Steel
GPU-accelerated rigid-body dynamics solver for robot learning
- Upvotes
- 0
- DR
- 67 no change since last check
- Followers
- 0
Steel is a GPU-accelerated physics solver built into Newton that enables efficient simulation of diverse robots, tasks, and environments together for robot
About Steel
Steel (Scalable Throughput-optimized Engine for Embodied Learning) is a GPU-accelerated rigid-body dynamics solver integrated into the open-source Newton Physics Engine. Developed in collaboration with NVIDIA, it accelerates robot learning in NVIDIA Isaac Lab by efficiently handling heterogeneous workloads combining different robots, tasks, and environments.
Unlike traditional physics solvers optimized for replicating single tasks thousands of times, Steel organizes dynamics and constraints per environment, groups articulations by joint count, and routes work dynamically. This design allows it to handle variable complexity within environments and across heterogeneous batches without padding or wasted computation.
Steel achieves 3.8× the simulation throughput of PhysX on 20 robot-learning benchmarks and up to 13.0× on heterogeneous batches mixing different robot types and tasks. The solver supports coupled physical interactions including cloth, cables, soft bodies, granular materials, and differentiable simulation for model-predictive control.
- reinforcement learning
- robot learning
- heterogeneous simulation
- gpu acceleration
- open-source
- physics simulation
Features worth knowing
- 3.8× PhysX throughput on 20 benchmarks geometric mean
- 13.0× performance on heterogeneous task mixtures
- Per-environment work scheduling for efficiency
- Reduced-coordinate articulation dynamics
- Differentiable simulation
- Rigid-deformable coupling support
- Compliant and hydroelastic contact
- 62% less GPU memory than PhysX at matched batch size
- 3.90 million environment steps per second peak throughput
Pricing
Pricing model: Open source
Follow Steel and hear about their milestones, new launches and demo videos by email. Stop any time.
Steel metrics
| Metric | Value |
|---|---|
| Domain Rating | 67 no change since last check |
| Trust Flow | 24 no change since last check |
| Citation Flow | 40 no change since last check |
| Referring domains | 2,061 no change since last check |
| External backlinks | 16,624 no change since last check |
Ahrefs Domain Rating scanned 10 Nov 2026, updated 9 Nov 2026. Main topic: Business/Financial Services (24). Majestic scanned 9 Nov 2026, updated 9 Nov 2026. Link stats via Majestic. Trust Flow, Citation Flow and Topical Trust Flow are trademarks of Majestic-12 Ltd.
Steel domain facts
Reveal the domain facts
| Domain age | 3 years |
|---|---|
| Registered | |
| Expires | |
| Registrar | NameCheap, Inc. |
| DNSSEC | Not signed |
Steel launch results
| Measure | At launch, then latest |
|---|---|
| Domain Rating (Ahrefs) | 67 to 67 no change since launch 9 Nov 2026, then 10 Nov 2026 |
| Trust Flow (Majestic) | 24 first measured 9 Nov 2026 |
| Referring domains (Majestic) | 2,061 first measured 9 Nov 2026 |
Day 1 since launch. Next receipt: day 30, 9 Dec 2026. The figures report a change; they do not say what caused it, and many things can. No figure here is a guess.
Launching something soon?
Free tips from Best Digital Launch for getting ready, launch day, the week after and what to do next.
- Set up
- On launch day
- The days after
- Beyond launch
Compare and explore
FAQ
- What is the performance advantage of Steel over PhysX?
- Steel delivers 3.8× PhysX's simulation throughput on 20 robot-learning benchmarks with geometric mean, and is faster on 19 of the 20 tasks. On heterogeneous batches combining different robot tasks, the advantage reaches 13.0× PhysX's peak measured throughput.
- How does Steel handle heterogeneous workloads?
- Steel organizes dynamics and constraints per environment, groups articulations with similar joint counts for efficient GPU execution, and routes contact-solving work according to each environment's current demands. This reduces padding and unnecessary solve work when different robots and tasks run together in one batch.
- Where can I access Steel?
- Steel is available as open-source software integrated into Newton. Researchers can inspect, reproduce, extend, and improve it through the Newton ecosystem.
- What physical interactions does Steel support?
- Steel supports rigid-deformable coupling, cloth manipulation, granular materials, compliant and hydroelastic contact, soft bodies, and differentiable simulation for model-predictive control.
Leave a review for Steel
Nobody has reviewed Steel yet. Start things off.
A review is always free and can never be paid for. Go to the reviews and discussion
Reviews and chat · Product news · Send it on · Run this company? Claim it here · Report it
Steel has a free listing. The owner can unlock the full profile with a paid launch for $5:
- Founder story (locked)
- Milestones (locked)
- Demo video (locked)
- Dofollow website link (locked)
Alternatives to Steel
S2Arm
Native ARM64 game porting for Mac and iPhone
Game dev tools0 votes. Log in to add your vote for S2ArmChessVox
Free chess game analysis and personalized training built from your own games
Game dev tools0 votes. Log in to add your vote for ChessVoxSchem Canvas
AI-powered Minecraft build generator for Java Edition with 3D preview and schematic export
Game dev tools0 votes. Log in to add your vote for Schem CanvasAI Sprite Generator
Create pixel-art characters, animate sprites and edit frames for your 2D game
Game dev tools0 votes. Log in to add your vote for AI Sprite Generator
