Figure AI vs NVIDIA GR00T / Isaac
Figure AI vs NVIDIA GR00T / Isaac: side-by-side comparison of two robotics foundation models & agent stacks systems — architecture, taxonomy, license, pricing, MCP/A2A support, and direct edges.
Figure AI · NVIDIA GR00T / Isaac
Cost & capability
| Figure AI | NVIDIA GR00T / Isaac | |
|---|---|---|
| Capability band | entry | entry |
| Capability composite | 30 | 35 |
| Cost tier | — | free |
| $/Mtok input | — | 0 |
| $/Mtok output | — | 0 |
Where they differ (11)
Rows where both sides have data and the values disagree — the shortlist of dimensions that actually distinguish these two systems.
| Figure AI | NVIDIA GR00T / Isaac | |
|---|---|---|
| Capability composite | 30 | 35 |
| Type | Humanoid robot maker (Figure 01 / Figure 02 / Helix VLA model) | Robotics foundation model + simulation stack (NVIDIA) |
| Created | 2022 (founded) | 2024-03 (GTC announcement) |
| Latest release | Helix VLA (Feb 2025); Figure 03 in development | GR00T N1 (2B) on HuggingFace Mar-2025 |
| Funding | $675M Series B Feb-2024 (Microsoft, OpenAI, Bezos, NVIDIA, Intel, LG Innotek; $2.6B val); reported $39.5B val funding Q2 2025 | NVIDIA parent — public company, $3T+ mkt cap |
| Backend storage | Internal Figure cloud | Caller-managed (on-robot + cloud) |
| Deployment | Direct customer deployment (humanoid robot) | On-robot (Jetson Thor) + simulation (Isaac Sim) + training cloud (Cosmos) |
| API surface | No public API; closed humanoid platform | Python; Isaac Sim; HuggingFace weights |
| Multi-tenancy | Per-customer fleet | Per-developer / per-robot |
| Optimised for | End-to-end humanoid VLA — full-body + dexterous hand control | Full-stack robotics FM: model + sim + compute, multi-partner |
| Anti-fit | Closed humanoid platform — no developer API; alpha-customer-only | Requires NVIDIA hardware (Jetson Thor / GPUs); Isaac Sim learning curve |
At a glance
| Figure AI | NVIDIA GR00T / Isaac | |
|---|---|---|
| Section | Robotics foundation models & agent stacks | Robotics foundation models & agent stacks |
| Tier | T1 | T1 |
| Type | Humanoid robot maker (Figure 01 / Figure 02 / Helix VLA model) | Robotics foundation model + simulation stack (NVIDIA) |
| Created | 2022 (founded) | 2024-03 (GTC announcement) |
| Latest release | Helix VLA (Feb 2025); Figure 03 in development | GR00T N1 (2B) on HuggingFace Mar-2025 |
| License | — | NVIDIA Open Model license |
| GitHub | — | github.com/NVIDIA/Isaac-GR00T |
| Pricing | — | OSS weights free; Cosmos / Jetson hardware sold separately |
| Funding | $675M Series B Feb-2024 (Microsoft, OpenAI, Bezos, NVIDIA, Intel, LG Innotek; $2.6B val); reported $39.5B val funding Q2 2025 | NVIDIA parent — public company, $3T+ mkt cap |
| Backend storage | Internal Figure cloud | Caller-managed (on-robot + cloud) |
| Deployment | Direct customer deployment (humanoid robot) | On-robot (Jetson Thor) + simulation (Isaac Sim) + training cloud (Cosmos) |
| API surface | No public API; closed humanoid platform | Python; Isaac Sim; HuggingFace weights |
| Multi-tenancy | Per-customer fleet | Per-developer / per-robot |
| Optimised for | End-to-end humanoid VLA — full-body + dexterous hand control | Full-stack robotics FM: model + sim + compute, multi-partner |
| Anti-fit | Closed humanoid platform — no developer API; alpha-customer-only | Requires NVIDIA hardware (Jetson Thor / GPUs); Isaac Sim learning curve |
Taxonomy
| Axis | Figure AI | NVIDIA GR00T / Isaac |
|---|---|---|
| storage | weight | weight |
| retrieval | parametric-recall | parametric-recall |
| persistence | parametric-permanent | parametric-permanent |
| update | agent-controlled | agent-controlled |
| unit | trajectory | trajectory |
| governance | opaque | opaque |
| conflict | training-time | training-time |
Pros & cons
Figure AI
Pros: Best-funded humanoid maker; Helix VLA is one of the highest-profile robotics-FM releases of 2025; BMW pilot.
Cons: Closed platform / weights; humanoid-only; production scale still small; valuation aggressive.
NVIDIA GR00T / Isaac
Pros: Full-stack robotics offering; major partner ecosystem; open-weights GR00T N1; NVIDIA distribution.
Cons: NVIDIA-hardware lock-in; Isaac Sim has steep learning curve; vendor sprawl across many SKUs.