Pathway has announced that its post-Transformer AI architecture, BDH (Dragon Hatchling), now runs on NVIDIA AI infrastructure and the Amazon Web Services (AWS) cloud stack. This technical integration is designed to enable a new class of adaptive and continuously learning AI systems that evolve with business operations, contrasting with the static knowledge states of traditional Transformer-based models.
Pathway's post-Transformer AI architecture, BDH, now runs on NVIDIA AI infrastructure and AWS.
The model emphasizes continuous learning, allowing it to adapt over time rather than remain static.
It promises greater efficiency, lower latency, and full observability into the model's live state.
The collaboration leverages AWS for scalable compute and NVIDIA Hopper architecture for performance.
BDH is designed for enterprise use cases requiring complex, adaptive thinking and low latency.
The model will be showcased at AWS re:Invent 2025, with design partnerships launching late 2025.
The initiative targets a fundamental shift in AI capabilities. While Transformer-based large language models (LLMs) have driven market growth, they maintain consistent knowledge after training. Pathway's BDH architecture, inspired by brain-like networks, is engineered for continuous learning through dynamic state management. This allows models to adjust their internal representations as new data arrives, aiming to unlock applications requiring intelligence that evolves alongside business processes.
Pathway positions BDH as an enterprise-first model, built from the ground up for new categories of complex applications. These include use cases demanding low latency, high observability, or sophisticated reasoning over time. The technical foundation combines AWS as the preferred cloud provider for scalable deployment with NVIDIA's Hopper architecture to support the high-performance, low-latency demands of continuous adaptation. “Our integration of NVIDIA and AWS technologies represents a fundamental shift from static to adaptive intelligence,” said Zuzanna Stamirowska, CEO and co-founder of Pathway.
The BDH architecture challenges conventional deep learning assumptions, particularly the trade-off between model size and interpretability. Pathway proposes that with its scale-free network design, increased scale can lead to greater clarity through neuron specialization, bridging concepts from neuroscience and machine learning. The company also claims BDH achieves competitive performance on general-purpose hardware while showing potential for significantly faster inference on specialized AI processors, which could reduce both operational costs and latency.
The core promise of this integration is to open entirely new categories of AI applications for enterprises. By moving beyond the static capabilities of current models, Pathway aims to enable systems that can learn continuously from live data streams, offering full observability into their reasoning process. Jason Bennett of AWS noted the partnership aligns with enabling innovators to build "entirely new categories of continuously learning AI applications."
The deployment of Pathway's BDH architecture on AWS and NVIDIA infrastructure marks a significant step toward commercializing a novel approach to AI. By focusing on continuous adaptation, observability, and enterprise-scale performance, Pathway is differentiating itself in a market dominated by Transformer-based models. This collaboration could pave the way for more dynamic, efficient, and transparent AI systems capable of tackling complex, real-time business problems that have remained out of reach for current generation AI.
About Pathway
Pathway is shaking the foundations of artificial intelligence by introducing the world’s first post-transformer model that adapts and thinks just like humans.
Pathway’s breakthrough architecture outperforms Transformer and provides the enterprise with full visibility into how the model works. Combining the foundational model with the fastest data processing engine on the market, Pathway enables enterprises to move beyond incremental optimization and toward truly contextualized, experience-driven intelligence. The company is trusted by organizations such as NATO, La Poste, and Formula 1 racing teams.
Pathway is led by co-founder & CEO Zuzanna Stamirowska, a complexity scientist who created a team consisting of AI pioneers, including CTO Jan Chorowski who was the first person to apply Attention to speech and worked with Nobel laureate Goeff Hinton at Google Brain, as well as CSO Adrian Kosowski, a leading computer scientist and quantum physicist who obtained his PhD at the age of 20.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.