AI Answers Could Soon Cost More When You Want Them Fast

A new paper on arXiv argues that the economic theory of LLM pricing treats tokens as a homogeneous commodity, focused on aggregate token count as the main feature buyers and sellers consider. Authors Ian McDougall and Karthikeyan Sankaralingam instead model inference as a service market where buyers have three-dimensional private information — willingness-to-pay, task volume, and time preference — and utility depends on latency slack alongside token quantities. Their main result is a separation theorem: discrete hardware tiers induce endogenous self-selection on time preferences, reducing three-dimensional screening to standard one-dimensional screening within each tier. They derive the cost structure from GPU inference physics (compute-bound prefill and bandwidth-bound decode) and characterize optimal tiered mechanisms via virtual-value techniques. Optimal per-task prices are volume-independent, providing theoretical grounding for flat per-token API pricing. Calibrating to 8-GPU clusters of H100 and B200 hardware, they find the separation theorem holds in 83% of 105 tested configurations overall, rising to 96% at economically relevant WTP scales. A seller adopting two-tier pricing under the optimal mechanism captures 26-66% higher profit than the best single-tier alternative, with gains driven by efficient cross-tier allocation in regimes where hardware costs are a significant fraction of per-request value.