AI Stocks Face a Bubble – Why Doubt Persists

The Illusion of AI's Dominance

The concept of "getting big or getting lost" has long been a guiding principle in Silicon Valley, where the belief is that only the largest companies can survive and thrive. This mindset often leads to aggressive growth strategies, sometimes at the expense of profitability. While this approach has worked for some, like Amazon, it has also led to significant losses, as seen during the dot-com bubble. Just like professional wrestling, where the spectacle often overshadows the reality, the current AI landscape is filled with grandeur and hype that may not match the actual outcomes.

Professional wrestling, much like the AI industry, is characterized by exaggerated performances and larger-than-life figures. One such figure is Paul Wright, known as The Big Show, who embodied the idea of going big. His iconic slogan, “Go big or get lost,” resonates with the current AI scene, where companies are pushing boundaries to gain market share, often without considering the financial implications.

The Reality Behind the Hype

Despite the excitement surrounding artificial intelligence, the revenue streams for major tech companies like Alphabet, Microsoft, Meta Platforms, and even Amazon are largely driven by their core businesses rather than AI. Their cloud services benefit from technologies that are not solely AI-based. The U.S. economy is being fueled more by investments in cloud infrastructure than by spending on cloud services themselves. This trend has surpassed the spending seen during the dot-com boom, indicating a growing reliance on cloud infrastructure.

Expectations of future AI profits have led to inflated valuations for AI startups, with companies like OpenAI, xAI, and Anthropic reaching staggering figures. However, these high valuations come with significant risks, as many of these companies are operating at a loss. For instance, OpenAI reported a $5 billion loss on $3.5 billion in revenue in 2024, while Anthropic faced a $5.6 billion loss on $1.1 billion in revenue. These figures are expected to increase, highlighting the challenges faced by AI companies.

The Struggle for Profitability

The get-big-or-get-lost mentality has resulted in companies struggling to cover their costs, particularly when it comes to cloud service providers. As cloud-service providers become more profitable, they invest heavily in infrastructure, driving up demand for data centers, chips, and electricity. This cycle creates a frenzy in the stock market, but it also raises concerns about sustainability.

AI software companies are facing similar challenges. They are operating with low prices, small revenues, and significant losses. The expectation is that they will eventually raise prices, which could lead to customer dissatisfaction. Anthropic, for example, has started to implement new pricing structures that limit continuous usage, potentially affecting users who rely on free services for academic and personal purposes.

Sam Altman, CEO of OpenAI, has hinted at raising prices soon, which could lead to a backlash from users accustomed to freemium models. As one article noted, the AI industry faces a critical issue: how to generate revenue when training and running AI models is significantly more expensive than search queries. Users may be in for a shock as companies start to increase prices to cover their costs.

The Future of AI and Its Implications

Investors may not fully grasp the unsustainable nature of the AI cloud and processor revenues for big tech companies. As AI prices rise, demand for AI services and related infrastructure is likely to decline. This could have a ripple effect on data centers, chips, and electricity providers, who are cautious about investing without assurances from tech companies.

Electricity providers, in particular, are aware of the looming problem and are seeking financial commitments from tech companies before building new generating plants. The extent of the demand drop will depend on how high prices go and whether AI companies can deliver reliable services that justify the increased costs. Given the current trajectory, optimism is in short supply.

In 2000, during the dot-com bubble, internet subscriptions, e-commerce, and PC sales generated about $1.5 trillion in revenue (in 2024 dollars), which is vastly higher than the projected $20 to $30 billion for AI software this year. It’s challenging to envision AI users paying more for services that may not be reliable, especially when the cost of mistakes is substantial.

As the AI industry navigates these challenges, it’s crucial to recognize the disparity between the hype and the reality. The lessons from past bubbles serve as a reminder that sustainable growth requires more than just ambitious goals; it demands a realistic approach to profitability and value creation.