Doubao's Daily Token Usage Exceeds 180 Trillion, Growing 10 Times Year-on-Year; Volcano Engine Leads

According to monitoring by Dongcha Beating, at the FORCE Power Conference held on June 23, 2026, Volcano Engine revealed that as of June this year, the daily average Token usage of the Doubao large model has surpassed 180 trillion, growing more than 10 times in the past year, making it the most widely implemented model in China's industry. In terms of usage breadth, Volcano Engine holds the leading position in China's public cloud MaaS service market with a 49.5% market share. In terms of usage depth, over 200 enterprises have cumulatively used more than 1 trillion Tokens on Volcano Engine in the past year, with a doubling in usage over the last six months. ByteDance CEO Liang Rubo emphasized at the conference that climbing the AI peak is currently the most important task for ByteDance, and the Volcano Engine's MaaS business is becoming a foundational business for the company, with long-term and firm investment planned. To support high-frequency model calls and the implementation of Agents, Volcano Engine has upgraded its AI cloud-native full-stack services. The MaaS layer has released the Ark CLI; the Agent development and operation layer has upgraded the enterprise-level Agent infrastructure AgentKit and AI cloud infrastructure, launching ArkClaw Enterprise Edition and HiAgent 3.0; and the security aspect has strengthened the AI Trust security system. In addition to the highlighted Doubao large model 2.1 Pro, Seedance 2.5 video generation model, and the native 4K version of Seedance 2.0, the conference also unveiled two new models for image and audio. The upcoming Doubao image creation model Seedream 5.0 Pro has upgraded features for interactive precise editing, multi-layer separation, high-density information expression, and native multilingual generation; the officially released Doubao audio generation model 1.0 (Seed-Audio 1.0) supports simultaneous arrangement of dialogue, emotions, dialects, background music, and sound effects from a single prompt, producing film-quality sound effects.