Morgan Stanley report reveals the top 10 investment truths about AI: Moore's Law is obsolete, token economy is emerging, and software moats are being reshaped

Glonghui July 6th｜Morgan Stanley Investment Management has released its Q2 in-depth research report titled "Artificial Intelligence: Ten Investment Truths." The report points out that AI is no longer just a technological race, but a systemic revolution that is rewriting the fundamental rules of global industries and organizational operations.1. Resonance of four forces ushers AI into a flywheel effect. The current leap in AI is not driven by a single model breakthrough, but by the simultaneous acceleration of four forces: algorithms, computing power, talent, and capital.2. Moore's Law no longer holds; the physical world is the new bottleneck. The previous belief that "computing power will become increasingly cheaper" is weakening, as the AI bottleneck continues to shift from chips to power supply, memory, networks, cooling equipment, and data center provisioning.3. The rise of the token economy: computing power is becoming revenue. Previously, data centers were cost centers, but they are now transforming into manufacturing facilities producing intelligent output, with tokens serving as the unit of measurement for this output.4. The era of agents is coming, shifting employees from execution to management. AI is evolving from answering questions to executing tasks, invoking tools, managing processes, and running autonomously.5. Software moats are being reshaped; data, expertise, and distribution are the keys. AI is quickly lowering the threshold for code and feature development, meaning software companies’ competitive advantages are shifting from product features to harder-to-replicate core assets, including enterprise data, industrial knowledge, client networks, compliance capability, and deeply embedded business process experience.6. AI is moving from the analysis economy to the operations economy. With the rapid advancement of embodied intelligence, AI has extended its reach from knowledge work to physical industries such as transport, logistics, manufacturing, healthcare, and energy.7. AI is a full-stack capital cycle spanning infrastructure, models, and application layers, not a single-industry story, covering an entire ecosystem including chips, data centers, model training, AI agents, robotics, and more.8. China and the US are forming parallel AI architectures, meaning future AI competition is not just about capital and model size, but will depend more on cost control, engineering capabilities, supply chain integration, application deployment speed, and organizational efficiency.9. AI is already strategic infrastructure, but governance seriously lags behind.10. Infrastructure comes first; the most rewarding applications may not have emerged yet. History shows that when infrastructure matures and costs fall rapidly, unforeseen new applications often follow. As token costs plummet, the real killer app that will consume massive AI computing resources may not have been born yet, and even the companies related to it may not exist today.