Architect Financial Technologies Launches Liquid Inference Platform to Trade AI Model Output
Architect Financial Technologies has launched Liquid Inference, an auction platform for large language model (LLM) inference that introduces exchange-style price discovery and competitive quoting. The platform allows buyers to set rules around cost, latency, data handling, and other parameters while enabling providers to compete on pricing and performance. Each request is awarded based on the lowest-priced offer meeting buyer criteria, with transparent order books and usage-based billing. The launch expands Architect's AI Exchange offerings beyond compute forwards, swaps, options, and futures into real-time inference trading. The platform supports existing coding tools like Claude Code and Codex while offering a chat-style interface for users. Brett Harrison, CEO of Architect, emphasized the growing cost of AI inference in tech budgets and how Liquid Inference introduces competitive pricing mechanisms similar to financial exchanges. Architect is also operationalizing its American Innovation Exchange as a U.S.-based futures market focused on GPU compute and AI commodities pending regulatory approval.
Topics: AI derivatives trading principles inference marketplaces LLM auction platform