The inference efficiency of GPUs has increased 40x since 2019 according to Jensen Huang, CEO of Nvidia.  This measurement represents the number and cost of tokens processed by an AI model running on a GPU, where a token is the unit of accounting for language models.

A million tokens is roughly equivalent to 2500 pages of documents, contracts, or spreadsheets.  A 4000% increase in horsepower since 2019, means that a task that costs $400 to process on a GPU in 2019, and costs $10 in 2025, will be a penny by 2030. Think about that, another 1000x projected increase in the next five years will mean reasoning costs will become effectively free. This change of epoch has implications for all businesses of any size: leverage this technology now, or risk falling farther behind. And every month counts.

iClerk is in the business of helping companies of all sizes, capitalize on the use of AI through the implementation of AI Agents.  Agents that let you leverage the best of the best AI technologies without worrying about vendor lock-in. By providing access to any AI model, iClerk Agents ensure optimized tool use based upon task requirements, cost, performance, and application characteristics.  These features combine with memory, traceability, and the management of data context to provide easily auditable access to your data, regardless of where that data comes from, or upon which IT system they are processed.  AI Agents automate the proprietary, tacit processes that drive daily business operations, for a fraction of your current cost to perform the work.

‍