Portkey raises $15 million to scale an in-path control plane handling $500K daily AI spend

Portkey said it raised $15 million in a Series A led by Elevation Capital, funds the San Francisco startup will use to scale a unified control plane that the company says already handles striking levels of production AI traffic and spend. The platform, Portkey says, processes more than 500 billion LLM tokens across 125 million requests per day and manages over $500,000 in AI spending daily for more than 24,000 organizations worldwide.

Krishna Mehra, a partner at Elevation Capital, framed the investment as a bet on an operational layer for enterprise AI: "Portkey is the only platform we’re aware of that provides this level of unified control across AI traffic at proven enterprise scale," Mehra said. "As AI becomes embedded across products and workflows, and agents begin taking actions autonomously, companies need one in-path system to keep AI reliable and accountable."

Portkey positions itself as an "in-path" LLMOps gateway where, in the firm's phrasing, "tokens become dollars." The product combines a high-performance gateway with governance, observability, reliability and cost-control features so engineering teams can enforce policies and maintain uptime while finance teams gain real-time visibility into model spend. Inc42 reported that Portkey offers a unified API gateway to more than 1,600 models, enabling routing, automatic fallbacks and load balancing across providers.

The company said it will use the Series A to expand product capabilities and scale go-to-market operations, with specific plans to enhance support for agent-based systems through additional controls around permissions, identity and budget management and to improve performance for low-latency uses. Portkey has made its core enterprise gateway available for free, the company told press outlets, intending to lower barriers for early governance and observability adoption.

Portkey names Postman and Snorkel AI among enterprise customers. The startup's claimed processing volumes and spend translate into material economic exposure for companies that rely on large language models: the reported daily spend of $500,000 extrapolates to roughly $182.5 million a year, a figure that mirrors an Inc42 estimate of more than $180 million in annualized LLM spending under management.

The funding builds on a $3 million seed round led by Lightspeed in August 2023. Inc42 says that with the new financing Portkey has raised more than $18 million to date; that total equals the published seed and Series A figures, though reporting on precise totals differs slightly among outlets. Sources also differ on founders. Inc42 lists Ayush Garg and Rohit Agarwal as founders of the 2023 startup, while Entrackr refers to Rohit Agarwal as founder; company materials distributed via GlobeNewswire provided the Series A details.

Portkey's pitch underscores a broader enterprise problem: as AI moves from experimentation to production, unpredictably rising API bills, model failures and the prospect of autonomous agents acting without adequate guardrails create both financial and operational risk. A single in-path control plane that can meter, route and police model traffic promises to centralize those safeguards, and investors appear willing to bet on that consolidation.

The company distributed its announcement via a paid GlobeNewswire release and has supplied the usage and scale figures cited in its materials. Portkey said the new capital will accelerate engineering and commercial expansion as enterprises grapple with the costs and governance of production AI.