Naïve Raises $28.5M Series A to Build Autonomous Company Infrastructure

Naive, a Palo Alto-based AI lab building the infrastructure AI agents need to run entire businesses, led by Sean Dorje and Dennis Zax, has announced a $28.5 million Series A led by Nexus Venture Partners, with participation from Y Combinator, Zetta, Liquid 2, and angel investors including Gokul Rajaram, Tim Zheng (Apollo), JD Sherman (former HubSpot), Gert Lanckriet (Amazon), Robert Chatwani (Docusign), and Zachary Sims (Codecademy).

The Problem: Apps Ship in an Afternoon, Businesses Take Months.

Coding agents such as Claude Code and Codex can produce a working application in hours. Turning that application into an operating business, however, remains largely manual, requiring developers to stitch together siloed APIs for payments, email, phone services, and incorporation, most of which were designed around human operators.

The deeper challenge is that today’s cloud infrastructure was built for human-driven software and was not designed for autonomous agents.

Naïve’s Solution: One Config File, One API, the Entire Operating Stack.

Developers can use a coding agent to generate a single configuration file, while Naïve provisions the underlying operating stack through one unified API, covering incorporation, payment cards, email, mobile phones, memory, models, and compute. A governance gateway oversees every action, enforcing budgets and approval requirements before execution and giving developers control over what agents can spend and do.

Key Platform Capabilities:

• Real-World Identity: KYC/KYB verification, LLC formation, email, phone numbers, and virtual cards, enabling agents to operate as distinct legal and economic entities.

• Cloud Infrastructure: Databases, hosting, compute, object storage and authentication.

• Serverless Agents: Runs Hermes, Claude Code, Codex and Vetta as lightweight, serializable state rather than dedicated machines.

• Multi-Agent Orchestration: Task management, agent-to-agent communication and scheduling.

• Inference & Model Routing: Routes requests to cost-efficient models, optimizes context, and reuses previously solved reasoning.

• Memory: A unified context layer that grows with the business.

• Third-Party Integrations: Connects with more than 10,000 business tools, including Stripe, GitHub, Supabase and QuickBooks.

• Governance: Enforces capability policies, maintains immutable audit logs and requires human approval for sensitive actions.

Abhishek Sharma, Partner at Nexus Venture Partners, said the past two years have demonstrated the potential of autonomous software, while the next decade could see the emergence of autonomous companies. He added that Naïve provides entrepreneurs and small businesses with turnkey infrastructure to build and operate AI-driven companies without requiring deep AI expertise.

Where the Capital Goes: Four Research Pillars.

The new capital will support frontier research focused on improving the efficiency of autonomous agent loops. Naïve’s research centers on four areas designed to maximize the work delivered per token spent:

• Serverless Runtimes: Minimize compute costs by running agents and their sandboxes as V8 isolates rather than dedicated virtual machines. With agents spending most of their time idle, the architecture charges primarily for active execution, with reported cold starts of 2.3 milliseconds and 1.2MB per agent.

• Inference: Reduce cost per token by routing requests to the least expensive model capable of handling each task, using smaller models for routine operations and batched inference for longer-running agents.

• Brain: Reduce input tokens through a unified memory layer that converts conversations into structured facts and provides agents with only the context required for each task. The system reportedly achieves comparable or better recall with around 11x fewer tokens per query.

• Orchestration: Reduce wasted agent turns by assigning tasks to the appropriate subagent and model tier rather than defaulting to frontier models. The orchestration layer is hosted serverlessly at a reported fraction of conventional costs.

Sean Dorje, CEO and co-founder of Naïve Labs, said agent spending could reach trillions of dollars over the next decade as companies increasingly rely on autonomous systems to complete work. He added that Naïve Labs aims to improve the efficiency of each token, helping make autonomous companies more economically viable.

𝐑𝐞𝐦𝐨𝐭𝐞 𝐇𝐢𝐫𝐞 𝐖𝐢𝐭𝐡 𝐔𝐬: 𝐌𝐞𝐧𝐥𝐨 𝐓𝐚𝐥𝐞𝐧𝐭 helps technology companies hire exceptional remote talent from India. Build your team with carefully curated engineers, product leaders, designers, GTM professionals, and more. Learn More At: https://www.menlotimes.com/menlo-talent