World Model Startups contains 22 companies across
5 categories.
This landscape covers "world models" a.k.a startups building learned representations of real or virtual systems that can predict how they evolve, ideally in response to actions. It excludes products that simulate or generate environments without learning their underlying dynamics.
General & Spatial Worlds
Models designed to understand and generate interactive environments that are not limited to one industry or physical system.
AMI Labs (Advanced Machine Intelligence)
Website: amilabs.xyz
AMI Labs is a frontier AI research lab building world models and other AI systems that can understand the real world, maintain persistent memory, reason and plan, and remain controllable and safe.
Category: General & Spatial Worlds.
Funding stage: Seed.
Founded: 2026.
Country: France.
Employees: 11-50 employees.
Landscape fit: It fits the World Model Startups category because its core product direction is explicitly world models that learn from real-world sensor data and predict consequences for action in interactive environments.
Funding rounds
-
seed
announced 2026-03-01
: 1030000000 USD
; investors: Cathay Innovation, Greycroft, Hiro Capital, HV Capital, Bezos Expeditions, Toyota Ventures, New Legacy Ventures, Temasek, SBVA, NVIDIA, Mark Cuban, Association Familiale Mulliez, Groupe industriel Marcel Dassault, Sea, Alpha Intelligence Capital, Eric Schmidt, Aglaé Lab, ZEBOX Ventures, Artémis, Xavier Niel, Publicis Groupe, Samsung, Bpifrance Digital Venture, Jim Breyer, Tim & Rosemary Berners-Lee, Mark Leslie
; source: AMI Labs - Updates
; https://amilabs.xyz/updates
Company timeline
-
2026-07-16:
Begin partner scouting in Asia for real-world world-model applications
(gtm)
. In a July interview, AMI Labs said it was scouting industrial partners, global companies, and researchers in Korea and Asia, and that it needs real-world access to train and validate its models outside the lab.
Source: TechCrunch.
https://techcrunch.com/2026/07/16/why-ami-labs-alexandre-lebrun-wont-call-his-ai-agi-or-superintelligence/
-
2026-03-10:
Official launch with public hiring call
(hr)
. AMI Labs publicly launched and invited candidates to join the team, pointing to an active early hiring phase.. Team size observed: 11-50 employees employees.
Source: AMI Labs Updates.
https://amilabs.xyz/updates
-
2026-03-10:
Official launch with world-model positioning and $1.03B seed round
(gtm)
. AMI Labs publicly launched as a frontier AI research lab focused on world models, persistent memory, planning, and controllable safe systems, while announcing a $1.03B seed round and a globally distributed operating footprint.
Source: AMI Labs - Updates.
https://amilabs.xyz/updates
-
2026-03-10:
Official launch of AMI Labs
(product)
. AMI Labs publicly launched and introduced its mission to build world-model-based AI systems with persistent memory, reasoning, planning, and safety.
Source: AMI Labs - Updates.
https://amilabs.xyz/updates
-
2026-03-01:
seed
(funding)
. Amount: 1030000000 USD. Investors: Cathay Innovation, Greycroft, Hiro Capital, HV Capital, Bezos Expeditions, Toyota Ventures, New Legacy Ventures, Temasek, SBVA, NVIDIA, Mark Cuban, Association Familiale Mulliez, Groupe industriel Marcel Dassault, Sea, Alpha Intelligence Capital, Eric Schmidt, Aglaé Lab, ZEBOX Ventures, Artémis, Xavier Niel, Publicis Groupe, Samsung, Bpifrance Digital Venture, Jim Breyer, Tim & Rosemary Berners-Lee, Mark Leslie
Source: AMI Labs - Updates.
https://amilabs.xyz/updates
World Labs
Website: worldlabs.ai
A frontier AI research and product company building spatial intelligence world models that can perceive, generate, reason about, and interact with 3D environments, including its Marble product for creating persistent 3D worlds from text, images, or video. ([worldlabs.ai](https://www.worldlabs.ai/about))
Category: General & Spatial Worlds.
Funding stage: Series C.
Founded: 2024.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits General & Spatial Worlds because its core product is a world model for generating and interacting with explorable 3D environments, not a narrow vertical application. ([worldlabs.ai](https://www.worldlabs.ai/about))
Funding rounds
-
Series C
announced 2026-02-01
: 1000000000 USD
; investors: AMD, Autodesk, Emerson Collective, Fidelity Management & Research Company, NVIDIA, Sea
; source: World Labs
; https://www.worldlabs.ai/blog/funding-2026
-
Series A extension
announced 2024-09-01
: 130000000 USD
; investors: Andreessen Horowitz, NEA, Radical Ventures
; source: World Labs
; https://www.worldlabs.ai/blog/funding-2026
Company timeline
-
2026-07-21:
World Labs acquired SceniX
(acquisition)
. World Labs announced it was acquiring SceniX, a robotics company, to advance spatial intelligence for robotics.. Acquirer: World Labs. Status: announced
Source: World Labs.
https://www.worldlabs.ai/blog/scenix
-
2026-07-01:
Team size observed
(hr)
. Public source shows a team size of 11-50 employees employees.. Team size observed: 11-50 employees employees.
