Naïve, a startup that provides infrastructure enabling AI agents to perform most of the work involved in creating and operating a business, has closed a $28.5 million Series A round led by Nexus Venture Partners. With prior financing, the company’s total capital raised is now roughly $32 million, according to information obtained by TechCrunch.
What the product does
Naïve packages many of the tasks required to stand up and run a business behind a single API: company formation (for example, preparing to form a U.S. LLC), payment and virtual card issuance, email accounts, phone numbers, cloud resources, storage, and integrations with services like Stripe and QuickBooks. The company supplies a prompt that developers can feed to tools such as Cursor, Claude Code, or Codex; those tools can call Naïve’s API to provision the necessary infrastructure.
While Naïve can orchestrate steps such as preparing formation details (state, industry code, business description and proposed names), customers still must complete KYC/KYB verifications and any required payments.
What agents can automate
According to Naïve, configured AI agents can handle:
- creating email inboxes and virtual cards
- provisioning phone numbers and databases
- allocating compute resources
- connecting to Stripe, QuickBooks and similar services
- using business templates (AI SEO, full‑stack SaaS apps, recruiting, accounting, customer support)
- operating mobile apps on emulators
The platform also includes a governance layer to set budgets, restrict agents’ permissions and require human approval for sensitive actions.
Who is using it
Sean Dorje, Naïve’s co‑founder and CEO, told TechCrunch that over 30,000 developers had signed up within months of launch. Customers are using Naïve to run autonomous businesses such as AI automation agencies, “face‑less” content channels on TikTok and YouTube, and even an automated rental‑car agency. Dorje cited an example of a TikTok channel posting AI‑generated videos of dancing and boxing cats and dogs that was running on Naïve’s infrastructure.
The company says it scaled annual run‑rate revenue by 10x to the low double‑digit millions over the past six months.
Where the new capital will go
Naïve plans to use the Series A proceeds to reduce the cost of running agents and improve efficiency. The key engineering projects named by the company include:
- a model router that directs queries to the most cost‑effective model for a task while preserving and replaying previously reasoned data
- a memory system that stores and surfaces business context as agents need it
- an orchestrator to divide work among agents
- a serverless runtime that executes agents inside lightweight JavaScript environments instead of assigning each agent a full virtual machine, so customers pay mainly when agents are active
Dorje said that while the automated company setup features are currently in high demand, optimizing inference costs and serverless agent execution are among the fastest‑growing sources of customer demand.
Team and investors
Naïve currently employs 10 full‑time people. The company will hire researchers and continue development on four infrastructure projects: virtualized sandboxes for agents; model routing and inference optimization; a memory layer; and governance and orchestration.
The Series A was led by Nexus Venture Partners and included participation from Y Combinator, Zetta, Liquid 2 and angel investors such as Gokul Rajaram, Apollo.io co‑founder Tim Zheng, and former HubSpot COO JD Sherman. With this round, Naïve’s total funding is about $32 million.
Why it matters
Naïve’s tooling lowers the friction of launching agent‑driven businesses and offers templates and governance that speed up deployment. The bigger commercial test will be whether the company can materially reduce the recurring cost of operating fleets of agents — a problem that becomes especially important as organizations scale and as inference costs rise.
