The Case Against Government Buying Into the AI Giants

A Unlikely Political Consensus Faces Economic Reality

An intriguing convergence of opinion has emerged between figures as disparate as Senator Bernie Sanders and former President Donald Trump. Both have publicly floated the idea of the federal government acquiring equity in leading artificial intelligence companies. For Sanders, this proposal is framed as a mechanism for the American populace to reap benefits from AI advancements, which he argues were built upon the uncompensated utilization of public creative works. This approach, he suggests, would also grant the government a voice in shaping the future trajectory of AI development, representing public interests.

The precise motivations behind Trump's endorsement of government AI investment remain less defined, leaving room for speculation. However, a fundamental challenge underpins this political sentiment: the current economic reality of these AI firms. Contrary to the narrative of burgeoning wealth generation, these companies are presently consuming vast sums of capital from investors. This influx of funds is heavily subsidizing user access, with a clear and profitable business model still conspicuously absent.

The assertion that AI follows the historical trajectory of transformative technologies like the internet or smartphones, where adoption leads to cost reductions and economies of scale, appears increasingly questionable. In the years since the AI boom began, the operational expenses associated with delivering AI's promised capabilities have, in fact, been escalating. This counterintuitive trend is driven by the very nature of resource-intensive AI models; each additional user necessitates greater computational power and infrastructure, thereby increasing, not decreasing, per-unit costs. This stands in stark contrast to traditional software, where the marginal cost of distributing an already developed product approaches zero.

The Specter of AI Hallucinations and Financial Unsustainability

Concerns about the fundamental viability of current AI models are mounting. Last November, a warning was issued regarding the potential for AI companies to necessitate a government bailout, a scenario stemming from inherent flaws in their models that may prevent them from delivering on their ambitious promises. Industry commentators highlight a significant gap between what AI is advertised to do and what it can demonstrably achieve in the present. When pressed to describe current, tangible AI capabilities without resorting to future projections, proponents often struggle to provide concrete examples.

Even in areas where AI is being integrated, such as search engines, the reliability of outputs is reportedly declining. Users are compelled to meticulously verify information, an indicator of diminishing trust in AI-generated content. This skepticism was recently amplified by an incident involving a report from KPMG, a major accounting firm. The report, which was intended to showcase AI's utility in corporate and governmental applications, contained significant factual inaccuracies, attributed to AI "hallucinations"-instances where the AI fabricates information.

The report's subsequent withdrawal, particularly given its purpose was to promote KPMG's AI integration services, underscores the precariousness of relying on current AI technology. Adding to the controversy, a powerful business lobby in Ohio is reportedly seeking legislative changes to eminent domain laws. Their aim is to allow AI companies to acquire land for data center energy projects without upfront payment, a move sparking fierce opposition from agricultural groups concerned about the precedent of seizing private property without immediate compensation.

Why Government Investment Now Is a High-Stakes Gamble

The financial underpinnings of the AI industry present a stark picture, suggesting that government equity acquisition at this juncture could be a financially perilous endeavor. The question arises: why would the leader of a purportedly dominant AI company willingly dilute shareholder value by inviting government ownership, unless anticipating a future need for financial rescue?

Two primary factors appear to be driving the industry's push for government intervention. Firstly, the AI sector has successfully cultivated a narrative among policymakers that positions AI as the indispensable future of technology, thereby creating an implicit understanding that the industry itself, distinct from the technology, cannot afford to fail. Secondly, the substantial capital expenditures within the AI industry are currently a significant engine for economic activity and a key driver of the stock market. A downturn among major AI players could trigger widespread economic and portfolio distress, making a bailout an attractive, albeit potentially misguided, option for those in power.

Senator Sanders has proposed a 50 percent government stake funded through a stock tax, while the specifics of a Trump administration's approach remain unclear. Regardless of the mechanism, acquiring AI company shares now risks investing at the peak of a speculative bubble. Governmental bodies are historically ill-equipped for astute financial market timing, often prioritizing policy objectives over shrewd investment strategy. Should the current AI bubble deflate, the temptation for the government to prop up its investment with taxpayer funds, leveraging its unique powers of taxation and currency issuance, will be immense. Furthermore, these new government partners, the AI company executives, will exert considerable lobbying pressure to secure bailouts, even in the absence of a viable path to profitability.

Reading Between the Lines

The current framework of Large Language Models presents a fundamental dilemma: the industry struggles to articulate a clear strategy for simultaneously reducing operational costs and eradicating the "hallucinations" that render AI unreliable and potentially dangerous for autonomous applications. Projections suggest the industry needs to generate approximately $2 trillion in new revenue over the next four years to validate its current and future investments, according to commentary from AI critic Ed Zitron. For context, the entire global software industry generated roughly $719 billion in revenue in 2025, offering products with near-zero marginal distribution costs.

Data indicates a significant disconnect between AI project ambitions and realized business value. Reports suggest that 80 percent of AI projects fail to deliver tangible business benefits. Even among the largest publicly traded companies, a mere 21 percent of S&P 500 firms can identify a measurable AI advantage, despite global AI spending on capital expenditures alone projected to reach $527 billion in 2026. While current AI models might offer some utility, the economic case for widespread adoption hinges on drastically higher user fees, potentially exceeding $100 per month for premium services, far beyond what most individuals or businesses currently pay or perceive as valuable, especially given the persistent need for human oversight due to errors.

The path forward, therefore, appears to be one where the market dictates AI's true value. Allowing the marketplace to determine AI's success or failure, without government intervention, is the most prudent course. If a genuinely profitable AI model emerges, capable of delivering sustainable value to businesses and consumers, it should be allowed to flourish organically. Presently, the industry seems to be seeking a governmental backstop to cover the financial miscalculations of its leadership, effectively shielding personal wealth from the inevitable consequences of an inflated market bubble.