Agent Bloodbath 2026: Big Tech Killed 10+ AI Brands in 6 Months — The $65B Revenue Wake-Up Call
In August 2026, Alibaba, ByteDance and Tencent consolidated agent products, killing 10+ AI brands overnight. Anthropic's $65B ARR proved agents can print money — and the $10.9B global market is entering a brutal second half.
The Bloodiest Product Purge in AI History
In August 2026, the AI industry witnessed something unprecedented—not a wave of product launches, but a massacre of brands. Alibaba merged QoderWork, Wukong, and MuleRun into "QwenWork," launching public beta on August 3. ByteDance folded Coze and TRAE into "Doubao Work," releasing it on August 25. Tencent consolidated multiple internal projects into WorkBuddy. Within six months, more than ten once-beloved AI products were killed by their own parents: Coze, TRAE, Feishu Aily, Wukong, MuleRun, QoderWork. They were not outcompeted by rivals—they were strategically sacrificed by the very companies that created them.
Why the Rush? A $65 Billion Revenue Alarm
Anthropic's annualized revenue hit $65 billion by July 2026—more than 7x growth since late 2025. When one company proves that agents can print money, every tech giant panics. Yet ByteDance's Doubao boasts 200 million+ DAU and 380 million MAU, while daily revenue remains under one million yuan. A free chatbot cannot carry the weight of enterprise agent monetization. Stuffing high-barrier office capabilities into a consumer product built on free chat creates the wrong user expectation: users do not see "great feature," they see "now you want my money too."
The Numbers Behind the War
The global AI agent market reached $10.91 billion in 2026 and is projected to hit $183 billion by 2033 at a 49.6% CAGR, according to Grand View Research. China's enterprise agent market alone stands at 44.9 billion yuan, nearly doubling year-over-year. But here is the gap that matters: while 79% of enterprises have adopted agents in some form, only 11% run them in production. That 68-percentage-point chasm is the battlefield of the second half of 2026. IDC forecasts agentic AI will absorb 26% of worldwide IT spending by 2029, reaching $1.3 trillion. The pie is real. The question is who can actually serve it.
Who Will Build the WeChat of Agents?
Tencent's WorkBuddy leverages ecosystem integration with Tencent Docs and enterprise WeChat. Alibaba's QwenWork rides on the Qwen model's technical strength. ByteDance's Doubao Work brings unmatched monthly active users and deep Feishu integration. But the real question is not which product wins today—it is which platform can embed agents into existing enterprise workflows without forcing users to migrate their entire operating system. OpenAI's Codex has surpassed 5 million weekly active users. Claude Code now commands a 1-million-token context window, letting a single agent ingest entire codebases. And in the most dramatic move of all, SpaceX acquired Cursor for $60 billion, folding its capabilities into Grok Bot. Models are just brains; the harness—tool access, planning logic, memory, retry mechanisms—determines who wins.
Signals You Cannot Ignore
GitHub Copilot has evolved into a multi-agent development platform. Baidu's Dazi agent hit 6.74 million MAU in July, surging 1,063% month-over-month. Gartner predicts 40% of enterprise applications will embed task-specific AI agents by end of 2026, up from under 5% just twelve months earlier. The competition has shifted from "which model talks better" to "which agent actually gets the job done." The second half of 2026 is where the real AI war begins.