Naïve, a startup building infrastructure for AI agents, has raised $28.5 million in a Series A funding round. Nexus Venture Partners led the round.
The company gives AI agents the tools to set up and run a business. This includes payments, email accounts, phone numbers, and cloud storage.
Users access these tools through a single API. Developers connect the API to tools like Cursor, Claude Code, or Codex.
The system can form a U.S. LLC by filling in details like state, industry code, and business description. People still need to complete identity checks and handle payments themselves.
How the Platform Works
Once a company is formed, AI agents can set up email inboxes, virtual cards, and phone numbers. They can also connect to services like Stripe and QuickBooks.
Naïve includes a governance layer. This lets users set spending budgets and limit what actions an agent can take without approval.
The company offers templates for different business types. These include AI search optimization, software apps, recruiting, accounting, and customer support tools.
One template lets agents operate smartphone apps using an emulator. This expands what tasks the agents can complete on their own.
Naïve says more than 30,000 developers have used its platform since launch. Annual revenue has grown ten times over in the past six months, according to CEO Sean Dorje.
Dorje said customers use the platform to run AI automation agencies and online content channels. Some post AI-generated videos to platforms like TikTok and YouTube.
He said one customer runs an entire rental car agency without direct human involvement. Automation agencies are currently the fastest-growing category among users, he added.
Cutting the Cost of Running AI Agents
Running AI agents around the clock can get expensive. Costs rise as agents call AI models, process data, and stay active while waiting for tasks.
Naïve plans to use part of the new funding to lower these costs. It is building a model router that sends each task to the most efficient AI model available.
The company is also developing a memory system. This stores business data so agents can access it without repeating work.
Naïve is building a serverless runtime that runs agents in lightweight environments instead of full virtual machines. This setup lets customers pay only when an agent is active.
Dorje said demand for this infrastructure is growing quickly. He said the cost of running agents is often a business's biggest expense.
He said interest is also coming from larger companies, though he did not name specific firms.
Naïve has 10 full-time employees. The company plans to use the funding to hire researchers and expand four projects: agent sandboxes, model routing, memory systems, and governance tools.
Other investors in the round included Y Combinator, Zetta, Liquid 2, and several angel investors. The round brings Naïve's total funding to about $32 million.