Source: World Labs | LinkedIn.
https://www.linkedin.com/company/world-labs/
-
2026-06-03:
World Labs deepens the world-model narrative
(gtm)
. A June taxonomy essay decomposed world models into renderers, simulators, and planners, reinforcing World Labs’ category leadership and technical authority.
Source: World Labs.
https://www.worldlabs.ai/blog/taxonomy-of-world-models
-
2026-04-14:
Spark 2.0 streaming and LoD backend
(product)
. World Labs described Spark 2.0, adding a Level-of-Detail system and streaming .RAD format for large 3DGS worlds on the web.
Source: World Labs blog.
https://www.worldlabs.ai/blog/spark-2.0
-
2026-04-14:
Spark 2.0 extends World Labs' web rendering motion
(gtm)
. World Labs published Spark 2.0, adding streamable level-of-detail rendering for very large 3D Gaussian splat worlds on the web and open-sourcing the renderer.
Source: World Labs.
https://www.worldlabs.ai/blog/spark-2.0
-
2026-04-02:
Marble 1.1 and 1.1 Plus model rollout
(product)
. World Labs added Marble 1.1 and Marble 1.1 Plus, surfaced model selection in Create, and exposed create-and-edit and asset visibility improvements.
Source: World Labs docs.
https://docs.worldlabs.ai/marble/release-notes
Odyssey
Website: odyssey.ml
Odyssey is an AI lab building general-purpose world models that understand, simulate, and generate interactive worlds in real time.
Category: General & Spatial Worlds.
Funding stage: Series B.
Founded: 2023.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits the General & Spatial Worlds category because its core product is a general world model for simulating interactive environments rather than a domain-specific vertical model.
Funding rounds
-
Series B
announced 2026-06-01
: 310000000 USD
; investors: Natural Capital, Amazon, GV, AMD Ventures, EQT, IQT
; source: Odyssey
; https://odyssey.ml/our-series-b
Company timeline
-
2026-07-01:
Team size observed
(hr)
. Public source shows a team size of 11-50 employees employees.. Team size observed: 11-50 employees employees.
Source: LinkedIn.
https://www.linkedin.com/company/odysseyml
-
2026-06-01:
Series B
(funding)
. Amount: 310000000 USD. Investors: Natural Capital, Amazon, GV, AMD Ventures, EQT, IQT
Source: Odyssey.
https://odyssey.ml/our-series-b
-
2026-05-18:
Starchild-1 and Agora-1 broadened the product surface
(gtm)
. Odyssey added a first real-time multimodal world model and a multi-agent world model, extending the platform from single-stream interactive video toward richer multimodal and multi-participant simulations. ([odyssey.ml](https://odyssey.ml/writing))
Source: Odyssey.
https://odyssey.ml/writing
-
2026-05-18:
Agora-1 launched multi-agent world simulation
(product)
. Odyssey released Agora-1, a real-time world model that lets up to four human or AI participants interact within the same generated simulation.
Source: Agora-1: The Multi-Agent World Model.
https://odyssey.ml/introducing-agora-1
-
2026-05-17:
Starchild-1 added real-time multimodal audio-video generation
(product)
. Odyssey previewed Starchild-1, its first multimodal world model, which generates synchronized audio and video in real time while responding to streaming user input.
Source: Starchild-1: The First Real-Time Multimodal World Model.
https://odyssey.ml/introducing-starchild-1
-
2026-05-12:
PROWL-1 introduced adversarial world-model training
(product)
. Odyssey released PROWL-1, an RL-driven framework where an agent searches for failures in world models and turns them into training curriculum.
Source: Introducing PROWL-1: Learning Through Discovery.
https://odyssey.ml/introducing-prowl-1
Decart
Website: decart.ai
Decart is a live AI lab that builds infrastructure and models for real-time, interactive AI experiences, including world models, live video transformation, and playable AI-generated environments. ([decart.ai](https://decart.ai/about))
Category: General & Spatial Worlds.
Funding stage: Series C.
Founded: 2023.
Country: Israel.
