Get your free personalized podcast brief

We scan new podcasts and send you the top 5 insights daily.

Erik Bernhardsson warns that buyers hoarding expensive multi-year GPU contracts face acute risks if hardware prices experience their historical downward pull. If prices drop, utilization dips and secondary resale markets disappear simultaneously, creating severe balance sheet shocks for startups that bought heavily at the top of the market.

Related Insights

Historically, infrastructure from tech bubbles (e.g., fiber optic cables) had long-term value for second-wave investors. AI's core infrastructure, GPUs, has a short 2-3 year shelf life, creating a unique and devastating "depreciation bomb" risk for investors caught in the hype cycle.

The call for a "federal backstop" isn't about saving a failing company, but de-risking loans for data centers filled with expensive GPUs that quickly become obsolete. Unlike durable infrastructure like railroads, the short shelf-life of chips makes lenders hesitant without government guarantees on the financing.

Despite the rapid pace of hardware innovation, the value of older NVIDIA GPUs like the H100 is holding strong. Cloud provider CoreWeave reports these chips are retaining 90-95% of their pricing power over a 5-6 year lifespan because compute demand far outstrips supply.

Don Wilson says the trillion-dollar question is what a B200 will be worth four or five years out. H100 on-demand prices fell for months, then rose after Anthropic's new model boosted inference demand, while token prices moved differently. That volatility is why he argues forward contracts are badly needed.

While the industry standard is a six-year depreciation for data center hardware, analyst Dylan Patel warns this is risky for GPUs. Rapid annual performance gains from new models could render older chips economically useless long before they physically fail.

Accessing next-generation GPUs at scale is no longer a simple purchase. The market now demands three-to-five-year commitments with a significant portion (20-30%) of the total contract value paid upfront. This makes a company's cost of capital a critical competitive factor in acquiring compute capacity.

The current GPU shortage is a temporary state. In any commodity-like market, a shortage creates a glut, and vice-versa. The immense profits generated by companies like NVIDIA are a "bat signal" for competition, ensuring massive future build-out and a subsequent drop in unit costs.

Despite reports of falling H100 spot rental prices, contract prices for sustained GPU workloads are rising. This indicates the market is shifting from short-term, experimental use to long-term, committed production deployments, reflecting stronger, not weaker, underlying demand for AI infrastructure.

Contrary to expectations of easing supply, the GPU shortage has intensified since 2023. With clearer AI business models, mega-customers like OpenAI and Anthropic are spending even more aggressively, creating a fierce bidding war that pushes startups out.

Despite narratives of scarcity, GPU utilization in some data centers is as low as 35-40%. This is due to colossal hoarding and double-ordering by companies terrified of being caught short in a future parabolic moment. This behavior creates a phantom demand and points to a future supply glut.