Data as of 31 July 2026. Source: Artificial Analysis. The Artificial Analysis Intelligence Index (composite score from 0 to 100) measures and ranks the overall intelligence and capabilities of large language models. It combines a comprehensive suite of evaluation datasets to assess language model capabilities across reasoning, knowledge, maths and programming.
August 2026
Chinese AI labs have closed the gap with leading US developers far faster than many expected. Moonshot AI's Kimi K3 is among the first open-source models to approach the capabilities of the industry's leading systems, highlighting how quickly advanced AI capabilities are spreading. According to the Artificial Analysis Intelligence Index, it now ranks among the world's leading AI models.
Much of the discussion today focuses on whether the US or China will win the AI race. However, investors may be asking the wrong question. Rather than a winner-take-all outcome, a more likely scenario is one where the most advanced models retain an edge in complex, high-value applications such as scientific research and drug discovery, while increasingly capable open models support a growing range of commercial and everyday uses.
As advanced AI capabilities become more widely available and performance differences between models narrow, value creation may increasingly shift away from the AI models themselves. Instead, some of the greatest beneficiaries could be the companies providing the infrastructure needed to deploy AI at scale. History suggests that when technologies become more affordable, usage often accelerates. As AI costs decline, broader adoption could drive even greater demand for computing power.
This reinforces the investment case for the foundational enablers of the AI ecosystem: Graphics Processing Units (GPU), advanced semiconductors, memory, networking equipment, power infrastructure and data centres. The key insight is simple: if AI models become increasingly commoditised, the most durable sources of value may lie not in the models themselves, but in the infrastructure and platforms that power their widespread adoption.
Stay informed: Sign up to get our latest insights delivered to your inbox.