Employees: 51-200 employees.
Landscape fit: It fits the World Model Startups / General & Spatial Worlds category because Decart explicitly builds 'world models' and responsive AI-generated environments that can transform video and create playable real-time worlds. ([decart.ai](https://decart.ai/about))
Funding rounds
-
series C
announced 2026-05-01
: 300000000 USD
; investors: Radical Ventures, Adobe Ventures, Andrej Karpathy, Atreides Management, Benchmark, Michael Eisner, Moritz Baier-Lentz, NVIDIA, Sequoia Capital, Toyota Ventures, Valor Equity Partners, Zeev Ventures, eBay Ventures
; source: Decart
; https://decart.ai/blog
-
Series B
announced 2025-08-06
: 100000000 USD
; investors: Sequoia Capital, Benchmark, Zeev Ventures, Aleph VC
; source: fortune.com
; https://fortune.com/2025/08/07/exclusive-decart-raises-100-million-at-a-3-1-billion-valuation-chasing-the-future-of-real-time-creative-ai/
-
Series A
announced 2024-12-19
: 32000000 USD
; investors: Benchmark, Sequoia, Zeev Ventures
; source: TechCrunch
; https://techcrunch.com/2024/12/19/decart-adds-another-32m-at-a-500m-valuation/
-
seed
announced 2024-10-01
: 21000000 USD
; investors: Sequoia Capital, Zeev Ventures
; source: Decart
; https://www.decart.ai/articles/sequoia-backed-decart-raises-21m-in-seed-funding
Company timeline
-
2026-07-16:
Lucy 2.5 rollout across realtime and batch surfaces
(gtm)
. Decart released Lucy 2.5 as its latest-generation video editing model and made it the default alias for lucy-latest across realtime and batch APIs.
Source: Decart blog.
https://decart.ai/blog
-
2026-07-16:
Lucy 2.5 latest-generation release
(product)
. Decart shipped Lucy 2.5 as the latest-generation video editing model with higher-quality edits, physically-aware VFX, stronger temporal consistency, and lower latency.
Source: Decart AI Lab | Resources.
https://decart.ai/blog
-
2026-07-01:
Team size observed
(hr)
. Public source shows a team size of 51-200 employees employees.. Team size observed: 51-200 employees employees.
Source: LinkedIn.
https://www.linkedin.com/company/decart-ai
-
2026-06-10:
Oasis 3 preview for physical AI
(gtm)
. Decart introduced Oasis 3 as a realtime interactive world model for physical AI, with driving-simulator and reinforcement-learning framing.
Source: Decart blog.
https://decart.ai/blog
-
2026-06-01:
Oasis 3 preview and realtime quality upgrades
(product)
. Decart introduced Oasis 3 Preview as a promptable realtime world model, added Lucy VTON 3, moved Python realtime SDKs to LiveKit, and added enterprise SSO plus better connection-quality tooling.
Source: Decart API Platform Changelog.
https://docs.platform.decart.ai/changelog
-
2026-05-18:
DOS and enterprise AI infrastructure repositioning
(gtm)
. Decart publicly reframed itself as an infrastructure company with DOS, Lucy, and Oasis as separate product lines, and said the funding announcement included DOS 2.0 plus upcoming model releases.
Source: Decart blog.
https://decart.ai/blog
General Intuition
Website: generalintuition.com
General Intuition is an AI research lab building world models and action models that learn from gameplay data to perceive, predict, and act in virtual and physical environments.
Category: General & Spatial Worlds.
Funding stage: Series A.
Founded: 2025.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits the General & Spatial Worlds category because its core product is a world-model approach to spatial and temporal reasoning, trained on interactive environments and game data rather than a single industry workflow.
Funding rounds
-
series A
announced 2026-06-01
: 320000000 USD
; investors: Khosla Ventures, General Catalyst, Hedosophia, Bezos Expeditions, Innovation Endeavors, Nico Rosberg
; source: Axios
; https://www.axios.com/2026/06/26/general-intuition-ai-gaming
-
seed
announced 2025-10-01
: 133700000 USD
; investors: Khosla Ventures, General Catalyst, Raine
; source: TechCrunch
; https://techcrunch.com/2025/10/16/general-intuition-lands-134m-seed-to-teach-agents-spatial-reasoning-using-video-game-clips/
Company timeline
-
2026-07-01:
Team size observed
(hr)
. Public source shows a team size of 11-50 employees employees.. Team size observed: 11-50 employees employees.
