Pathway, an AI startup, has secured a $500 million valuation ahead of public benchmarks for its Post-Transformer models. The company argues that the fundamental architecture, rather than inputs like data, GPUs, and capital, is the limiting factor in AI. In October, Pathway published research on BDH, a Post-Transformer architecture built on a novel, scale-free, biologically inspired state-space sequence architecture, which was ranked as the #1 Paper of the Day on Hugging Face. The BDH architecture is designed for continuous learning from small datasets, eliminating retraining cycles and avoiding escalating compute costs. In a recent test, BDH solved 97.4% of Sudoku Extreme puzzles, a task where leading reasoning models reportedly scored near zero. The company has received additional funding, bringing its total seed funding to $30 million, primarily to enhance compute capacity, including new GB300s. Key investors include Id4 Ventures, TQ Ventures, Red Bridge Ventures, Kadmos Capital, and WS Investment Co., alongside angel investor Jonathan Frankle, Chief AI Scientist at Databricks. Pathway has expanded its team with Adam Kurzrok, formerly a Group Product Manager for Gemini at Google DeepMind, joining as Chief Product Officer. The company is formalising its advisory group, which includes Łukasz Kaiser, co-inventor of the Transformer architecture; Jonathan Frankle; Martín Farach-Colton, Chair of Computer Science and Engineering at NYU Tandon; and Jacques Attali,