Fears of rapid GPU depreciation are overwrought. While the newest chips are needed for frontier model training, older architectures have an 'incredibly fat tail' of productive uses. CoreWeave is contracting 2020-era A100s through 2029 for tasks like batch computing and medical research, proving their long-term value.

Countering the narrative of rapid hardware obsolescence, CoreWeave is still signing new contracts for NVIDIA's 2020-era A100 GPUs that extend to 2029. This proves the long-term useful life and economic viability of older chip generations for specific, optimized AI workloads, challenging the idea that only the latest hardware is valuable.

Separating inference into "prefill" (memory-bound) and "decode" (bandwidth-bound) tasks is a game-changer for hardware longevity. It allows older GPUs to be used for prefill tasks indefinitely, extending their useful economic life from 3-4 years to 10-15 years, a boon for data centers and their financiers.

AI software is improving so rapidly that older hardware, like a three-year-old NVIDIA inference chip, is now more profitable than it was when new. This phenomenon, where software advancements outpace hardware depreciation, is unprecedented and makes existing infrastructure increasingly valuable.

The fear of chip depreciation is mitigated by repurposing older GPUs. While the latest models are crucial for high-speed training, older, less powerful chips are perfectly suitable and cost-effective for running inference, extending the hardware's useful life and long-term value.

Andreessen highlights a unique economic phenomenon: the pace of AI software improvement outstrips hardware depreciation. This means a three-year-old NVIDIA inference chip can generate more revenue today than when it was new, a complete reversal of typical tech hardware value cycles.

Contrary to fears of rapid obsolescence, new domain-specific accelerators (DSAs) can be paired with older GPUs to handle specific tasks. This disaggregated approach extends the useful life of GPUs to 10-15 years, lowering financing costs for compute providers and invalidating bear cases.

Contrary to the belief that AI hardware becomes obsolete quickly, older GPUs like A100s will have a long depreciable life. As companies optimize costs, they'll use model routing to send simple queries to older, cheaper hardware, extending its utility for six to eight years.

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.