Synthszr Charts — die großen AI-Marken im Wettkampf ums Podium

Marble 1.1

World Labs · v1.1 · depuis 2026-04-02 · 13× · vu le 30 juin 2026

Momentum

Marble 1.1 is the currently recommended default model of World Labs' generative world platform Marble, co-founded by Fei-Fei Li. It generates persistent, navigable 3D worlds from text, images, video, or 360° panoramas using Gaussian Splatting, and delivers improved lighting, contrast, and significantly fewer visual artifacts compared to Marble 1.0. Generated worlds can be exported as Gaussian splats (.spz/.ply), meshes, or videos, and integrated into engines such as Unreal Engine or Unity. The model is accessible via a web app as well as a public REST API.

Historique du momentum

19.05.17.08.

Fonctionnalités

Fine-Tuning	No dedicated fine-tuning; instead editing tools like Chisel, Expand, and Compose for post-hoc adjustments
Generation Time	approx. 2-5 minutes per world
Max Resolution	Export as Gaussian Splats: full resolution (2 million splats, .ply/.spz) or lightweight 500k version
Platform	Web app (marble.worldlabs.ai) and World API (api.worldlabs.ai)
Price	Marble 1.1: 1,500 credits/world (fixed); 1.1-Plus: 1,500 base + 0-1,500 variable (300/Dynamic Cube, up to 5)
Release Date	April 2, 2026

Plus de produits dans cette catégorie: Génération de mondes 3D

Preuves (13)

Company Analysis: World Labs

As of 04/07/2026

SELLSynthszr Vote

On a data-driven basis, the balance of evidence currently favors a cautious, slightly negative stance for non-insider investors. World Labs is a technically impressive, well-funded spatial-AI leader with a strong founder, a differentiated product (Marble), and rising mindshare, as reflected in growing search interest and a high-profile $1B Series C anchored by Autodesk and other strategic investors. However, the company remains early in commercialization with no disclosed revenue metrics, profitability path, or clear proof of large-scale enterprise adoption, while its valuation in the mid–single-digit billions already prices in substantial future success. At the same time, it faces intense competition from hyperscalers and other well-capitalized players pursuing similar world-model and physical-AI ambitions. Given this combination of rich private-market pricing, limited financial transparency, and elevated execution and competitive risks, the risk/reward skew appears unfavorable at current implied levels, supporting an overall SELL recommendation for investors who can exit, while acknowledging that high-risk, long-horizon capital could still view the name as an option on the world-model theme.

Key Takeaways

World Labs is a private spatial-intelligence AI company founded by Fei‑Fei Li that builds Large World Models (LWMs) to generate and reason about interactive 3D environments; its first commercial product, Marble, turns text, images, or video into persistent, explorable 3D worlds for creators, simulation, and robotics use cases. (worldlabs.ai)
The company has rapidly scaled its capital base: after earlier rounds that pushed valuation above $1B in 2024, it closed a $1B Series C on February 18, 2026, led by Autodesk with a $200M strategic investment and participation from major investors (a16z, NEA, Nvidia, AMD, Fidelity, Cisco, others), implying a valuation in the ~$5–7B range. (techcrunch.com)
World Labs is positioned as a category leader in “world models” and spatial AI, competing with Google DeepMind (Genie/Veo), Meta’s AMI Labs, Nvidia’s Cosmos platform, and OpenAI’s video/3D efforts; its perceived edge is focus and Fei‑Fei Li’s computer-vision pedigree, but it faces intense competition from far better-capitalized incumbents. (index.businessinsurance.com)
Commercial traction is still early-stage: Marble launched commercially in late 2025 and has since seen iterative upgrades (e.g., Marble 1.1), integrations into creative workflows (Autodesk partnership), and growing visibility in creator/enterprise communities, but there is no disclosed revenue, unit economics, or clear evidence yet of large-scale, recurring enterprise deployments. (time.com)
Secondary-market data suggest an implied valuation of roughly $6.8B (about $99 per share equivalent) with modest price appreciation year-to-date but flat over the last year, indicating that while investors recognize strategic potential, they are also cautious about execution risk, monetization, and the broader AI-cycle volatility. (stockanalysis.com)

Action Ideas

SELL

For existing secondary-market holders, the risk/reward at an implied ~$6–7B valuation looks stretched relative to the company’s still-unproven revenue model and heavy competitive pressure. The $1B Series C led by Autodesk and others priced in strong expectations for Marble’s adoption across gaming, VFX, design, and robotics, but there is no public visibility into ARR, margins, or retention, and the business remains effectively pre-scale. Meanwhile, hyperscalers (Alphabet, Meta, Nvidia, OpenAI) are aggressively investing in world models and physical AI, potentially compressing World Labs’ long-term pricing power and strategic independence. If secondary prices are near recent indications (~$99/share) and liquidity is available, de‑risking by selling part or all of a position appears prudent until there is clearer evidence of durable monetization and enterprise uptake.

Horizon: 12 mo.

BUY

For investors with access to primary or secondary shares at a meaningful discount to the latest implied valuation and a high risk tolerance, World Labs offers leveraged exposure to the emerging world-model and physical-AI theme. The company combines a top-tier technical founder (Fei‑Fei Li), strong research credentials, and a focused product (Marble) that directly targets high-value workflows in entertainment, design, simulation, and robotics. The $1B Series C and strategic partnership with Autodesk, plus backing from Nvidia, AMD, Cisco, and others, validate both the technology and its integration potential into existing 3D and industrial software stacks. If Marble becomes a de facto standard for 3D content generation or robotics simulation, upside from current private valuations could be substantial over a 5+ year horizon.

Horizon: 60 mo.

HOLD

For investors already committed and unable or unwilling to adjust exposure immediately, a neutral stance is warranted while monitoring key validation milestones. The company has secured blue-chip investors and partners, launched a differentiated product in Marble, and continues to iterate technically (e.g., Marble 1.1), but the lack of transparent financial metrics and the frothy funding environment make it difficult to underwrite risk-adjusted returns with confidence. Maintaining a position while tracking indicators such as enterprise contract wins, integration depth with Autodesk and other partners, and any disclosures on recurring revenue can help inform a more decisive BUY or SELL stance over the next 12–24 months.

Horizon: 24 mo.

Google Trends · ↗ rising

Proxy data from trend-tracking tools and newsflow suggest that global search interest for “World Labs” has risen sharply over the past two years, with a first step-up around its public unveiling and early funding in 2024, followed by a much steeper increase from late 2025 onward as Marble launched commercially and the company announced its $1B funding round and Autodesk partnership. Interest appears to have remained elevated through 2026 year-to-date, with only modest pullbacks between news cycles, consistent with a company moving from niche research topic to mainstream AI narrative. (axios.com)

Contrarian Insights

• Despite the hype around world models as the “next LLMs,” World Labs’ most durable moat may not be its core model performance but its workflow integrations and data flywheel with partners like Autodesk and robotics/industrial ecosystems. If Marble becomes deeply embedded in 3D design, CAD, and simulation pipelines, switching costs and proprietary usage data could matter more than raw model benchmarks, which competitors can eventually match. (techcrunch.com)
• While many investors frame World Labs primarily as a content-creation tool for gaming and media, the larger long-term value could come from robotics and physical AI, where accurate world models are critical for training and simulating embodied agents. In that scenario, World Labs might evolve into an infrastructure provider for robot learning and industrial autonomy rather than a pure creative-tool company, potentially justifying infrastructure-like multiples if it can secure key platform positions. (nvidianews.nvidia.com)