Is the AI-Driven Tech Sector Inflating Earnings Through Accounting Manipulation?

The AI-driven tech sector has become a dominant force in global markets, with valuations soaring as companies pour billions into AI infrastructure, data centers, and R&D. However, beneath the surface of this innovation lies a growing concern: Are these companies inflating earnings through accounting practices that obscure the true economics of their AI investments? For contrarian value investors, this question is not just academic-it's a critical lens through which to assess the sustainability of current valuations.

The AI Boom and the Earnings Gap

According to a report by the Stanford HAI 2025 AI Index, global private AI investment hit $252.3 billion in 2024, a staggering figure that underscores the sector's allure. Yet, as noted by the McKinsey Global Survey, most organizations remain in experimental or pilot phases of AI adoption, with only 39% reporting any enterprise-level EBIT impact according to the survey. This disconnect between investment and earnings raises red flags. A 2025 analysis by highlights that while AI infrastructure spending contributed 0.5 percentage points to U.S. real GDP growth, many analysts question whether revenue growth justifies the scale of capital outlays.

Accounting Practices Under Scrutiny

One of the most contentious practices involves depreciation schedules for AI hardware. Tech investor Michael Burry has accused hyperscalers like Oracle and Meta of artificially extending the estimated useful lives of their AI computing equipment-despite the fact that AI chips typically become obsolete within 2-3 years according to Burry. By spreading depreciation expenses over longer periods, these companies reduce annual depreciation charges, inflating reported profits. Burry estimates this manipulation could overstate Oracle's profits by 26% and Meta's by 21% by 2028 according to Burry. Such tactics obscure the real costs of maintaining cutting-edge AI infrastructure, creating a misalignment between reported earnings and actual asset performance.

Another area of concern is revenue recognition in interdependent contracts. For instance, NVIDIA and Microsoft have entered multi-year agreements with AI startups and data center operators, creating circular financing loops that boost valuations and revenue recognition. These relationships can inflate demand signals while masking underlying vulnerabilities, such as the risk of overbuilding data centers without commensurate revenue streams.

Case Studies of Misrepresentation

The risks of accounting manipulation are not theoretical. In 2024, Kubient's former CEO was charged with inflating the company's revenues and misrepresenting the capabilities of its AI-powered fraud detection tool. Similarly, the FTC sued Air AI in 2025 for deceptive claims about business growth, earnings potential, and refund guarantees according to the FTC. These cases illustrate a pattern of exaggerated AI-related claims that distort financial signals for investors.

Even larger players are not immune. A 2025 case study by The Dando Diary revealed how a tech executive manipulated revenue recognition by overstating the value of AI-driven services according to the case study. Such practices highlight the sector's susceptibility to creative accounting, particularly in a market where hype often outpaces tangible results.

Contrarian Value Investing: A Path Forward

For value investors, the key lies in scrutinizing non-GAAP metrics and revenue recognition practices. While AI has undeniably improved efficiency in areas like invoice processing and fraud detection, these benefits are often overstated in earnings reports. For example, AI-powered accounting systems reduced manual data entry by 75% in some firms, but such operational gains do not necessarily translate to sustainable revenue growth.

Moreover, economic projections for AI's long-term impact-such as a 1.5% GDP boost by 2035-come with significant caveats according to a 2025 model. Current data on AI's productivity gains is limited, and future outcomes depend on factors like regulatory shifts and technological breakthroughs. This uncertainty underscores the need for a contrarian approach: focusing on companies with transparent accounting, verifiable revenue streams, and measurable productivity improvements rather than those relying on speculative AI narratives.

Conclusion

The AI-driven tech sector is at a crossroads. While innovation is real, the risk of earnings inflation through accounting manipulation is equally tangible. For contrarian investors, the path forward requires skepticism, rigorous due diligence, and a focus on fundamentals. As the sector matures, those who avoid the hype and prioritize transparency will likely outperform in the long run.