Kong turns gateway into multiple budgets

Kong is trying to turn one control point into many product budgets. A platform team that already routes REST APIs through Kong can buy adjacent modules to govern LLM calls, MCP tool access, agent to agent communication, and Kafka event streams, without swapping out the core gateway. That matters because each traffic class is usually owned by a different internal team with its own security, platform, or AI spend.

• Kong built AI Gateway on top of the same gateway runtime and plugin model used for API traffic. In practice that means an existing Kong customer can extend from HTTP API policy enforcement into model routing, token controls, observability, and AI specific guardrails, instead of adopting a separate AI ops stack.
  3 konghq 7 konghq 8 konghq
• The newer products map to distinct traffic types. MCP Registry governs which tools agents can discover and use. Agent Gateway governs agent to agent messages. Event Gateway governs Kafka style event streams. Context Mesh turns existing APIs into agent ready tools, then registers them back into the same catalog.
  2 konghq 4 konghq 5 konghq 6 konghq 9 konghq
• This also expands who signs the PO. Traditional API gateway spend often sits with platform engineering. AI gateway and MCP controls can pull in AI platform and security budgets, while event governance can attach to data infrastructure spend. Competitors like Gravitee and WSO2 are pitching similar consolidation, which makes breadth a key selling point in renewals and expansions.
  1 sacra 10 sacra

The next step is for gateway buying to follow traffic growth inside the enterprise. As more work moves from app to model, tool, agent, and event flows, the vendors that already sit in the request path have the cleanest shot at adding new governance layers and turning one infrastructure foothold into a wider control plane.