The Casino's House of Cards

Silicon Valley is building a $600 billion casino with chips that expire in three years, a bubble that may be more destructive than all previous tech bubbles combined.

TL;DR
• Silicon Valley is spending $600 billion on AI infrastructure, with a 6-7x capacity-to-revenue ratio, worse than the 1990s fiber optic bubble.
• The AI infrastructure boom is a bubble that may be more destructive than all previous tech bubbles combined.
• The technology is real, but the investment is insane, with a 13x cost structure and a 30-50% annual depreciation rate.

Summary
The technology industry is currently experiencing a significant bubble, with $600 billion being spent on AI infrastructure to generate $40 billion in revenue. This spending spree is unsustainable, with companies like Amazon, Microsoft, and Google increasing their capital expenditures to build GPU infrastructure that may become obsolete in just three years. The business models behind this spending are flawed, with companies like OpenAI losing $5 billion despite growing revenue to $3.7 billion. The industry's focus on short-term gains and excessive spending is reminiscent of previous tech bubbles, including the dot-com crash and the telecom meltdown.

The consequences of this bubble bursting could be severe, with potential losses for investors, including private credit funds and major lenders like Blackstone and BlackRock. The hyperscalers, such as Amazon, Microsoft, and Google, may survive due to their diversified revenue streams and strong balance sheets, but they will likely take massive write-downs on their stranded GPU capacity.

Key concerns include:
• The rapid depreciation of AI infrastructure, which becomes obsolete in just three years.
• The unsustainable business models of AI companies, which are losing billions of dollars.
• The risks of a bubble burst, which could lead to significant losses for investors and lenders.
• The potential impact on the broader economy, including the risk of correlation across portfolios.

The AI infrastructure boom is following a familiar pattern, with excessive spending and flawed business models. The industry's focus on short-term gains and speculation has led to a situation where companies are lighting $15 on fire for every $1 earned. This approach is not only unsustainable but also poses significant risks to investors and the broader economy.

As the bubble bursts, likely between 2026 and 2028, the consequences will be severe, with many companies facing financial difficulties and potential collapse. The industry's failure to learn from previous bubbles and adopt a more measured approach to AI infrastructure buildout has led to this point. The outcome will be a significant destruction of capital, with investors and lenders facing substantial losses.

The AI infrastructure bubble is a ticking time bomb, and it's only a matter of time before it bursts, leaving a trail of destruction in its wake.