Worldbuild Launches $30M Fund I to Back Early-Stage AI and Infrastructure Startups
Worldbuild just launched Worldbuild I, a $30M debut fund for early-stage founders building when, as Sumeet Singh puts it, “conviction cannot be outsourced.” In venture, consensus is often just a very expensive way to arrive late.
Sumeet Singh, Founder and Managing Partner of Worldbuild, is backing category-creating companies and long-term “infinite game” builders. Not pitch-deck theater. Not spreadsheet cosplay. Builders with original thought, technical teeth, and the nerve to make polite committees check their pulse.
The fund began investing in January 2025, with the formal close announced in May 2026. Worldbuild Ventures I, LP is a Delaware limited partnership with $30M in committed capital across creative technology, AI infrastructure, compute, developer tools, and space energy. The hard corners. The strange corners. The corners where tomorrow shows up wearing a hoodie and a burn rate.
Sumeet Singh’s path gives the thesis scar tissue. BlackRock. Nyca Partners. Brigit, where Sumeet Singh led strategy after advocating for the company as an investor. Andreessen Horowitz, where Sumeet Singh was one of the first New York hires. Then came the obvious lesson: at the earliest stage, waiting for everyone to agree can turn conviction into committee theater.
Worldbuild has already partnered with founders across SF Compute, Browserbase, Truffle AI, Aetherflux, Interfere, Flower Computer Company, Fractal Power, Armory Medical, Company Picnic, Actual Computer, and Athanor. That is not just a portfolio. That is a map of where the next infrastructure arguments are going to get settled.
Look closely at the companies already inside the orbit. SF Compute is chasing AI compute infrastructure while Browserbase is building headless browser tooling for the AI era. Aetherflux is working on space-based solar power. Truffle AI is deep in AI agent infrastructure and developer tooling. Flower Computer Company sits in the strange and fascinating lane of environmental computing and consumer AI. None of these bets feel optimized for cocktail-party consensus. That is the point.
The business lesson is clean. Worldbuild did not raise by trying to be louder than the market. It raised around a point of view. Sumeet Singh publishes theses through Every, Notion, Discord, Reddit, and the Worldbuilders podcast, then lets original thinking attract the right founders before the room gets crowded.
Ben Casnocha of Village Global is a confirmed limited partner. A top-20 university endowment and several prominent multi-stage VC firms are also in the LP mix, though not publicly named. Institutions do not back taste because it sounds cute. They back it when taste starts looking like signal.
Worldbuild is the right name. The job is not to admire the future from a safe distance. The job is to find the people laying the first ugly, brilliant bricks while everyone else is still debating the floor plan.