GPU access once depended on each lab's own budget. Now 110 students share simulation GPUs, training robots before risking real hardware.
A university robotics program spanning 19 core faculty, 40 affiliated faculty, and about 400 students and researchers training everything from small autonomous cars to humanoid soccer robots.
Robots are expensive and scarce, and every physical experiment risks breaking hardware during a reset. Realistic training requires simulators with...
“There’s just no way I would be able to run these experiments at scale without these compute nodes.”
Public research university with diverse academic programs and research initiatives.
NVIDIA is a technology company that specializes in semiconductors, graphics processing units, and artificial intelligence for applications in data centers, gaming, and more.
The University of Texas at Austin's Robot skills training is part of this use case:
Related implementations across industries and use cases
Robotics labs capped at 6–7 training runs/day on desktop GPUs. Shared campus compute lifted that to 100+ runs daily.
Compute limits capped labs at 1–2 targets/year. A shared 760-GPU cluster let one team screen 10M molecules in a week.
Traditional courses left students unprepared to navigate massive codebases. Using AI, they bypass boilerplate to build sophisticated apps.
Simulations were stuck in 2D and took hours. Physics-informed AI now models 3D urban airflow in seconds.
CFD and FEA simulations once took hours or days per run. Now surrogate models evaluate engine design changes in seconds.
Deal data scattered across systems kept sellers from districts. AI unified it—leadership moved from retrospective reports to live intel.
Ops teams combed Gong, Zendesk, Slack, and Salesforce to answer routine questions. Now one query surfaces cited answers.
Tournaments running simultaneously meant an hour of manual checks each. AI agents now run them in minutes, freeing the team to be proactive.
Large AI training jobs meant fighting for preemptible slots or leaving campus. Marlowe gave any lab guaranteed multi-node access on demand.
GPU access once depended on each lab's own budget. Now 110 students share simulation GPUs, training robots before risking real hardware.
A university robotics program spanning 19 core faculty, 40 affiliated faculty, and about 400 students and researchers training everything from small autonomous cars to humanoid soccer robots.
Robots are expensive and scarce, and every physical experiment risks breaking hardware during a reset. Realistic training requires simulators with...
“There’s just no way I would be able to run these experiments at scale without these compute nodes.”
Public research university with diverse academic programs and research initiatives.
NVIDIA is a technology company that specializes in semiconductors, graphics processing units, and artificial intelligence for applications in data centers, gaming, and more.
The University of Texas at Austin's Robot skills training is part of this use case:
Related implementations across industries and use cases
Robotics labs capped at 6–7 training runs/day on desktop GPUs. Shared campus compute lifted that to 100+ runs daily.
Compute limits capped labs at 1–2 targets/year. A shared 760-GPU cluster let one team screen 10M molecules in a week.
Traditional courses left students unprepared to navigate massive codebases. Using AI, they bypass boilerplate to build sophisticated apps.
Simulations were stuck in 2D and took hours. Physics-informed AI now models 3D urban airflow in seconds.
CFD and FEA simulations once took hours or days per run. Now surrogate models evaluate engine design changes in seconds.
Deal data scattered across systems kept sellers from districts. AI unified it—leadership moved from retrospective reports to live intel.
Ops teams combed Gong, Zendesk, Slack, and Salesforce to answer routine questions. Now one query surfaces cited answers.
Tournaments running simultaneously meant an hour of manual checks each. AI agents now run them in minutes, freeing the team to be proactive.
Large AI training jobs meant fighting for preemptible slots or leaving campus. Marlowe gave any lab guaranteed multi-node access on demand.