Source: LinkedIn.
https://www.linkedin.com/company/generalintuition
-
2026-06-01:
series A
(funding)
. Amount: 320000000 USD. Investors: Khosla Ventures, General Catalyst, Hedosophia, Bezos Expeditions, Innovation Endeavors, Nico Rosberg
Source: Axios.
https://www.axios.com/2026/06/26/general-intuition-ai-gaming
-
2025-10-16:
General Intuition launches as a gameplay-data AI lab with a seed round
(gtm)
. The company publicly introduced itself as a frontier research lab using Medal gameplay data to train foundation models and general agents for spatial-temporal reasoning, alongside its $133.7M seed financing.
Source: TechCrunch.
https://techcrunch.com/2025/10/16/general-intuition-lands-134m-seed-to-teach-agents-spatial-reasoning-using-video-game-clips/
-
2025-10-01:
seed
(funding)
. Amount: 133700000 USD. Investors: Khosla Ventures, General Catalyst, Raine
Source: TechCrunch.
https://techcrunch.com/2025/10/16/general-intuition-lands-134m-seed-to-teach-agents-spatial-reasoning-using-video-game-clips/
Runway
Website: runway.com
Runway is an AI company building real-world intelligence: multimodal models and products that can understand, simulate, generate, and act in the world, including creative, developer, and robotics tools.
Category: General & Spatial Worlds.
Funding stage: Series D+.
Founded: 2018.
Country: United States.
Landscape fit: It fits the General & Spatial Worlds category because Runway explicitly says it is building general world models and real-world intelligence that can simulate the world and support robotics, media, and developer use cases.
Funding rounds
-
series_e
announced 2026-02-01
: 315000000 USD
; investors: General Atlantic, NVIDIA, Adobe Ventures, AllianceBernstein, AMD Ventures, Fidelity Management & Research Company, Mirae Asset, Emphatic Capital, Felicis, Premji Invest
; source: Runway
; https://runway.com/news/runway-series-e-funding
-
series_d
announced 2025-04-01
: 300000000 USD
; investors: General Atlantic, Fidelity Management & Research Company, Baillie Gifford, NVIDIA, SoftBank Vision Fund 2
; source: Runway
; https://runway.com/news/runway-series-d-funding
-
series_c_extension
announced 2023-06-01
: 141000000 USD
; investors: Google, NVIDIA, Salesforce Ventures, Amplify Partners, Lux Capital, Coatue, Compound
; source: Runway
; https://runway.com/news/runway-raises-141-million-to-continue-building-the-future-of-creativity
-
series_c
announced 2022-12-01
: 50000000 USD
; investors: Felicis, Amplify Partners, Lux Capital, Coatue, Compound, Madrona
; source: Runway
; https://runway.com/news/runway-raises-50-series-c
-
series_b
announced 2021-12-01
: 35000000 USD
; investors: Coatue, Amplify Partners, Lux Ventures, Compound
; source: Runway
; https://runway.com/news/runway-raises-35-series-b
Company timeline
-
2026-07-23:
Runway Media Router launches inside Runway Dev
(gtm)
. Runway introduced a preference-optimized model router that automatically selects video, image, or audio models inside Runway Dev.
Source: Introducing Runway Media Router.
https://runway.com/news/company-news/introducing-runway-media-router
-
2026-07-20:
Runway publishes AI Media Report
(gtm)
. Runway released a longitudinal enterprise report quantifying cost and time savings and showing deeper adoption across broadcasters, agencies, retailers, and other enterprises.
Source: The AI Media Report: Cost, Speed and What Comes Next.
https://runway.com/news/company-news/ai-media-report
-
2026-07-08:
Runway Dev launches for developers and enterprise teams
(product)
. Runway launched Runway Dev, an AI media platform with one API for image, video, audio, and real-time character models. ([runway.com](https://runway.com/news/company-news/introducing-runway-dev))
Source: Runway News.
https://runway.com/news/introducing-runway-dev
-
2026-07-08:
Runway launches Runway Dev
(gtm)
. Runway introduced Runway Dev as a unified AI media platform for developers and enterprise teams with image, video, audio, and real-time character models.
Source: Introducing Runway Dev.
https://runway.com/news/company-news/introducing-runway-dev
-
2026-07-01:
Runway and Bertelsmann expand their partnership
(gtm)
. Runway announced a creative partnership with Bertelsmann to integrate Runway's models across RTL Group, BMG, and Bertelsmann Marketing Services.
Source: Runway Announces Creative Partnership with Bertelsmann.
https://runway.com/news/runway-announces-creative-partnership-with-bertelsmann
-
2026-06-29:
Japan expansion included a Head of Japan hire
(hr)
. Runway said its recent Japan expansion included opening a Tokyo office as Japan headquarters and hiring a Head of Japan to lead the effort.
