AI Infrastructure Startups Raise Seed Rounds
Venture capital continues to flow to AI infrastructure and agentic platform startups. Manufact, a company building MCP-based agent infrastructure, raised a $6.3 million seed round. In a related space, Cashmere, a generative AI publishing platform, secured $5 million in seed funding.
- Manufact is building infrastructure for AI agents based on the Model Context Protocol (MCP), an open standard from Anthropic for connecting large language models to external data and tools. The company provides an open-source SDK to help developers move agentic AI prototypes into production. - The seed round for Manufact was led by Peak XV and included participation from Y Combinator, Liquid 2 Ventures, and the co-founder of Supabase. The company's CTO, Enrico Toniato, is a former tech lead from IBM Research specializing in AI reasoning. - Cashmere provides an infrastructure layer that allows premium publishers to license, monitor, and monetize their content for use in AI systems. The platform gives publishers token-level access control and usage tracking to protect their IP while creating new revenue streams from AI companies. - Cashmere's lead investor, Reach Capital, is a venture firm with a background in education technology that has expanded its thesis to include workforce development and lifelong learning. Other investors in the round included content distributor Ingram Content Group and AI search company Perplexity, which was also Cashmere's first client. - The investment trend reflects a market shift from funding AI models themselves toward building the operational infrastructure for agentic AI. As enterprises move from AI pilots to production, the focus is turning to security, orchestration, and compliance layers that allow autonomous systems to operate at scale. - This focus on agentic infrastructure is driven by the complexity of moving beyond chatbots to AI systems that can perform tasks and complete workflows autonomously. Venture capital is concentrating on these foundational systems that are required to turn experimental AI into operational enterprise software.
