--- title: "Current Status of AI Applications: High Penetration Rate, but Extremely Low Monetization Rate" type: "News" locale: "zh-CN" url: "https://longbridge.com/zh-CN/news/270782275.md" description: "McKinsey's research on AI applications shows that despite an AI adoption rate of 88%, most companies have not achieved large-scale deployment, with only 39% of respondents believing that AI has a substantial impact on corporate profits. The real winners are those companies that reshape workflows and invest more than 20% of their digital budget. Investors should focus on companies that dare to reconstruct business processes, rather than those that remain at the pilot stage" datetime: "2025-12-25T09:12:00.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/270782275.md) - [en](https://longbridge.com/en/news/270782275.md) - [zh-HK](https://longbridge.com/zh-HK/news/270782275.md) --- > 支持的语言: [English](https://longbridge.com/en/news/270782275.md) | [繁體中文](https://longbridge.com/zh-HK/news/270782275.md) # Current Status of AI Applications: High Penetration Rate, but Extremely Low Monetization Rate Three years after the advent of generative AI, a significant gap has emerged between market hype and the actual implementation by companies. According to the Chasing Wind Trading Desk, in December, McKinsey's AI application research report pointed out that **while everyone is talking about AI, almost no one is really making big money from it.** Data shows that nearly 90% of companies have normalized the use of AI, but the vast majority remain trapped in "pilot purgatory" and have not achieved large-scale deployment. More critically, only 39% of respondents believe that AI has had a substantial impact on their company's earnings before interest, taxes, depreciation, and amortization (EBITDA), with most contributions being less than 5%. McKinsey believes that one should not be misled by the "AI content" in corporate earnings calls. Current AI applications exhibit a typical characteristic of being "widespread but superficial." Capital expenditures are occurring, but returns are severely lagging. The real winners are not just those buying software, but those reshaping workflows and pursuing growth, rather than merely cutting costs through layoffs. When evaluating targets, investors should be wary of companies that are still in the "pilot" stage and look for those that are truly willing to reconstruct business processes and invest more than 20% of their digital budget—these are the "true players." The penetration rate has not translated into scale: Most companies are still in the "trial" phase. Although the reach of AI seems impressive, the depth is severely lacking. McKinsey's research shows that 88% of companies report normalizing the use of AI in at least one function, up from 78% last year. However, this is merely surface prosperity. - Scaling dilemma: Nearly two-thirds of respondents admit that their companies have not fully initiated large-scale AI deployment. Most organizations remain in exploratory or pilot phases. - Size matters: The larger the company, the closer it is to scaling. Among large enterprises with revenues exceeding $5 billion, nearly half have entered the scaling phase; whereas for small enterprises with revenues below $100 million, this proportion is only 29%. - Outstanding performance in mainland China: In mainland China, 45% of companies have achieved large-scale or comprehensive deployment of AI, higher than the global average of 38%, and 83% of companies have normalized the application of generative AI, demonstrating stronger execution capabilities. The hype around "intelligent agents" versus reality: still in the early experimental stage. The market has high hopes for "intelligent agents," but actual applications are still in the very early stages. - High trial, low implementation: 62% of respondents indicate they are trialing intelligent agents, but only 23% of companies have initiated large-scale applications in at least one function. - Functional limitations: Even among companies advancing towards scaling, applications are limited to one or two areas. In any single function, the proportion that has moved from pilot to scale does not exceed 10%. - Industry differentiation: The technology, media, and telecommunications, as well as healthcare industries, are the leaders in the application of intelligent agents. IT service desk management and knowledge management are the most mature application scenarios. The truth about profits: perception outweighs reality, and monetization capabilities are in doubt This is the part of the report that most alerts investors. Although 64% of respondents believe that AI drives innovation, the impact of AI on financial statements' bottom line is minimal. - Weak EBIT impact: Only 39% of respondents felt a substantial impact at the EBIT level of their companies. This means that over 60% of companies are currently in a phase of "losing money to gain attention" or are merely in the investment stage. - Limitations of cost reduction and efficiency improvement: Although costs in software engineering, manufacturing, and IT have decreased, and marketing and sales have brought revenue growth, these localized improvements have not yet converged into a company-wide profit explosion. 6% of winners take all: What did high-performing companies do right? The survey defined about 6% of companies as "AI high-performing companies," meaning their EBIT increased by over 5% due to AI and created significant value. These companies are fundamentally different from the other 94%: - Refusal to simply cut costs: 80% of ordinary companies focus solely on efficiency and cost reduction, while high-performing companies often pursue growth or innovation goals simultaneously. - Restructuring workflows: High-performing companies are nearly 3 times more likely to fundamentally reshape workflows than other companies. This is not just about installing a plugin, but changing the way work is done. - Real monetary investment: One-third of high-performing companies allocate over 20% of their digital budget to AI, a ratio nearly 5 times that of other companies. - Executive involvement: Management's involvement is not just verbal support but establishing clear human-machine collaboration processes and manual verification mechanisms. Employment and risks: Layoff expectations rise, with hallucination risk at the forefront As AI deployment deepens, the labor market and corporate risk control face dual pressures. - Diverging layoff expectations: Although the employee count has remained stable over the past year, expectations for the next year have become pessimistic. 32% of respondents expect a decrease in employee count in the coming year, while only 13% expect growth. Large companies are more inclined to believe that AI will lead to workforce reductions. - Structural talent shortages: On the other hand, the demand for software engineers and data engineers remains strong. - The primary risk is "nonsense": 51% of application companies have encountered at least one negative event related to AI. The most common risk is "inaccurate results," also known as the hallucination problem. While companies are beginning to focus on privacy and compliance, attention to "explainability" remains insufficient, posing a potential tail risk. For investors, the AI narrative in 2025 must shift from "who is using it" to "who is making money." Current data indicates that, aside from a few top players willing to invest heavily and restructure business processes, most companies are still in the "growing pains" phase of AI investment and have not yet seen a clear profit turning point ## 相关资讯与研究 - [MedPal AI Wins Strong Shareholder Backing at AGM as It Expands AI Health Platform](https://longbridge.com/zh-CN/news/281501359.md) - [Letter from the Editor Introducing AI Intelligence on American Banker](https://longbridge.com/zh-CN/news/281266312.md) - [Folk are getting dangerously attached to AI that always tells them they're right](https://longbridge.com/zh-CN/news/280988775.md) - [Insig AI Plans Growth Drive and Eyes Nasdaq Dual Listing](https://longbridge.com/zh-CN/news/281311983.md) - [17:23 ETRafay Systems Transforms GPU Providers Into AI Factories By Empowering Them to Monetize Token-Metered Access to AI Models](https://longbridge.com/zh-CN/news/281577708.md)