While focus is on massive supercomputers for training next-gen models, the real supply chain constraint will be 'inference' chips—the GPUs needed to run models for billions of users. As adoption goes mainstream, demand for everyday AI use will far outstrip the supply of available hardware.

When major infrastructure like AWS or Cloudflare goes down, it affects many companies simultaneously. This creates a collective "mulligan," meaning individual startups aren't heavily penalized by users for the downtime, as the issue is widespread. The exception is for mission-critical services like finance or live events.

The initial deployment of a new AI cluster sees a high failure rate, with 10-15% of new-generation GPUs like Blackwell needing to be returned or reseated. This "infant mortality" is a standard operational challenge for data centers, underscoring the physical difficulties of scaling AI infrastructure with bleeding-edge chips.

The focus in AI has evolved from rapid software capability gains to the physical constraints of its adoption. The demand for compute power is expected to significantly outstrip supply, making infrastructure—not algorithms—the defining bottleneck for future growth.

The critical constraint on AI and future computing is not energy consumption but access to leading-edge semiconductor fabrication capacity. With data centers already consuming over 50% of advanced fab output, consumer hardware like gaming PCs will be priced out, accelerating a fundamental shift where personal devices become mere terminals for cloud-based workloads.

Hyperscalers face a strategic challenge: building massive data centers with current chips (e.g., H100) risks rapid depreciation as far more efficient chips (e.g., GB200) are imminent. This creates a 'pause' as they balance fulfilling current demand against future-proofing their costly infrastructure.

Countering the narrative of rapid burnout, CoreWeave cites historical data showing a nearly 10-year service life for older NVIDIA GPUs (K80) in major clouds. Older chips remain valuable for less intensive tasks, creating a tiered system where new chips handle frontier models and older ones serve established workloads.

While power supply is a current data center bottleneck, a more significant long-term risk is technological disruption. Chip innovations promising 10-1000x more power efficiency could make today's massive, power-centric data center investments obsolete or oversized before they are fully utilized.

Responding to the AI bubble concern, IBM's CEO notes high GPU failure rates are a design choice for performance. Unlike sunken costs from past bubbles, these "stranded" hardware assets can be detuned to run at lower power, increasing their resilience and extending their useful life for other tasks.