leadership, and more!

Our AI agents also provide portfolio managers with the ability to generate many different types of artifacts, such as real-time dashboards, institutional-grade reports, and comprehensive visualizations.

Traction

Kimpton is deployed and generating trade proposals on billions of dollars of assets under management today.

Portfolio managers are using Kimpton to monitor their exposures, understand investment thesis drift, earnings coverage, manage positioning across teams, investor reporting, and as an investment team collective brain.

Portfolio managers are underserved today when the potential to refactor the way they work is massive. Kimpton is focused on improving the lives portfolio managers who are actively making decisions in the public markets.

Ask

If you know anyone who manages assets a hedge fund, family office, mutual fund, ETF manager, or other public markets manager — we’d love to speak with them.