Pilot Protocol has emerged from stealth with $4.5 million in seed funding to develop a network designed specifically for autonomous artificial intelligence agents. Version One Ventures led the round, with participation from Precursor Ventures, Night Capital, Todd & Rahul Capital, and angel investors Lenny Rachitsky and Ben Tossell. The San Francisco-based company is positioning its platform as infrastructure that allows agents to discover, communicate, and transact with one another without continuous human coordination.
Building a Network for Autonomous Agents
Most AI agents currently operate within isolated environments, tied to a single user, machine, or software stack and dependent on conventional websites for information and services. Pilot Protocol seeks to address that fragmentation through a Network OS that assigns each agent an address and enables direct interaction with trusted agents. The company argues that this model can reduce repetitive web searches, lower computational costs, and allow useful capabilities to move between specialized systems.
The founding team includes Razvan Roman, Artemii Amelin, and Teodor Ioan Calin, with Roman serving as co-founder and chief executive. Roman previously built commerce infrastructure company TwoTap, which was acquired by Honey before Honey was acquired by PayPal. That experience in online transactions and platform development informs Pilot’s effort to create an economic and communications layer for software agents.
Early Network Growth
Pilot says approximately 250,000 agents are connected to its network and collectively generate about two billion requests each day. According to the company, the network expanded by as much as 10 percent daily during its early development, including one period when 16,000 agents joined within 24 hours. It also reports that roughly 70 percent of connected agents now begin tasks through Pilot rather than first using a traditional search engine.
The company attributes this growth largely to autonomous adoption rather than conventional customer acquisition. Pilot says agents discovered the software on the open web, installed it through a single line of code, and recommended it to other agents without direct marketing. Version One Ventures said this self-directed onboarding created unusually rapid network effects because software participants can continuously find, recruit, and interact with one another.
App Store and Agent Payments
Alongside its public launch, Pilot has opened an App Store where companies can publish tools that agents independently discover and use. Available capabilities include communication services, company and people searches, and coding environments, while packages are reviewed before distribution to reduce risks associated with compromised code. Pilot reported more than 30,000 autonomous installations of partner applications during the store’s first two weeks.
Pilot is also introducing wallets and agent-to-agent payments using the x402 standard. The system is intended to let agents spend credits on tools or services while receiving value for useful work performed for other participants. This approach could support a marketplace where companies provide agent-native products, agents purchase resources directly, and software systems earn credits through productive activity.
Market Opportunity and Challenges
The funding arrives as technology companies increasingly explore agents capable of completing multistep tasks with limited supervision. Pilot believes a dedicated network will become necessary as autonomous systems grow, while citing projections that US agent-driven commerce could reach between $300 billion and $500 billion by 2030. Adoption will also depend on security, identity verification, payment controls, interoperability, and whether organizations permit agents to install services or transact independently.
Pilot Protocol’s seed round gives the company financial backing as it moves from experimental network growth toward a broader commercial platform. Its early figures suggest demand for infrastructure connecting autonomous agents, but the durability of that activity and reliability of self-reported metrics require continued scrutiny. The next phase will test whether its network, application marketplace, and payment tools can become foundational infrastructure for an emerging agent economy.