Alibaba said memory cut its agents' token use up to 67 percent.
Alibaba Cloud made the claim on 22 September 2026 at its Apsara Conference, with TelecomTV reporting it the same day.
The agents are AI assistants that carry out tasks on their own.
Tokens are the units of text a model reads and writes, and model providers commonly charge by the token. Alibaba said the reduction applies to knowledge-heavy customer service and coding work. Data analysis is also covered.
The saving comes from reusing background material instead of giving an assistant the same information for every task.
The service behind it, Agent Context, gives assistants access to a company's documents and business systems. Chat records feed into the same shared layer. Agents can then pull current company information and retain context from earlier work.
Alibaba also announced AgentCore, which it said provides tools to deploy and run agents, with monitoring included.
The 67 percent is Alibaba's own figure. The company cited no independent testing.
Token count is only part of the bill. What does one finished task cost you end to end?
Image: AI-generated.
Sources
Our file on Alibaba
All 3 Alibaba storiesFor your next steering committee
- 1Ask which AI tools already touch your ERP or finance data, including the ones nobody approved.
- 2Ask what each tool is allowed to change, and who reviews what it changed.
- 3Ask how the result is measured, and who decides to switch a tool off.
Issue 01 of the ERP LEADERS brief puts a timeline, the three-number review and six questions on one page. Read issue 01
Welcome back. · Issue 01
