Cognite and NVIDIA Operationalize NV-Tesseract to Transform Industrial Forecasting

Strategic integration combines Cognite’s Industrial Knowledge Graph and AI agents with NVIDIA’s time-series AI model to drive operational efficiency and predictive accuracy at scale

SAN JOSE, Calif.--(BUSINESS WIRE)--Cognite, the global leader in Industrial AI, today announced the integration of NVIDIA’s NV-Tesseract family of models into the Cognite AI and Data Platform to operationalize foundational forecasting models for heavy industry. This integration leverages the context-rich data within Cognite’s platform to power NVIDIA’s advanced time-series AI, delivering unprecedented predictive accuracy for critical manufacturing processes.

Unlocking the full potential of industrial AI requires more than just powerful models; it requires unified data that understands the physical world. This integration bridges that gap by connecting the Industrial Knowledge Graph within Cognite Data Fusion^®—the platform's core data foundation that makes data AI-ready—to NVIDIA’s NV-Tesseract models packaged as NVIDIA NIM microservices. Furthermore, the solution transcends raw predictions by leveraging Cognite Atlas AI^™—the platform’s low-code industrial agent workbench—and other no-code analytics and collaboration tools to visualize and engage with forecasting results, turning model outputs into actionable insights directly within daily operational workflows.

Celanese, a global chemical and specialty materials leader, has deployed this advanced predictive modeling solution at its Clear Lake, Texas, facility. Together with Cognite, Celanese identified a high-value opportunity to improve the prediction of reaction water levels in certain production units—a metric critical to stable and efficient plant operation. Currently, Celanese relies on manual samples to adjust existing prediction models. However, operators frequently observe "bias jumps" when new sample data is entered, indicating that current legacy models could more accurately predict real-time conditions. By feeding historical time series data from Cognite into NVIDIA’s NV-Tesseract forecasting model, Celanese aims to eliminate these inconsistencies.

"By integrating the Cognite AI and Data Platform with NVIDIA Tesseract NIM, we have transitioned from reactive monitoring to a system of intelligence that predicts operational states before they impact the bottom line," said Chirayu Shah, Chief Product Officer at Cognite. "Our real-world work with Celanese demonstrates that we aren’t just solving technical hurdles, we’re unlocking higher throughput and asset efficiency. When contextualized industrial data meets advanced modeling, we empower our customers to move beyond historical data to a proactive, exception-driven future."

“NVIDIA NV-Tesseract NIM delivers state-of-the-art forecasting and foundational time-series capabilities, but these models require high-fidelity data to be effective in complex physical environments,” said Marc Spieler, Senior Managing Director of Global Energy Industry at NVIDIA. “Cognite’s Industrial Knowledge Graph provides the necessary context and domain expertise to ground the models and ensures that our advanced forecasting drives tangible value on the plant floor.”

Cognite will be demonstrating the NV-Tesseract NIM integration at NVIDIA GTC, taking place this week in San Jose, California. Learn more here: cognite.com/en/partners/nvidia.

About Cognite

Cognite makes AI work for industry. Leading energy, manufacturing, and power & renewables enterprises choose Cognite to deliver secure, trustworthy, and real-time data to transform their asset-heavy operations to be safer, more sustainable, and profitable. Cognite provides a user-friendly, secure, and scalable industrial data & AI platform that makes it easy for all decision-makers, from the field to remote operations centers, to access and understand complex industrial data, collaborate in real-time, and build a better tomorrow. Visit us at www.cognite.com, and follow us on LinkedIn.