MUMBAI, India, March 13 -- Intellectual Property India has published a patent application (202521133888 A) filed by Dr. Tabrez Qureshi; Dr. Sandeep Gupta; Dr. Mohd Dilawar; Dr. Mahazbeen Sayed; Dr. Akash Dwivedi; Dr. Ashish Kumar Sinha; Dr. Vijay Pratap Singh; and Mr. Anas Ali, Bhopal, Madhya Pradesh, on Dec. 30, 2025, for 'system and method for thermal management in immersion-cooled high-performance computing devices.'
Inventor(s) include Dr. Tabrez Qureshi; Dr. Sandeep Gupta; Dr. Mohd Dilawar; Dr. Mahazbeen Sayed; Dr. Akash Dwivedi; Dr. Ashish Kumar Sinha; Dr. Vijay Pratap Singh; and Mr. Anas Ali.
The application for the patent was published on March 13, under issue no. 11/2026.
According to the abstract released by the Intellectual Property India: "The present invention discloses an autonomous thermal management platform for immersion-cooled HPC devices comprises a thermal-aware SRAM array with embedded on-die thermal sensors and adaptive body-biasing circuitry, a two-phase immersion cooling assembly equipped with an array of fast-response micro-valve actuators and variable-speed pumps, and an edge-resident multi-modal artificial-intelligence controller. The controller continuously fuses on-die thermal telemetry, workload performance counters, and coolant-state telemetry into a hybrid predictive model embedding governing equations of transient heat conduction and incompressible fluid flow. In a single inference pass, the controller simultaneously issues body-bias voltage commands and zoned coolant-flow commands within a closed-loop cycle of less than one second, while an online learning mechanism continuously retrains the predictive model using only locally computed prediction residuals. The platform thereby achieves coordinated microsecond-scale leakage suppression and millisecond-scale phase-change optimisation without external connectivity or intervention."
Disclaimer: Curated by HT Syndication.
