NVIDIA (NASDAQ: NVDA) announced the DSX platform at its GTC Taipei conference, providing infrastructure builders with integrated tools for designing and operating AI facilities. The platform combines open source software libraries, application programming interfaces, reference designs, and partner technologies into a unified framework for AI factory development.

The platform includes two new software components. DSX MaxLPS optimizes token performance per megawatt through 45-degree Celsius liquid cooling and in-rack technologies, enabling operators to run up to 40% more GPUs at their most energy-efficient operating point. DSX OS provides open source, modular software for lifecycle management, intelligence scheduling, runtime consistency, and multi-tenant operations.

"We're not just shipping chips — we're giving every infrastructure builder a complete playbook to build AI factories," said Jensen Huang, NVIDIA's founder and CEO. "With the DSX platform, you can simulate the entire factory before you spend a dollar, validate performance before a single rack is installed and operate with the kind of reliability that production AI demands."

The platform also incorporates existing features including DSX Reference Design for validated AI factory architectures, DSX Sim for high-fidelity simulation, DSX Flex for connecting AI factories to power-grid services, and DSX Exchange for integrating compute and energy signals.

Cloud partners CoreWeave, Crusoe, Firmus, IREN, Lambda, Nebius, Nscale and Yotta Data Services are deploying DSX platform components. System manufacturers including Dell Technologies, HPE, Lenovo, Supermicro, ASUS, Foxconn, GIGABYTE, Pegatron, Quanta Cloud Technology, Wistron and Wiwynn are building DSX-ready systems.

DSX Flex is being tested in a commercial pilot with Emerald AI and Silicon Valley Power to demonstrate grid-responsive AI factories that can adjust power consumption based on utility signals while maintaining AI workload performance.