Pathway has raised additional funding at a $500 million valuation as the AI company prepares to publish benchmarks for its Post-Transformer models. The financing brings Pathway’s total seed funding to $30 million, with most of the new capital expected to go toward additional computing capacity, including new NVIDIA GB300 systems.
The financing includes participation from Id4 Ventures, TQ Ventures, Red Bridge Ventures, Kadmos Capital, and WS Investment Co., the investment arm of Wilson Sonsini. Databricks Chief AI Scientist Jonathan Frankle also participated as an angel investor.
Pathway is developing BDH, a Post-Transformer architecture that the company believes can address some of the limitations created by continually scaling Transformer-based AI models with more data, GPUs, energy, and capital.
The company describes BDH as a scale-free, biologically inspired state-space sequence architecture operating in latent space. Pathway said the architecture can learn continuously from relatively small amounts of data without requiring repeated retraining cycles.
Pathway published research on BDH last October. In subsequent testing against approximately 250,000 difficult Sudoku puzzles, Pathway said BDH solved 97.4% without chain-of-thought, backtracking, or external tools.
The company’s thesis is that the next stage of AI development may depend less on continually increasing computational scale and more on developing fundamentally different architectures capable of combining reasoning, memory, and continual learning more efficiently.
Pathway is now working to expand BDH’s capabilities and train multipurpose models for production environments where continual learning, personalization, and reliable reasoning are particularly important, including financial services, healthcare, and technology.
The company also named former Google DeepMind Gemini Group Product Manager Adam Kurzrok as Chief Product Officer. Kurzrok will oversee product direction, including how models based on BDH are packaged, evaluated, and deployed.
Pathway has also formalized an advisory group that includes Transformer co-inventor Łukasz Kaiser, Jonathan Frankle, NYU Tandon Computer Science and Engineering Chair Martín Farach-Colton, and economist Jacques Attali.
KEY QUOTES:
“The industry has spent several years proving how far scale can take you, and that was worth establishing. The harder question comes next. When scale stops being the answer, does the next era of AI belong to whoever assembles the most compute, or to whoever is willing to reconsider the design? We have made our bet clear enough. The next few months will settle it.”
Zuzanna Stamirowska
