Kimpton AI

Kimpton is the AI-native investment research platform for portfolio managers. Agents do the research and propose the trades. It was built by two ex-Goldman engineers and an ex-Vistra Energy engineer who left to start a quantitative systematic hedge fund with $10M raised from a family office at the ages of 21 and 22 years old. They scaled it to $35M at peak over four years, and productized the technology stack they ran it on. Portfolio managers spend most of their day on work that isn't investing: filings, earnings prep, 13Fs, peer benchmarking, thesis validation, trade writeups. Kimpton runs all of it autonomously and delivers structured Trade Proposals with full rationale, written as diffs to the portfolio and fitted to each manager's mandate and strategy. It is built to disagree. Most AI flatters the user, and consensus is poison for investment decisions. Kimpton is grounded in the firm's own mandate, thesis, and positions, so it surfaces breaches and risks the PM would otherwise miss instead of telling them what they want to hear. Risk profile, style, and constraints live in plain-text mandate.md and strategy.md files that drive every analysis the system performs. The rest of the desk sits alongside: Deep Research cited to source, natural-language Dashboards, Agentic Charting and Backtests, scheduled Skills for recurring workflows, and auto-generated Reports. Coverage includes FactSet pricing, fundamentals, estimates, transcripts, ownership, and M&A data, plus live prediction markets from Polymarket. Humans make the decisions. AI does everything else.