StartupVideo world modelUnited States
Spatial-intelligence company building generative 3D world models (Marble, Atlas) that it applies to real-to-sim environments for training and evaluating robot policies.
World Labs is a spatial-intelligence company founded by Fei-Fei Li, Justin Johnson, Ben Mildenhall and Christoph Lassner that came out of stealth in September 2024 with $230M in funding. It released the Marble 3D world model in November 2025 and the Atlas world model in September 2026, acquired robotics company SceniX in July 2026, and applies its models to real-to-sim-to-real robot training. On 2026-09-28 it signed a definitive agreement to join AMD, expected to close by the end of 2026.
Sources: worldlabs.ai (19 Oct 2026) · techcrunch.com (13 Sep 2024) · worldlabs.ai (12 Nov 2025) · worldlabs.ai (1 Sep 2026) · worldlabs.ai (21 Jul 2026) · worldlabs.ai (28 Jul 2026) · worldlabs.ai (28 Sep 2026)
Private company · Definitive agreement to join AMD signed 2026-09-28; expected to close by end of 2026 subject to regulatory approvals. · $1.23B disclosed in USD across 2 of 2 listed rounds
| Date | Round | Amount | Lead | Source |
|---|---|---|---|---|
| 13 Sep 2024 | Two early rounds (combined, announced at stealth exit)Valuation: over $1B | $230M | n/a | techcrunch.com, 13 Sep 2024 |
| 18 Feb 2026 | Funding round | $1B | n/a | worldlabs.ai, 18 Feb 2026 |
#69 of 104 model companies · score 14.7 / 100 · updated 19 Oct 2026
Each bar shows an input from 0 to 100% before weighting. 0 of 2 listed models have open weights.
The score (0–100) measures public, checkable evidence about a company's robot models. It does not measure private capability. A company that publishes no benchmark results and no weights can still have strong models. Every input is computed from the data on each company page, and every number there links to its source.
| Input | Weight | How it is measured |
|---|---|---|
| Published benchmarks | 35% | For each scored benchmark, the company's best result divided by the best result any company has published, averaged over the benchmarks it reports. No published results scores 0. Results come only from the model's own paper, model card or the official benchmark leaderboard. Scored: LIBERO (4-suite average success, %); SimplerEnv Google Robot, visual matching (%); SimplerEnv Google Robot, variant aggregation (%); SimplerEnv WidowX / Bridge (%); CALVIN ABC→D (average length, 0–5); RoboArena (pairwise rating). |
| Open weights | 20% | Share of the company's listed models whose weights the public can download. |
| Adoption | 20% | GitHub stars on the company's most-starred official model repository, on a log scale relative to the most-starred repository in the list. Hugging Face downloads are shown on company pages but not scored, because gated repositories report zero. |
| Released models | 15% | Number of publicly announced, named models, capped at 5. |
| Disclosed funding | 10% | The sum of the funding rounds listed on the company's profile that were disclosed in USD, on a log scale from $10M to the largest sum in the list. Rounds reported only in other currencies are listed but not converted or counted. This input applies only to private startups and robot makers. For big-tech divisions, research labs and listed companies it is n/a, and its weight is shared across the other inputs in proportion. |
Sources. arXiv papers and official model cards for benchmarks and parameters; official leaderboards (RoboArena); the GitHub and Hugging Face APIs for stars and downloads; company announcements and established press for funding. Missing values are shown as n/a, never estimated.
Updated 19 Oct 2026. Benchmark protocols differ between papers, for example in fine-tuning and number of trials, so small gaps between close scores are not meaningful.
Closed weightsWorld model
Multimodal autoregressive diffusion-transformer world model (text, image, video, 3D) that World Labs positions for real-to-sim robot navigation and manipulation, generating the RGB and depth a simulated robot would observe; in early access.
No benchmark results on a shared public benchmark have been published for this model.
Closed weightsWorld model
Generative 3D world model that builds editable worlds from text, images, video or coarse layouts and exports Gaussian splats or meshes; World Labs lists robotics simulation among its uses.
No benchmark results on a shared public benchmark have been published for this model.
