Requests for startups are a way for investors to communicate what might work and where they are actively looking for startup teams. It’s not perfect of course, and the majority of opportunities will come from entrepreneurs and not from investors, but it’s a good barometer of the current Zeitgeist in tech. This is a follow up to a similar post I published in 2024 and 2025 part 1 and part 2.
AI startups continued to attract record funding amounts in 2025. For example check out these 55 startups that raised over $100M or more in 2025. For the Spring 2026 batch, Y Combinator highlights a fundamental shift: AI-native companies can now be built “faster, cheaper, and with more ambition than ever”. While previous years focused on coding assistants, the focus has moved upstream to AI-native workflows, new financial primitives, and modernised industrial systems.
- Cursor for Product Managers: YC is looking for AI-native systems that help teams decide what to build, rather than just how to build it. While tools like Cursor assist with coding, there is a gap for tools that synthesize customer feedback, propose features, and create product requirements, effectively closing the loop on product discovery.
- AI-Native Hedge Funds: There is an opportunity to build the next generation of quantitative funds using swarms of AI agents. Instead of just adding AI to existing strategies, these funds would use agents to analyze financial filings and earnings calls to develop entirely new trading strategies.
- AI-Native Agencies: YC sees potential for agencies that operate with software-like margins by using AI to produce the work rather than just selling tools to clients. Examples include design firms, ad agencies, or law firms that use AI to deliver finished products instantly and at scale.
- Stablecoin Financial Services: With regulations like the GENIUS and CLARITY Acts opening a window, YC is interested in startups that bridge the gap between decentralized finance (DeFi) and traditional finance. This involves building compliant financial services, such as yield-bearing accounts or cross-border infrastructure, that utilize stablecoins.
- AI for Government: As AI tools allow the private sector to fill out forms and applications faster, government agencies need corresponding AI tools to process this influx of data efficiently. This sector offers “sticky” customers and large contracts for startups willing to navigate government sales.
- Modern Metal Mills: There is a request for “software-defined” mills, particularly in aluminum and steel, that utilize AI-driven planning and modern automation to reduce lead times and energy costs. The goal is to make domestic metal production cheaper, more flexible, and more profitable by replacing legacy systems.
- AI Guidance for Physical Work: YC is looking for startups that use multimodal models and wearables (like smart glasses) to provide real-time coaching for physical tasks. This technology would allow workers in fields like healthcare or manufacturing to “see, reason, and guide” their actions, effectively upskilling the workforce immediately.
A16Z’s Speedrun
Andreessen Horowitz is focusing on the “consumer AI renaissance” and creative tools, with a SR006 cohort running through April 2026:
- Vibe Creation Platforms: Moving from AI chat to “Vibe Coding” where agents build entire apps or visual stories (games, video) from a prompt.
- AI Companions in Cozy Simulations: Moving beyond chatbots into, persistent, multi-modal, agent-driven worlds (similar to a 3D Stardew Valley with AI).
- AI-Native Education: The first AI-native universities.
- Multiplayer AI: The shift from single-user AI tools to collaborative, multi-player AI agents.
- “The Year of Me”: AI that perfectly personalizes experiences to the individual user.
Conviction VC & Sarah Guo
Rather than a traditional static list, Conviction VC’s 2026 outlook emphasises the transition from “AI talkers” to “AI doers”. The theme for 2026 is Year of the Agent Harness:
- The Agent Harness: Startups building the infrastructure (the “harness”) that allows AI agents to operate reliably and autonomously in production environments.
- Economically Useful Evals: Moving beyond simple benchmarks to evaluation systems that prove real-world business ROI.
- 3D Generative AI: Systems that can generate functional 3D assets and environments without the “gamble” of unpredictable generative outputs.
- Energy-Efficient AI: Startups focused on “intelligence per watt,” squeezing maximum capability out of limited energy units.
I’ll continue to update this list with additional requests for startups published in 2026 by leading investors.
Part of VC Cafe’s Requests for Startups hub
For more investor theses, category maps, and related editions, visit the Requests for Startups hub. You can also read Requests for Startups in 2025 and Requests for Startups: Summer 2026 edition.
