Singapore – September 12, 2026 -- Accels has launched a combined AI infrastructure and financial services platform aimed at businesses building autonomous AI agents, betting that token consumption will replace seats and subscriptions as the core unit of AI economic activity.
Goldman Sachs projects a 24-fold surge in global AI token consumption by 2030
Separate industry forecasts cited by the company put annual token-related spending rising from roughly $200 billion today to as much as $2.5 trillion a year, driven largely by AI agents operating continuously rather than responding to occasional human queries.
Accels bundles model routing, token lending and merchant acquiring into one platform
The company's API gateway connects developers to 150+ AI models from 15 leading providers, spanning both major international and domestic Chinese model families. Its financial products include payment terms allowing customers to pay for token usage on account rather than upfront, plus compute credit lines -- termed "compute loans" -- that give businesses access to AI compute and token capacity without large upfront cash outlays.
Merchant acquiring services target billing complexity for AI-native businesses
The platform's third pillar lets businesses accept payments and manage billing globally through a single system, a move Accels positions as removing commerce friction for companies scaling agentic applications. The company argues that as the number of active AI agents multiplies, each will consume far more tokens per unit than individual human users historically did, reshaping demand for both compute infrastructure and the financial tools needed to fund it.