Naïve has raised $28.5 million in a Series A led by Nexus Venture Partners to build infrastructure that lets AI agents provision and operate entire companies. The startup signed more than 30,000 developer customers within months of launch and grew annual run-rate revenue 10x over the past six months to the low double-digit millions, according to CEO and co-founder Sean Dorje. The round brings total capital raised to roughly $32 million and lands with a team of just 10 full-time employees.
The product packages the tedious plumbing of starting a business — U.S. LLC formation, Stripe accounts, QuickBooks, email inboxes, virtual cards, phone numbers, databases, and compute — behind a single API. Developers hand a prompt to tools like Cursor, Claude Code, or Codex, which then call Naïve's API to stand up the stack. A human still has to complete KYC and KYB checks and clear payments, but the rest of the setup runs through the agent.
A governance layer sits on top of the runtime. Customers set budgets, restrict what their agents can do, and require human approval for sensitive actions. Naïve also ships templates for common use cases — AI SEO, full-stack SaaS apps, recruiting, accounting, customer support — and a mobile emulator that lets agents operate smartphone apps on virtual devices.
Key facts
- 01Naïve raised $28.5M in a Series A led by Nexus Venture Partners, bringing total capital to roughly $32M.
- 02The startup signed more than 30,000 developer customers within months of launch.
- 03Annual run-rate revenue grew 10x over the past six months to the low double-digit millions.
- 04The team is 10 full-time employees; Y Combinator, Zetta, Liquid 2, and Gokul Rajaram joined the round.
- 05Naïve packages LLC formation, Stripe, QuickBooks, email, phone numbers, and cloud infrastructure behind a single API for AI agents.
Dorje said the customers using the platform range from AI automation agencies and faceless TikTok and YouTube content channels to at least one autonomous rental-car business. In one case, Naïve's infrastructure was powering a TikTok account posting AI-generated videos of cats and dogs dancing and boxing.
The autonomous-company toolkit is the front door, but the more interesting bet is on the back end. Running agents at any scale gets expensive fast — model calls, context passing, and idle compute all pile up — and Naïve is using the new capital to build infrastructure aimed at that cost line.
Four projects are in flight: virtualized sandboxes for agents, a model router that sends queries to the cheapest sufficient model while replaying already-reasoned data, a memory layer that stores and surfaces business context on demand, and an orchestrator that divides work among agents. The company is also building a serverless runtime that runs agents inside lightweight JavaScript environments rather than assigning each a full virtual machine, so customers pay primarily when agents are active.
Dorje said inference optimization and serverless agents are the fastest-growing source of demand on the platform, and that enterprises — which he declined to name — are already showing interest. That is where the larger business likely sits. Founders using Naïve to skip the paperwork of incorporation are a top-of-funnel motion; enterprises running fleets of agents in production care about the recurring bill.
Y Combinator, Zetta, Liquid 2, and angel investors including Gokul Rajaram, Apollo.io co-founder Tim Zheng, and former HubSpot COO JD Sherman also participated in the round. Proceeds will go toward hiring researchers and pushing the four infrastructure projects forward.
The obvious risk is concentration. A platform that provisions bank connections, corporate identity, phone numbers, and compute for tens of thousands of semi-autonomous agents is a target for abuse — fraud rings, spam operations, and low-quality content mills are exactly the kind of customer that finds this workflow attractive. Naïve's governance controls and KYC handoff are the first line of defense, but the trust-and-safety surface area scales with the customer base, and the company is small.
The bet worth watching is whether Naïve becomes the plumbing layer for a category of software that does not exist yet: businesses whose entire operations, not just their code, are written by agents. If inference costs keep falling and agent reliability keeps rising, the operators who win will be the ones who reduce the marginal cost of running each agent to something close to zero. A 10-person team with $32 million and a serverless runtime is a reasonable shape for that fight — provided the enterprise pipeline Dorje hinted at actually converts.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
