The prevailing consensus in tech media is that the AI industry is a financial bubble waiting to burst. However, a new analysis published on November 8, 2026, argues that this view is fundamentally flawed. The author, Matteo Raso, contends that the massive capital expenditure seen in AI is driven by one-time infrastructure builds rather than unsustainable operational costs. For developers and infrastructure builders, understanding this distinction is critical for predicting the long-term viability of the tools we rely on daily.

Infrastructure Costs Are One-Time, Not Operational

The primary argument against the bubble theory is the nature of AI spending. Companies have committed approximately $3 trillion to data center rollout, with estimates suggesting $1.9 trillion has already been spent. While $5 trillion in total expenses sounds terrifying, these are capital expenditures for hardware, not recurring operational costs. Once the data centers are built, the spending requirement plummets. Crucially, Anthropic’s S-1 filing reveals that the company generates $2.08 in revenue for every $1 of compute cost. This unit economics suggests that once the initial infrastructure capex is absorbed, AI services are highly profitable.

Debt Service Is Manageable for Big Tech

Skeptics often point to the massive debt taken on to fund these data centers. If the industry had to repay $5 trillion over 20 years, that would require $250 billion in annual payments. While this is a staggering sum, it is not impossible for major players. For context, Alphabet reported nearly $162 billion in revenue in 2019 alone, which adjusts to over $213 billion today. Furthermore, much of this debt is structured as convertible notes, which can be settled in equity rather than cash. Anthropic, for example, recorded a $72 billion loss in the first half of 2026 through share issuance to lenders, leveraging its estimated $1 trillion valuation to cover obligations without draining cash reserves.

Revenue Growth Outpaces Skepticism

Anthropic’s financial trajectory provides concrete evidence of scaling demand. Revenue grew from $41 million in 2023 to $386 million in 2024, then to $4.585 billion in 2025. In the first six months of 2026, the company reported $16.3 billion in revenue. This exponential growth suggests that unmet demand for AI automation is vast, particularly in programming and mathematical fields where AI swarms can outperform human labor. The argument that AI is just 'slop' that no one will pay for contradicts the reality of institutional lock-in and the sheer volume of token consumption by enterprise clients who are integrating these tools into their core workflows.

Public Metrics Don't Look Like a Bubble

Traditional bubbles, like the dot-com era, are characterized by inflated P/E ratios and companies with no path to profitability. The current AI leaders do not fit this profile. Alphabet, Meta, and NVIDIA have seen their market capitalizations explode, yet their P/E ratios remain stable in the 20s. This indicates that their valuations are backed by current earnings, not just speculative future growth. Alphabet’s breakdown shows that while Google Cloud revenue is growing by 36%, Google Services revenue is growing by 12% in percentage terms but by a much larger absolute amount ($37.8 billion vs. $15.5 billion). This proves that the core business is monetizing AI effectively, independent of data center rental margins.

Key Takeaways

  • AI infrastructure spending is largely one-time capex, not recurring opex, with unit economics showing profitability once hardware is paid off.
  • Anthropic’s revenue hit $16.3 billion in H1 2026, demonstrating massive scaling potential beyond initial hype.
  • Major AI players like Alphabet and NVIDIA maintain healthy P/E ratios in the 20s, unlike the inflated valuations of typical bubbles.
  • Debt obligations are being managed through convertible notes and equity issuance, reducing immediate cash flow pressure.

The Bottom Line

While the AI industry faces genuine risks from regulatory barriers and complex debt structures, the current data does not support the 'bubble' narrative. Investors are betting on real revenue growth and stable valuations, not just speculative hype.