A startup called Naïve wants AI software to run your entire company, and 30,000 developers signed up to try it
Naïve just raised $28.5 million to let AI agents handle the paperwork, phone numbers, payments and cloud setup that normally eat a founder's first weeks. The real prize, though, is cutting the runaway cost of keeping those agents running.
Key points
- Naïve raised $28.5 million in a Series A funding round led by Nexus Venture Partners, bringing total capital raised to roughly $32 million.
- More than 30,000 developers signed up within months of the company's launch.
- Annual run-rate revenue, the amount the company would earn if current monthly income held steady for a year, grew 10x to low double-digit millions in six months.
- Naïve has 10 full-time employees and plans to use the funding to hire researchers and build out four infrastructure projects.
- Customers are already using the platform to run autonomous TikTok channels, AI automation agencies and at least one rental car agency.
Setting up a new business means weeks of tedious admin: registering a company, opening bank accounts, buying phone numbers, wiring up payment systems, spinning up computer servers. A two-year-old startup called Naïve wants AI software agents, programs that can carry out long chains of tasks on their own, to handle nearly all of it without a human touching a keyboard.
The company packages that whole setup process behind a single API, a connecting point that lets software talk to other software. A developer feeds a short instruction to an AI coding tool like Cursor, Claude Code or OpenAI's Codex, and the tool reaches through Naïve's API to register a U.S. LLC (a limited liability company, one of the most common business structures in the United States), create email inboxes, issue virtual payment cards, claim phone numbers, set up databases and link accounting tools like QuickBooks and payment tools like Stripe.
Humans still need to complete identity checks and sign off on payments. Everything else, the company says, the AI can handle.
What does this actually look like in practice?
Customers are already running some unusual businesses on top of it. TechCrunch first reported that CEO Sean Dorje found one customer's TikTok channel using Naïve's infrastructure to post AI-generated videos of cats and dogs dancing and boxing. Others run AI automation agencies, selling agent-powered services to small businesses. One customer operates a rental car agency with almost no human staff.
"The first business that a lot of people start is genuinely just selling agents to other small businesses," Dorje said.
A governance layer lets users cap spending, limit what their agents are allowed to do, and require a human to approve anything sensitive before it happens.
Why did investors put in $28.5 million?
The traction is hard to ignore. Thirty thousand developer customers in the first few months, and revenue up 10x in six months, convinced Nexus Venture Partners to lead the round. Y Combinator, Zetta, Liquid 2 and several angel investors also joined.
But the bigger prize may not be the company-setup toolkit at all.
Running AI agents around the clock gets expensive fast. Every time an agent calls a large language model (the AI brain behind tools like ChatGPT), that call costs money. Pass enough data between enough tasks, and costs spiral. Naïve is building four pieces of infrastructure designed to cut those costs:
| Project | What it does |
|---|---|
| Model router | Picks the cheapest AI model capable of each task |
| Memory layer | Stores business context so agents don't re-read the same information repeatedly |
| Orchestrator | Splits work intelligently across multiple agents |
| Serverless runtime | Runs agents in lightweight code environments instead of full virtual machines, so customers pay mainly when an agent is actually working |
Dorje says demand for that cost-reduction infrastructure is now growing faster than demand for the company-setup tools. Enterprises with existing businesses are already asking about it, even if he would not name any.
What does this mean for ordinary people?
For now, Naïve's customers are mostly technical founders and developers. But if the cost of running an AI agent drops far enough, more small businesses could eventually hand over routine operations: customer emails, scheduling, accounting updates. The guardrails Naïve is building, the approval gates and spending caps, will matter a lot when that happens. An agent with an unlimited budget and no oversight is a very different product from one a small shop owner can actually trust.
