VLA / robotics foundation models: NVIDIA Cosmos + Figure Helix 2 — general-purpose robotics vs humanoid-specific VLA

3 sources

NVIDIA Cosmos (world foundation model for robotics + autonomous systems with Cosmos Reason vision-language-action model + Cosmos Transfer simulation-to-real pipeline) and Figure Helix 2 (vision-language-action model for humanoid robots with dual-system architecture — System 1 fast reactive + System 2 slow reasoning) are two VLA / robotics foundation model surfaces that ship in the same week. The two are not equivalent: Cosmos is the general-purpose robotics + autonomy stack; Helix 2 is the humanoid-specific VLA. Together they make VLA / robotics foundation models a creator-researchable surface: Cosmos for general robotics, Helix 2 for humanoid.

Why now

Both shipped in the same week — VLA / robotics foundation models are now a creator-researchable surface with general-purpose + humanoid-specific bets.

Why it is worth publishing

Demo potential: side-by-side of a Cosmos-driven simulation task and a Helix 2 humanoid task.

Evidence basis

Two independent vendor primary sources + The Decoder weekly roundup

“NVIDIA Cosmos and Figure Helix 2 both shipped this week — and together they make VLA / robotics foundation models a creator-researchable surface: Cosmos for general robotics, Helix 2 for humanoid-specific.”

Angle

Frame the two as 'general-purpose vs humanoid-specific' VLA surfaces — Cosmos for general robotics + autonomy, Helix 2 for humanoid.

Format

Long-form explainer

Demo idea

Record a 10-minute explainer: 3 min intro on 'general-purpose vs humanoid-specific' framing, 3 min on Cosmos (world foundation model + Cosmos Reason + Cosmos Transfer), 3 min on Helix 2 (dual-system VLA for humanoid), 1 min on the bet comparison.

Platform notes

Specific robotics benchmark numbers for Helix 2 and Cosmos beyond the captured summary are not extracted; confirm specific capability or benchmark against the underlying vendor docs before stating them on the record.

Usable claims

NVIDIA shipped Cosmos — world foundation model for robotics + autonomous systems, with the Cosmos Reason vision-language-action model and the Cosmos Transfer simulation-to-real pipeline.
Figure shipped Helix 2 — vision-language-action model for the Helix humanoid robot, with the dual-system architecture (System 1 fast reactive + System 2 slow reasoning).

Evidence pipeline

From the news

Breakdown

Cosmos is the general-purpose robotics + autonomy stack; Helix 2 is the humanoid-specific VLA. This explainer keeps the two robotics foundation model surfaces distinct, rather than treating 'robotics foundation model' as a single category.

Sources

Risks

Use The Decoder and IT之家 as media-type corroboration, but read the underlying vendor docs for any specific capability or benchmark claim before stating it on the record. Verify specific capability claim against the underlying vendor docs and the actual license / pricing matrix before stating it on the record; do not paraphrase per-platform pricing or license terms into specific dollar figures or commercial-use clauses.
Vendor docs confirm feature existence but specific benchmark numbers beyond the captured summary are not in this pass. Verify specific capability claim against the underlying vendor docs and the actual license / pricing matrix before stating it on the record; do not paraphrase per-platform pricing or license terms into specific dollar figures or commercial-use clauses.

Demo ideas

Side-by-side robotics task: Cosmos-driven simulation vs Helix 2 humanoid task.
Architecture matrix: plot Cosmos vs Helix 2 on general-purpose vs humanoid-specific, simulation-to-real vs dual-system.
Use-case decision: 'which VLA for which robot' (general-purpose → Cosmos, humanoid → Helix 2).