Morgan Stanley Report: Open Weight Models Reduce AI Costs, Computing Demand May Continue to Grow

Morgan Stanley's report titled "Open Weight Models and Three Future Scenarios" indicates that open weight models may not weaken the demand for AI computing power. On the contrary, the reduced usage costs could accelerate the adoption of AI, creating a "Jevons Paradox": as the cost of inference decreases, companies will apply AI to more tasks, ultimately increasing the total demand for tokens, computing power, electricity, and infrastructure. The report emphasizes that open weights do not mean completely free; companies still need to bear costs related to GPUs, cloud services, operations, and security, with actual economic viability depending on the application scenario. Morgan Stanley believes that regardless of how the openness of models changes, companies like NVIDIA are likely to benefit.