What if an independent trader could manage their own investment fund with a team of AI agents?

Backed by Y Combinator, Instinct is building a team of AI agents to help retail traders analyze markets and automate their research.

One analyst to monitor the markets, another to comb through financial results, a third to track news and social media. In an investment fund, these responsibilities are divided among several professionals. Instinct aims to replicate this structure using AI agents coordinated by a single trader. Backed by Y Combinator, the American startup is advancing an ambitious idea: giving an independent investor a portion of the research and execution capabilities traditionally reserved for large financial institutions. A continuously operating team of analysts

Financial markets generate a volume of information that is difficult to track individually. Economic releases, market data, price movements, news, and online discussions constantly feed investment decisions.

Instinct proposes to divide this monitoring among several autonomous agents capable of working simultaneously across different sources. Each agent operates within its own computing environment to research information, develop tools, and delve deeper into opportunities.

The platform presents itself as an agent harness—in other words, an infrastructure tasked with organizing the agents' work, their tools, and the sequence of their actions.

The goal is not simply to summarize financial news. Instinct aims to enable these agents to formulate analyses, cross-reference information from multiple markets, and pursue their investigations without human intervention at every step. From research to strategy execution

Instinct intends to cover a broader workflow than that of a conversational financial assistant. Its agents could notably design analytical tools, explore investment hypotheses, monitor events, and assist in executing trades.

This approach specifically targets traders who develop their own strategies but lack a team of analysts, developers, and quantitative researchers.

The startup believes that smaller portfolios could exploit opportunities that are too specific or too modest to attract the attention of large institutions.

However, public documents do not yet specify the markets actually connected, the compatible brokerage intermediaries, or the level of autonomy granted to agents to place orders. No verifiable financial performance has been published to demonstrate an advantage over traditional strategies. Researchers with backgrounds in quantitative finance

Instinct is part of Y Combinator's Winter 2026 batch. Based in San Francisco, the company has five employees according to its public profile.

It was founded by Joon Won Song, a former quantitative developer at Squarepoint Capital, and Eunju Moon, whose background includes research in quantum computing at the University of Southern California.

The team also claims experience in particle physics, engineering, and financial services.

This background partially explains Instinct's positioning. Rather than designing a consumer assistant intended to comment on stock prices, its founders are trying to replicate the workflows of investment organizations by replacing part of their human coordination with software infrastructure. A $1 billion fund run by a single person?

The most dramatic prediction comes directly from its founders: the first $1 billion investment fund managed by a single person could emerge within the next two years, powered by Instinct's infrastructure.

This is a commercial ambition, not a proven trajectory. Automating research and certain operations eliminates neither financial risks, regulatory obligations, nor the need to oversee decisions.

Managing a large portfolio also involves liquidity, compliance, security, and oversight constraints that go beyond the capabilities of agents alone.

Artificial intelligence can multiply the available analyses, but it does not guarantee their relevance. Multiple agents working in parallel can also replicate the same errors, misinterpret signals, or make decisions based on incomplete information. The trader becomes a team leader

Instinct is currently opening early access requests. The public website does not yet detail pricing, commercial terms, or a general availability schedule.

The project reflects an evolution in financial tools: instead of offering another assistant that answers a user's questions, Instinct seeks to build an entire organization around them. The trader defines their convictions, objectives, and risk tolerance, while the agents handle a portion of the work required for their implementation.

It remains to be seen whether this setup can produce reliable decisions under real market conditions. Instinct's promise is not to make every investor more skilled, but to give them access to a workload capacity previously reserved for specialized teams. A significant distinction that does not do away with the need to know what to do with the generated information.