Interview
How Kong Was Born: APIs, Hustle, and the Future of AI Infrastructure
- The company aims to achieve an IPO at a future, undefined point in time.
- Market dynamics are expected to shift by 2030, with machines consuming the internet programmatically through interfaces rather than human-centric browsing.
- API traffic and AI traffic are predicted to converge within a two-to-three-year timeframe as an evolutionary step.
- Intelligence is projected to be monetized via tokens, with revenue releases centered on token usage.
- A unified API and AI connectivity platform designed to manage classic API, agent, MCP, and LLM traffic is planned for development over the next two to three years.
- MCP traffic is anticipated to grow on a quarterly basis in the near future.
- Enterprises are expected to transition from relying on a single large language model (LLM) to utilizing hundreds of small, medium, and large LLMs.
- A governance pattern for the LLM space is predicted to emerge, abstracting rate limiting, token authentication, and connectivity logic into a gateway layer similar to the microservices model.
- Strategic AI connectivity infrastructure is expected to become a standard component in next-generation apps, internal tools, and customer-facing applications.
- Building the company is viewed as a long-term endeavor requiring 10 to 20 years to grow into a trend, often taking longer than anticipated.
- A low burn rate will be maintained during early operations to prevent capital depletion.
- Failure to persist is characterized as an unacceptable outcome resulting in a return to Italy.
- Historical market consolidation in the API space, such as the 2016 acquisitions of MuleSoft and Apigee, is noted as a precedent preceding the subject company's breakout.
- Following the late 2016 Series B round, the company's annual recurring revenue (ARR) grew from under $1 million to $10 million within a single year.
- An annual Founders Award of 2,555 stock is planned to continue as a symbolic ritual commemorating seven years of struggle.
- Adoption of "banal use cases" including key management, monitoring, and billing alongside agentic workflows is expected.
- Automation of key provisioning, rotation, and authentication is anticipated to enable AI agents to operate freely without access control constraints.