Prices & markets

Follow what AI costs—and what the market believes

Track model prices, AI-related markets, cloud and inference economics, hardware supply and the commercial signals behind the AI build-out.

Use cost

Model and API pricing

Track published model prices and the changes that alter the economics of production AI systems.

AI pricingCurrent published API input and output prices.Pricing historyPermanent record of verified pricing changes.Models & BenchmarksPut cost beside capability, access and deployment fit.

API pricessubscriptionsfree tierstoken costsbatch pricingcached inputenterprise pricing

Market pulse

AI companies and public markets

Separate the commercial market signal from product hype by tracking selected AI-related companies and the events that move expectations.

AI marketsSelected United States AI-related market closes and context.AI industryCompany, product and infrastructure developments behind market moves.AI compute as a project-finance assetHow Nvidia’s financing partnerships connect capital, customer contracts, power, permitting and hardware-refresh risk.AI’s $1 trillion data-center lease pipelineWhy future data-center lease commitments matter without treating them as the same thing as debt.Why AI demand and AI stocks can divergeInfrastructure economics, capacity and market expectations explained.

AI stocksvaluationscapital spendingrevenuemarginspartnershipsacquisitions

Unit economics

Cloud and inference economics

Serving AI at scale depends on utilization, memory, latency, model size, power and infrastructure—not only the advertised token price.

Why AI inference is splitting into two hardware marketsCerebras, AWS, AMD and NVIDIA show why compute-heavy prefill and memory-heavy decode are increasingly being optimized separately.AI model routing and the cost-quality-latency trade-offWhy production systems may choose different models per request instead of paying frontier-model prices for every task.AI pricingStart with the published model price layer.AI compute project financeWhy outside capital, contracted demand, power and accelerator refresh cycles now matter to compute economics.Data-center lease commitments explainedFollow the long-duration capacity commitments behind the AI infrastructure build-out.AI infrastructure economics explainerUtilization, inference economics, cloud margins and depreciation.Industry & PolicyFollow data-center, chip and cloud capacity developments.

inference costutilizationlatencycloud marginspowerdepreciationcapacity

Physical economics

Chips, hardware and local AI economics

Hardware ownership can change privacy, control and recurring costs, but only when memory, performance and utilization justify the investment.

Local AI hardwarePublished hardware prices, memory and practical model fit.N3 and CoWoS supply constraintsWhy fabrication and packaging capacity affect AI accelerator supply.Open and local modelsConnect hardware choices to the models you want to run.

GPU pricesacceleratorsHBMpackaginglocal ownershipcloud vs localhardware supply