AI agents, the autonomous bots designed to act on behalf of users, are quickly moving from science fiction to everyday reality. But while the technology underpinning these AI agents has matured rapidly, the infrastructure that connects them to digital services remains fragmented and outdated.
That gap is exactly what Paris- and San Francisco-based startup Alpic aims to address, having closed a $6 million pre-seed round led by Partech and a consortium of investors with deep ties to AI and cloud infrastructure.
“The real potential of agents lies in their ability to interact with the digital world around them. For that, we need infrastructure built from the ground up for agents, not retrofitted for them,” said Pierre-Louis Theron, co-founder and CEO of Alpic.
The investment will fuel the development of Alpic’s platform, which already counts dozens of MCP server deployments with early customers and has entered public beta. The timing reflects broader market momentum: as AI agents increasingly handle tasks from booking flights to managing CRM data, enterprises need scalable, reliable infrastructure to support this new user layer.
The story behind Alpic
Founded by Pierre-Louis Theron (CEO), Nikolay Rodionov (COO), Frédéric Barthelet (CTO), Charles Sonigo (CPO), and Erica Beavers (CMO), the team combines deep expertise in infrastructure, cloud computing, AI, and media technologies. This experienced core has a proven track record from their previous venture, Streamroot, through acquisition and leadership roles in the media and cloud sectors.
“Authentication, user experience, payments, and developer workflows look completely different when the user is an AI agent. We’re only beginning to understand how much the internet will change. Alpic’s mission is to provide the infrastructure to make those connections seamless,” added Theron.
Reflecting on the motivation behind the startup, Theron shares with TFN, “We kept seeing the same pattern in early agent projects: great demos, then brittle integrations in production that hardly worked. Teams were wiring models to human interfaces, which was slow, fragile, and hard to secure. MCP provides the right abstraction, yet companies still need a reliable way to deploy, secure, and observe this new agent interface.
We started Alpic to make agent access safe, measurable, and easy to run at scale. Our previous experience building and operating large-scale video delivery infrastructure helped us identify this emerging need and the inefficiency of current approaches.”
From the perspective of a woman in tech, Beavers notes, “Speaking personally: progress is real, but bias still surfaces, often around technical credibility. What has made the biggest difference for me is learning that confidence from others starts with confidence in yourself. If you trust your skills, your product, and your vision, it becomes much easier to carry that conviction into every meeting. Surround yourself with people who value your contribution and respect outcomes over optics.”
The world’s first cloud platform purpose-built for the MCP
Alpic is building the world’s first cloud platform purpose-built for the Model Context Protocol (MCP), an open standard gaining traction among leading AI players for enabling secure, direct communication between AI agents and external services.
Theron explains to TFN: “Alpic is the first full-service cloud platform built specifically for the Model Context Protocol. We give developers everything they need to make their services agent-accessible in one place (hosting, security, analytics, and operations), without requiring new frameworks or custom plumbing.”
He adds, “What sets us apart is the developer experience. We abstract away deployment and infrastructure so teams can spin up MCP servers in minutes and focus on building useful tools. Security and observability are built in from the start, so companies can run agent workflows with confidence.”
The startup’s offering allows businesses and developers to deploy MCP servers quickly, with built-in security frameworks and analytics tailored to agent interactions, bypassing the brittle and inefficient workarounds of AI having to “navigate” human-centred interfaces.
Alpic’s technology differentiates itself by focusing not just on connectivity but also on operational ease and security, offering one-click deployment, fine-grained authentication models, and observability features tailored to production environments.
Boris Golden, General Partner at Partech, adds, “Agent-first protocols like MCP could be as foundational to AI as HTTP was two decades ago to the internet. Alpic has the deep infrastructure experience and timing to shape how they will be adopted.”
The startup’s approach foresees a future where AI agents will routinely interface with the digital world directly, replacing many traditional user interfaces.
