AI investors say cheaper open-weight models can handle most routine tasks

Aaron Jacobson, a partner at New Enterprise Associates, explained that the AI industry is moving away from the assumption that bigger, more powerful models are always required. He said open-weight models are improving rapidly, offering lower costs and easier customization, which makes them suitable for a wide range of standard tasks. Frontier models from companies like OpenAI and Anthropic remain valuable for highly complex or long-horizon problems, such as extensive planning or secure code generation.

However, many enterprises are hesitant to adopt Chinese-origin open-weight models due to data-security worries, leading them to either use less capable domestic models or pay for closed-source options. Recent pricing data shows Anthropic’s lower-cost Opus 5 gaining traction, underscoring the competitive pressure on premium AI services. Jacobson predicts the next decade will focus on efficient model use rather than continual scaling.