From the perspective of the capital market, AI is entering a critical transition period from "potential narrative" to "result verification". Deutsche Bank Research Institute pointed out that 2026 may be the most difficult year for the development of AI so far, with the triple pressures of disillusionment, dislocation and distrust emerging simultaneously. At the same time, the global CEO survey report released by PwC also shows that there is a huge gap between enterprises’ ambitions and reality in AI applications, and the lack of underlying logic in technology implementation is becoming a core constraint on industry development.

Deutsche Bank warns: AI faces triple pressures of disillusionment, dislocation and distrust

Wall Street is generally betting that 2026 will be the "moment of reckoning" for AI technology, and that the market will no longer be satisfied with conceptual hype, but will demand that this technology and related trading networks deliver tangible returns. Recently, the volatility of software stocks has intensified, and the US competition for Greenland has caused market turmoil. Many heavyweight AI concept stocks have suffered significant declines. The S&P 500 technology sector fell more than 2%, AI chip leader Nvidia fell nearly 4%, Google parent company Alphabet fell 2%, and Broadcom fell nearly 5%.

Technology implementation encounters "disillusionment period"

The benefits of generative AI are currently concentrated only in Silicon Valley and a few early adopters, and have not brought significant revenue growth to most enterprises. Companies are confronting the technology's inherent limitations as they move pilot projects into production. These limitations include insufficient accuracy, difficulty in applying in real-life scenarios, and higher cost than human labor. Cox said bluntly that generative AI will eventually bring about changes, but not now.

The imbalance between supply and demand exacerbates the "dislocation" dilemma

Bottlenecks such as energy grid constraints and talent shortages have caused the gap between demand and production capacity of AI technology to continue to widen. Private AI companies, represented by OpenAI, are under tremendous pressure as they step up their fundraising efforts to compete with large cloud computing giants. Apple's move to choose Google's Gemini model to support its AI capabilities has made OpenAI's situation even worse. Cox believes that this year is a critical year for the success of independent AI model manufacturers. OpenAI’s front line is too long and it has not yet found a viable business model, making it difficult to cover high cash consumption. In comparison, competitors such as Google rely on their own data centers and internal funds to launch comparable models, making OpenAI's moat appear relatively shallow.

Industry anxiety breeds "distrust" sentiment

The distrust surrounding AI continues to rise and is reflected in multiple aspects. These aspects include AI-driven job replacement, copyright and privacy litigation, the consumption of electricity and water resources by data center investment, geopolitical competition, etc. As AI becomes a tool for countries to pursue self-sufficiency, concerns about the global AI race are growing. Cox predicts that this year anxiety about AI will turn from a low hum to a deafening roar.

PwC survey: The gap between AI ambition and reality is rooted in implementation capabilities

PwC's 29th Global CEO Survey Report, released at the opening of the Davos Annual Meeting, further confirmed the plight of the AI ​​industry. The report is based on feedback from 4,454 CEOs in 95 countries and regions around the world, revealing the huge gap between enterprise AI application ambitions and reality. From a cognitive perspective, the business world has completed key changes between 2024 and 2025. The focus of enterprises has changed from "whether AI should be adopted" to "all employees are involved in AI layout". The value of AI has become an industry consensus. But judging from the actual results, this consensus has not translated into performance growth. Only 10% to 12% of companies said that AI applications have brought actual benefits in revenue growth or cost control, and as many as 56% of companies bluntly stated that "AI investment has no return."

The above conclusion coincides with previous research from MIT, which pointed out that 95% of generative AI pilot projects in the corporate world ended in failure. Kander, the relevant person in charge of PricewaterhouseCoopers, said that the root of this contradiction is not the technology itself, but that the company has ignored the underlying logic of implementation. AI technology develops too fast, causing companies to forget that technology implementation must return to the basics. These foundations include consolidating data governance, improving business processes, establishing compliance frameworks, etc. They are the prerequisites for AI to exert its value. The survey found that companies that have benefited from AI applications have, without exception, built a solid infrastructure. The core of the problem lies in execution ability, and the level of execution ultimately depends on the management level and leadership of the enterprise.

Under the changing circumstances: CEO confidence is low and corporate transformation is imminent

The environment full of uncertainty has also given rise to an emotional paradox in the business world. Although CEOs are confident about global economic trends, only 30% believe their companies can achieve growth. This proportion has dropped sharply from 38% in 2025 and 56% in 2022, setting a record for the lowest CEO confidence in the company's own revenue prospects in five years. Kander pointed out that the current changes are a real test for corporate executives, requiring them to break away from day-to-day tactical matters and achieve rapid change and flexible adaptation. It is worth noting that while short-term confidence is low, a large number of business leaders are still laying out multi-year growth opportunities. They are trying to promote corporate transformation and reshaping through AI, technological innovation and cross-industry expansion.

In addition, the development of AI is forcing companies to redesign career development paths. The traditional "apprenticeship model" is being subverted. In the future, the focus of the enterprise training system needs to shift from task execution capabilities to systematic thinking capabilities. Kander suggested that business executives look beyond the short-term perspective of the past five years and look at the present from the historical perspective of the past 50 to 100 years. He cited the railway era and the early infrastructure construction boom of the Internet as examples, and firmly believed that the current wave of AI investment will usher in a new era of innovation. The PwC report also defined the future as "a decade of innovation and industrial restructuring" and pointed out that companies that gain more revenue from emerging fields tend to have higher profit margins and CEOs are more confident in future growth. Kander concluded that humans tend to fear things they do not understand, and the best way to resolve fear is to proactively seek the truth. In the wave of AI changes, only by embracing the changes can we seize a new round of development opportunities.