ASUS and Poesis Test Autonomous AI Agents in Cryptocurrency Trading

Taiwanese hardware maker ASUS has partnered with crypto firm Poesis to test autonomous artificial intelligence agents that can conduct market research, manage risk, and execute trades. The week-long experiment used an ASUS ExpertCenter Pro ET900N G3 laptop equipped with NVIDIA technology. This trial marks a significant shift in the development of AI trading bots, as these agents no longer merely offer advice but operate independently, collecting data, developing strategies, and trading continuously.

The always-on nature of cryptocurrency markets and their programmable infrastructure make them an ideal environment for this shift. ASUS is making a case for running such agents locally rather than in the cloud, citing 'The Token Trap', a scenario where continuous agent activity leads to unpredictable API bills and usage caps imposed by cloud providers.

The cost differential between local deployment and cloud-based workloads was highlighted by ASUS, with an upfront investment of roughly $4,000 plus annual electricity costs of about $384, compared to $6,000 to $8,000 in annual API fees for a comparable cloud-based workload. The experiment reflects the broader evolution of financial AI, which is moving from rule-based algorithmic trading to machine learning prediction and now to agentic finance.