Palo Alto-based Naïve Inc., an artificial intelligence lab that’s developing autonomous agents capable of setting up and running entire businesses, said today it has closed on a $28.5 million Series A round of funding. Nexus Venture Partners led the round, which saw participation from Y Combinator, Zetta, Liquid 2 and a host of angel investors.

The startup is all about automation, noting that AI tools that can lighten people’s workloads have proven to be especially popular with people such as developers, enabling them to avoid repetitive and tedious tasks. Now it wants to advance this and use AI automation to perform basically all of the work that’s required to get an established business up and running. It’s an enticing concept, and so far Naïve has already attracted more than 30,000 developer customers to its platform.

Naïve goes further than traditional vibe coding. It has developed the infrastructure for AI agents to automate all of the steps required to establish a company and then run it on a day-to-day basis, so long as the customer provides the actual agents and the token budget. Essentially, it packages together the process of creating payments infrastructure, email accounts, phone numbers, cloud infrastructure, storage and business incorporation, making it all available through a simple application programming interface.

The company supplies a set of prompts that developers can use with third-party agentic tools such as Claude Code, Codex and Cursor. When prompted, these agents will connect to Naïve’s API to provision all of the infrastructure and other things needed to establish a company. For instance, it can tell agents how to orchestrate the creation of a U.S.-based limited liability company, providing details such as the state it’s based in, a business description, industry code and a proposed name. However, the owner of the business must still perform the Know Your Customer and Know Your Business processes themselves and make any payments that are required.

But just about everything else can be automated by AI agents. This includes tasks such as setting up email inboxes, phone numbers, databases, compute resources and links to payment processors like Stripe and accounting platforms like QuickBooks. Naïve has also built a comprehensive governance layer that can be used to restrict the capabilities of user’s AI agents, establish budgets and decide which actions require human approval before they’re taken by the agents. In addition, there are templates for things such as AI SEO, full-stack software-as-a-service applications, customer support, recruiting and accounting. It even has a mobile emulator that can be used by AI agents to navigate smartphone apps on virtual devices.

Naïve says it has enjoyed impressive revenue growth, with sales increasing more than tenfold over the last six months, achieving an annual revenue run-rate in the low double-digit millions. Customers have used its agentic infrastructure to create and run a range of autonomous businesses, including content channels on apps such as TikTok and YouTube, car rental firms that are fully managed by AI agents and more.

However, Naïve believes that while this toolkit is extremely useful to entrepreneurs, it has an even bigger opportunity in terms of reducing the costs associated with AI automation. That’s important, because while an almost fully-automated business might sound great to the person that owns it, the cost of running those agents can escalate dramatically as they make hundreds of calls to expensive models, pass around large volumes of context and eat up resources while sitting idle.

That’s why Naïve is also looking for ways to improve the efficiency of AI agents. Its innovations include a model router that instructs agents to always choose the most cost-effective AI model for each task and a memory system that allows them to store and reuse business context. It has also built a serverless runtime environment that allows agents to run in lightweight JavaScript environments, further reducing costs. In this way, it says it can help to slash the operational expenses of automated businesses.

Co-founder and Chief Executive Sean Dorje said this kind of optimization is vital because agent spending is likely to grow to trillions of dollars in future as more companies embrace automation. “Our vision at Naïve is to make each token do more, so autonomous companies can become a cost-efficient reality,” he said.

Ultimately, this could turn out to be a much more valuable business for Naïve than simply helping company founders automate the grunt work of setting up a new company. The platform will certainly be useful in getting a business up and running, but once that’s done, most entrepreneurs will hope to expand and grow that business for years to come. If Naïve can significantly reduce the cost of running the hordes of agents that automate their businesses, those customers will likely want to keep using its platform for as long as their companies are around.

Abishek Sharma of Nexus Venture Partners said autonomous software has already become an established fact, and the next step is to have fully autonomous companies. “Naïve gives millions of entrepreneurs and small businesses worldwide the turnkey infrastructure needed to build and run autonomous companies without needing to become AI experts,” he said.

Image: Naïve