Businesses Losing Faith in AI Systems

The artificial intelligence sector is approaching a critical juncture. After years of soaring expectations, investors, executives, and corporate boards are growing increasingly skeptical that the technology will deliver significant productivity gains in the near term. This shift in sentiment is driven by a widening gap between the massive capital being deployed and the lack of measurable returns.

According to the Silicon Data Token Expenditure Index, spending on large language models has doubled since 2025. While the cost per token has fallen by more than 90% over the past three years, usage has grown so rapidly that overall expenses continue to climb. The Federal Reserve Bank of Atlanta estimates that firms will increase AI spending by 50% in 2026, reaching a total of $280 billion. This surge has exposed a fundamental asymmetry where companies face mounting pressure to justify costs while providers rely on consumption to sustain their valuations.

High Failure Rates and Data Risks

A 2025 MIT report found that 95% of generative AI pilot programmes have failed to achieve their stated objectives. Many of these projects were intended to accelerate revenue growth but have not delivered the promised efficiency. Beyond the financial disappointment, there are serious concerns regarding intellectual property. Companies deploying AI tools risk handing over proprietary data and business know-how, accumulated over decades, to major model providers like OpenAI and Anthropic.

For business executives, this dynamic creates a difficult dilemma where the drive for operational efficiency must be weighed against the potential loss of trade secrets. The tools designed to streamline operations might ultimately empower competitors to replicate proprietary processes that took years to refine.

The dispute between Anthropic and Figma illustrates this tension clearly. Figma partnered with Anthropic to develop AI-powered design assistants, but the relationship strained when reports emerged that Anthropic was developing a competing tool, which it subsequently launched as Claude Design. History is replete with examples of technology companies leveraging their scale to enter adjacent markets, most notably Microsoft’s displacement of Lotus 1-2-3 and WordPerfect in the 1990s. Companies are advised to keep this history in mind as they deepen their reliance on AI.

Strategic Shifts in Spending

Palantir CEO Alex Karp recently captured the wariness of the industry, accusing OpenAI and Anthropic of “stealing the weights and alpha of my business” through token-based pricing. In response to these risks, business leaders should consider four steps. First, companies should review and restructure AI contracts so that model providers have more skin in the game. If LLM providers are confident they can deliver billions of dollars in productivity gains and cost savings, they should be willing to tie at least part of their compensation to the value they create.