Students Showcase AI Prototypes on ROG Strix G16 at RTX AI PC Day

At the ITB campus event, NVIDIA collaborated with ASUS to distribute ASUS ROG Strix G16 laptops to student participants in the GeForce Campus Leaders program. Students utilized machines powered by GeForce RTX 5070 Laptop GPUs to run experimental projects spanning computer vision, logistics forecasting, and automated recycling sorting. Lenny Lin, Country Manager ASUS Indonesia, noted that the cooling systems inside the ROG hardware—featuring vapor chambers and triple fans—were engineered specifically to handle heavy AI workloads without thermal throttling. Edward Ferdian from the Pusat AI ITB added that the partnership aims to position students as active builders and innovators rather than passive software consumers.

Engineering students at ITB demonstrated three distinct prototypes developed entirely on the ROG Strix G16 hardware configuration. Izhar Alif Akbar from the Sekolah Teknik Elektro dan Informatika (STEI) ITB built REBA.ai, a computer vision system that analyzes factory CCTV footage to evaluate worker posture against Rapid Entire Body Assessment standards, leaving final intervention decisions to human supervisors. Tsabita Khoirunnisa developed ColdTrack AI to monitor refrigerated transport vehicles by predicting cargo temperature fluctuations and calculating remaining time before spoilage risks escalate. Meanwhile, Muhammad Taufiq Al Fikri designed RecycledFiberAI to detect metal and plastic contaminants in paper recycling streams in real time. NVIDIA noted that running these local training models on laptop GPUs significantly cut down experiment cycles compared to traditional workflows.

Adrian Lesmono, Country Consumer Business Lead NVIDIA Indonesia, outlined the strategic differences between running artificial intelligence workloads locally versus utilizing cloud infrastructure. Local processing on GeForce RTX 50 Series laptops—which feature Blackwell architecture and fifth-generation Tensor Cores—keeps sensitive data directly on the device, avoids recurring subscription fees, and operates independently of internet connectivity. However, Adrian emphasized that local output remains bound by the physical limits of the hardware and the specific model size loaded onto it. Conversely, cloud-based AI offers virtually unlimited capacity for massive foundational models, though intensive daily use drives up operational expenses.