🔥The Multi-Agent Future of VC: Will We See One-Person Billion-Dollar Funds?
How LLM Agents Are Reshaping the Venture Landscape — One Task at a Time
👋 Hi, I’m Andre and welcome to my newsletter Data Driven VC which is all about becoming a better investor with Data & AI. Join 34,320 thought leaders from VCs like a16z, Accel, Index, Sequoia, and more to understand how startup investing becomes more data-driven, why it matters, and what it means for you.
ICYMI, check out our most read episodes:
Brought to you by Affinity - Exclusive Dealmaking Benchmarks
From market trends to performance benchmarks, Affinity’s 2025 European venture capital benchmark report reveals key insights into the dealmaking activities of European venture capital firms.
With exclusive data on 640+ VC firms, view the report to see how your firm stacks up against the competition and what you can do to help your firm rise up the rankings.
From Single Prompts to Agent Intelligence
In 2023, the VC world started playing around with LLMs. I wrote about “how to 10x your productivity with ChatGPT”, “Top 10 AI Co-Pilots for VCs”, “how to create custom GPTs”, and a lot more, ending up with over 300 deep dives & how-to articles.
By 2024, we’ve gained a good understanding of model capabilities and created prompts to summarize decks and meetings, conduct competitive landscape analyses, assess founders, automate outreaches and interactions, draft investment memos, and a lot more.
The problem? We ended up with dozens of prompts that either needed to be copy & pasted into your chatbot of choice or put into CustomGPTs that then needed to be individually called, one at a time.
While we found a nice way to organize the most powerful prompts for deal sourcing, screening & evaluation, deal winning & closing, etc. in our VC prompt database (get it here), it’s still extremely inefficient to go back and forth with copy & pasting or calling dozens of CustomGPTs.
Fast forward to today, we stand at the edge of the next evolution: multi-agent systems and MCPs as connective tissue. I wrote about how to build your an MCP server for investors few weeks ago, adding another piece to the equation.
In today’s episode, I want to zoom out and look at how the different pieces fit together and how it will reshape the venture landscape.
Provokative thought: Will we see the first one-person billion-dollar fund?
What Are Multi-Agent Systems?
Multi-agent systems are networks of autonomous LLM-powered entities ("agents") that can:
Take on specialized roles (e.g., researcher, memo writer, planner)
Share context and collaborate with each other
Use external tools like web search, Notion, Airtable, CRMs, etc. to deeply integrate into existing stacks and workflows
Instead of prompting a chatbot over and over, you can now deploy a network of bots to:
Watch the market
Surface deals
Vet competition
Write the first draft of your memo
Flag key risks based on your investment thesis
Popular orchestration tools include LangGraph, CrewAI, AutoGen, and Dust. With the rise of open models (like open-weight Claude or Llama 3), we’re seeing an explosion of decentralized, composable agent systems designed for internal VC workflows.
VC Workflow Deconstructed: What Can Be Agentified?
Here’s how the traditional VC stack maps to agent-driven processes: