Unlike durable infrastructure like railways or fiber optic cables, AI's core component—expensive GPUs—becomes obsolete in just 2-3 years. This creates a permanent, recurring cost, a 'tax on innovation,' making profitability much harder to achieve compared to previous tech revolutions.
Major tech shifts like canals, railways, and fiber optics consistently wiped out their initial funders who financed essential but unprofitable infrastructure. A second wave of companies then acquired these assets for pennies on the dollar and built enormously valuable businesses on top.
Just as they did with subprime mortgages, large banks are repackaging risky AI data center debt—backed by rapidly depreciating hardware—into complex financial products. These are then sold to pension funds, insurers, and private credit, transferring risk away from the banks and onto the public.
To appear more financially viable, major AI companies are accused of booking their GPUs with a 5-6 year lifespan, despite experts claiming the real functional obsolescence is 2-3 years. This accounting maneuver intentionally hides massive losses and inflates valuations ahead of IPOs.
The wave of AI companies going public is presented as a growth opportunity, but it functions mechanically as an "exit" for early investors. It allows insiders to cash out and pass the immense financial risks of unprofitable, capital-intensive businesses onto the public market, dubbed "dumb money."
Given the high probability of a crash wiping out individual companies, the smart play is to bet on the entire AI sector's long-term success. Avoid debt, ignore short-term volatility, and hold a diversified portfolio for 20+ years, a timeframe that has historically overcome even major depressions.