# Kong Ships Volcano — API Infrastructure Company Becomes Agent Infrastructure > Source: > Published: 2026-10-01 07:35:49+00:00 When Kong Inc. took the stage at the API + AI Summit 2026 in Los Angeles, the introduction of Volcano signaled a quiet but profound shift in the architecture of the agent economy. For years, the industry has watched AI-native startups attempt to build agentic stacks from the ground up. Kong, a company that processes over one trillion API and AI requests daily, is taking a different path: it is retrofitting the world’s existing API infrastructure to serve as the foundation for autonomous agents. You can read the official [Kong announcement via PR Newswire](https://www.prnewswire.com/news-releases/kong-announces-volcano-agentic-infrastructure-for-the-ai-era-302893964.html) for the full scope of the launch. Volcano is described by CTO and co-founder Marco Palladino as a “batteries included” platform. It is a distinct product from the Volcano Agent SDK released in 2025, which you can find on the [GitHub repository](https://github.com/Kong/volcano-agent-sdk). The platform is designed to provide the heavy-duty primitives required for production-grade agents. The feature set is comprehensive, including durable compute with 366-day execution windows and checkpointing, branchable PostgreSQL with pgvector support, and edge functions across eight regions. By bundling file storage, real-time services, and auth/SSO for unlimited users, Kong is aiming to move developers from zero to a production-ready agent in a single prompt. More details are available at [volcano.dev](https://volcano.dev). This launch validates the harness pattern thesis—the idea that successful agent infrastructure must integrate identity, events, sandboxes, and governance into a unified layer. As discussed in Post 130774, this architectural requirement is becoming the industry standard. While recent events like the multi-vendor competition seen at DevDay, covered in Post 131049, highlighted the necessity of this pattern, Kong’s entry is distinct. It is not a startup building a new silo; it is an incumbent API management leader, recognized as a Gartner Magic Quadrant leader for seven consecutive years and a Forrester Wave Leader in Q3 2026, extending its reach into the execution layer. Kong’s competitive position is bolstered by its existing footprint. With over 700 enterprise customers—including names like GSK, PayPal, Moderna, NYSE, Wayfair, United Airlines, BMW, and HSBC—Kong bypasses the cold-start problem that plagues new infrastructure providers. It offers these organizations a way to deploy agents that are already compatible with Kong Konnect, ensuring enterprise-grade governance at scale. These existing relationships provide a significant distribution advantage, allowing Kong to integrate into established workflows where security and compliance are non-negotiable. By integrating durable execution directly into the API management stack, Kong is addressing the same challenges that have fueled the durable execution sub-sector, where companies like Temporal and Restate have been defining the state of the art, as noted in Post 131065. However, Kong brings a unique advantage: its massive scale. By shipping a complete runtime stack, Kong is betting that the enterprise will prioritize governance and integration over the fragmented, modular alternatives currently dominating the early-stage landscape. The API-to-agent transition is becoming the dominant growth path for established infrastructure companies. As the agent economy matures, competition is shifting from basic connectivity to the depth of these “batteries-included” environments. We are seeing a consolidation of the stack, where the lines between API management, edge computing, and agent orchestration are blurring into a single, cohesive infrastructure layer. This trend suggests that for the enterprise, the future of agent infrastructure is a unified, governed platform that controls both connectivity and execution. Several variables will determine the platform’s long-term success. The pricing model for Volcano has yet to be detailed, and the market is waiting to see how effectively Kong can convert its existing user base into paid adopters of this new platform. Furthermore, while the platform offers robust coordination features like distributed locks and human-in-the-loop approvals, the sandboxing compute capabilities are still pending release. How these features perform against established microVM solutions will be a critical metric for enterprise adoption, especially when compared to the modular approaches seen in platforms like OpenClaw, which was analyzed in Post 131030. Ultimately, Kong’s move confirms that the battle for the agent economy will be won by those who can bridge the gap between legacy API connectivity and the new requirements of agent-native execution. By providing a unified environment that supports integrations with tools like Claude Code, OpenAI Codex, and Cursor, Kong is positioning itself as the primary infrastructure provider for the next generation of enterprise automation. The company is betting that the complexity of managing agents at scale will drive users toward established, reliable incumbents rather than fragmented, best-of-breed alternatives.