TMTPost publishes opinion article on AI industry

TMTPost published an opinion article on the AI industry on June 23, 2026.

Publication of an opinion article does not mean the claims it discusses about technology or markets were independently verified. When reading about robot operating metrics or descriptions of simulations, proposed ideas and explanations should be distinguished from measurement results with clearly stated conditions.

The publication of an opinion article is not primary evidence supporting the performance of the technology or industry trends it discusses. Regardless of whether readers accept the article’s views, it is important to check who measured what for each technology and under what conditions the results were presented. See the original source for details. The same distinction applies to articles we have published about robots’ real-world operations. AgiBot proposes “Worken,” a metric for robot utilization, leaderobot.com reports introduced a metric combining the number of operating robots, autonomous work hours and an output efficiency ratio, while noting that its definition, measurement method and application to pricing could not be confirmed. Even when a metric’s name or formula is provided, it cannot immediately serve as a basis for comparing the operations of different robots or sites unless the measurement targets and calculation conditions are defined. The same distinction should be made between opinion and technical explanations concerning computing infrastructure. leaderobot.com publishes opinion article on computing resources for embodied AI presents views; that does not show they were independently verified. By contrast, NVIDIA explains world models for interactive simulation using Cosmos identified adapting models to sensors, tasks and data as a challenge, according to our report. Discussions of computing resources, explanations of simulations and records of operations in real-world environments should be treated as separate forms of evidence. The next questions are whether measurement data and conditions supporting each claim have been made public, and whether models or metrics are evaluated by the same standards across different environments. The publication of this article alone does not allow conclusions about the performance of specific models, deployment records, computing demand or business results. Clearly distinguishing reported views from reproducible test results and operational data gives readers a basis for making their own assessments.