Portkey just pulled in $15M in Series A funding, and if you work anywhere near production AI, you might want to sit up a little straighter. Founded in 2023, Portkey.ai has been moving with intent. Rohit Agarwal, Co-founder and CEO, operating out of San Francisco, and Ayush Garg, Co-founder and CTO, building from Bengaluru, saw something early. Everyone was racing to build AI apps. Few were thinking about what happens when those apps actually matter. When they touch revenue. When they touch customers. When they break.
So Portkey built the control plane. Not a shiny wrapper. Not another chatbot demo. A unified control plane for production AI. An in-path AI gateway that sits where the real traffic flows. Every request. Every token. Every model call. Watched. Routed. Governed. Measured.
The result is simple in theory and complex in execution. Connect to multiple large language models through one operational layer. Monitor performance. Track costs. Enforce policies. Manage prompts. Improve applications without crossing your fingers and hoping your cloud bill behaves. Enterprises are learning the hard way that AI without guardrails is just expensive curiosity. Portkey steps in where experimentation ends and accountability begins.
Investors noticed. Elevation Capital led the $15M Series A. Lightspeed doubled down after leading the $3M seed round. That brings total funding to $18M. Not hype capital. Infrastructure capital. The seed round also included senior executives from OpenAI, Cloudflare, Postman, and Asana. That kind of cap table does not chase noise. It backs picks and shovels when a gold rush starts overheating. And Portkey is very much in the picks and shovels business.
Millions of requests per day already flow through the platform, according to coverage from The Economic Times. That is not a sandbox. That is live fire. The kind of scale where latency, reliability, and cost visibility are not technical details. They are boardroom topics.
The timing makes sense. Enterprises are moving from playing with AI to depending on AI. Customer support. Internal operations. Developer tooling. When AI becomes mission critical, someone has to own the traffic control tower. That is the quiet power of a control plane. It does not scream for attention. It just makes sure the planes land.
Rohit Agarwal and Ayush Garg are not selling magic. They are selling discipline in a market that desperately needs it. Governance. Observability. Cost control. The unglamorous backbone of serious AI adoption.
This round is not just capital. It is validation that the next chapter of AI is not about bigger demos. It is about better systems. If you are running AI in production and still treating it like a science experiment, the question is not whether you need a control plane.
