800 million weekly active users, token consumption has surged 320 times...OpenAI's latest report: Enterprise AI has entered the "deep water zone"

OpenAI recently released the "Current Status of Enterprise Artificial Intelligence (AI) Report", which shows that AI is fully penetrating various industries, with technology, healthcare, and manufacturing being the fastest growing sectors. The tech giant revealed that the analysis data for this report comes from two main sources: the actual usage data of OpenAI's enterprise clients, and a survey of AI application patterns conducted on 9,000 employees from nearly 100 companies. The report also revealed that the weekly active user base of ChatGPT has exceeded 800 million. OpenAI emphasized that the current application of AI by enterprises is showing a dual characteristic of accelerating penetration and deep integration. Over the past year, the weekly interaction volume of ChatGPT Enterprise Edition has grown by about 8 times, and the weekly command volume sent by individual users has increased by 30%. In terms of application patterns, the frequency of using structured workflows such as project management and customized GPT has increased 19 times this year, indicating a shift from scattered ad hoc queries to systematic and reusable business processes. The computational power consumption for model reasoning at the enterprise level has also significantly increased, with the average token consumption for inference in a single enterprise growing by about 320 times in the past 12 months. OpenAI interpreted this data as confirming that smarter large models are systematically being integrated into expanding product and service systems within enterprises. The report further points out that AI is being used to complete increasingly complex professional tasks. In terms of industry distribution, the application of AI covers all sectors, with professional services, finance, and technology industries showing particularly strong growth due to their advantage in large-scale implementation. From a global market perspective, the number of enterprise clients in Australia, Brazil, the Netherlands, and France has all increased by over 140% compared to the previous year, making them the fastest growing regional markets. In terms of technical interfaces, over the past six months, the number of international enterprise clients using OpenAI's application programming interfaces (APIs) has increased by over 70%, with Japan having the highest number of enterprise API clients outside the United States. Research data shows that 75% of employees in surveyed companies stated that their work output efficiency or quality significantly improved after using AI tools. In terms of time costs, average users save 40 to 60 minutes of work time per day, while high-frequency users save over 10 hours of work time per week. Notably, the amount of programming-related commands sent by non-technical employees has increased by 36%; at the same time, 75% of users have successfully unlocked tasks that they were unable to complete independently in the past with the help of AI. The report also highlights a key trend: the gap between advanced users, leading companies, and industry averages is continuing to widen. From a user perspective, the weekly command volume of top users (top 5%) is 6 times that of average users, and they more deeply utilize the advanced functions of the models; from an enterprise perspective, the average interaction volume per seat of leading companies is double that of the industry median, and their AI technology deployment has achieved deep cross-team collaboration. OpenAI pointed out that this gap is of significant practical importance - research shows that the more intelligent computing power users access, and the more diverse business scenarios they cover, the more significant the time-saving benefits. In a blog post released by OpenAI on Monday, it emphasized, "We introduce a new feature or capability every three days on average. Nowadays, the core bottleneck that hinders enterprises from embracing AI is no longer the performance of the model or the tools themselves, but the adaptability of the organizational structure and the implementation capability on the ground."