DeepSeek, the Chinese AI disruptor, released striking adoption numbers for its AI agent infrastructure as it continues to challenge Anthropic and OpenAI with an entire swath of applications built around open-source AI technologies.
On Sunday, the Hangzhou-headquartered startup disclosed on X that DeepSeek Harness, its open-source agent application, had reached 1.54 million daily active users by Nov 4, less than three months after its Aug 13 debut.
Daily users were at about 271,000 at the end of September, right after it released a desktop version of the app, based on the number of users connecting to its own servers.
DeepSeek, which jolted Silicon Valley with its low-cost reasoning model R1 last year, is staking claim to the next frontier of the industry: infrastructure for AI agents, software that lets AI act autonomously on a user’s behalf.
DeepSeek is planning an official marketplace where developers can publish plugins and users can discover them. It’s also planning a support programme for plugin creators including token grants, technical assistance and promotional opportunities. No further details have been disclosed.
The new moves reflect a broader shift in the AI industry where the battle for users is increasingly being fought over ecosystems and integration rather than model performance. DeepSeek is positioning Harness as an open, developer-driven agent ecosystem to challenge OpenAI and Anthropic not just on model performance and cost, but on the breadth of applications built around its AI technology.
The AI disruptor is betting that third-party applications and specialised tools will build traction around its open-source AI agent infrastructure and become a decisive source of competitive advantage against Silicon Valley’s proprietary technologies, as the artificial intelligence industry moves beyond chatbots toward autonomous agents.
AI agents carry out multi-step tasks on behalf of the user such as writing and modifying code, analysing files, conducting research and using other tools with minimal human supervision.
Harness’ appetite for computational power is notable. The application is consuming approximately 92 trillion tokens daily, which works out to about 60 million tokens daily for each active user – providing among the first available evidence of AI agents steeply escalating the demand for processing power.
The scale shows that the hunger for compute power in AI agents is vastly higher than that of chatbots, as AI agents run for multiple hours as they read or write code, use other tools and go on the web, all with little human intervention.
The startup is building Harness as a “everything is a plugin” architecture. A plugin adds a special feature to a piece of software without changing it, giving it added functionality. DeepSeek counts about 35,000 plugins and 60% of users have built at least one of them, though the startup said that the sophistication of the plugins varies.
Leave a Reply