Source: Runway News.
https://runway.com/news/runway-and-mixi-announce-strategic-partnership
Luma
Website: lumalabs.ai
Luma is a private AI company building multimodal generative models and creative tools for video, image, and 3D content, with an emphasis on understanding and operating in the physical world.
Category: General & Spatial Worlds.
Funding stage: Series C.
Founded: 2021.
Country: United States.
Employees: 150+ employees.
Landscape fit: It fits the World Model Startups category because Luma’s products and mission center on multimodal models that can generate, understand, and interact with real-world-like scenes, including video, 3D, and physical-world reasoning.
Funding rounds
-
new funding round
announced 2025-11-19
: 900000000 USD
; investors: Humain, Advanced Micro Devices' venture arm, Andreessen Horowitz, Amplify Partners, Matrix Partners
; source: www.cnbc.com
; https://www.cnbc.com/2025/11/19/luma-ai-raises-900-million-in-funding-led-by-saudi-ai-firm-humain.html
-
Series B
announced 2024-01-09
: 43000000 USD
; investors: Andreessen Horowitz
; source: techcrunch.com
; https://techcrunch.com/2024/01/09/luma-raises-43m-to-build-ai-that-crafts-3d-models/
-
Series A
announced 2023-03-31
: 20000000 USD
; investors: Amplify Partners, General Catalyst, NVIDIA Ventures, Matrix Partners, South Park Commons
; source: www.contextvc.com
; https://www.contextvc.com/essay-blog/congratulations-to-luma-ai-for-raising-their-20-million-series-a
Company timeline
-
2026-06-16:
Luma launches Luma Skills for repeatable creative workflows
(gtm)
. Luma introduced Skills as a way to build a creative workflow once and reuse it repeatedly, reinforcing a packaged workflow automation direction.
Source: Luma News.
https://lumalabs.ai/news
-
2026-06-16:
Luma Skills launch
(product)
. Luma introduced Skills, a repeatable workflow system inside Luma Agents that lets users build a workflow once and run it repeatedly on new assets.
Source: Luma.
https://lumalabs.ai/news/luma-skills
-
2026-06-09:
Luma launches Ray3.2 as a model and API for professional video workflows
(gtm)
. Luma released Ray3.2 with frame-level control, longer HDR outputs, and first-time API access for integration into enterprise products and custom workflows.
Source: Luma News.
https://lumalabs.ai/news/introducing-ray-3-2
-
2026-06-09:
Ray3.2 launch
(product)
. Luma released Ray3.2 with frame-level control, stronger professional workflow support, and API availability for video generation.
Source: Luma.
https://lumalabs.ai/news/introducing-ray-3-2
-
2026-05-05:
Luma launches Uni-1.1 API for image generation and editing
(gtm)
. Luma opened access to the Uni-1.1 API and framed it as a developer surface for production creative workflows, with usage tiers for Build and Scale.
Source: Luma News.
https://lumalabs.ai/news/uni-1-1-api
-
2026-05-05:
Uni-1.1 API launch
(product)
. Luma opened the Uni-1.1 API for image generation and natural-language editing, extending its developer platform beyond video into controllable image workflows.
Source: Luma.
https://lumalabs.ai/news/uni-1-1-api
SpAItial
Website: spaitial.ai
SpAItial builds physically grounded world models that generate and reason about 3D environments, turning text, images, and panoramas into interactive spatial scenes.
Category: General & Spatial Worlds.
Funding stage: Seed.
Founded: 2024.
Country: United Kingdom.
Employees: 2-10 employees (LinkedIn).
Landscape fit: It fits the World Model Startups / General & Spatial Worlds category because its core product is a world model that creates and understands explorable 3D environments with physics-aware spatial reasoning.
Funding rounds
-
seed
announced 2025-05-01
: 13000000 USD
; investors: Earlybird Venture Capital, Speedinvest, angel investors
; source: SpAItial Blog
; https://spaitial.ai/blog/announcing-spaitial
Company timeline
-
2026-07-01:
Team size observed
(hr)
. Public source shows a team size of 2-10 employees (LinkedIn) employees.. Team size observed: 2-10 employees (LinkedIn) employees.
Source: LinkedIn.
https://uk.linkedin.com/company/spaitial-ai
-
2026-06-01:
Startup program introduced for spatial AI builders
(product)
. SpAItial launched a startup program offering credits, priority API access, early Echo models, and feedback loops for startups building with spatial AI.
Source: SpAItial homepage.
https://spaitial.ai/
-
2026-05-01:
May 2026 platform expansion: Echo HQ and developer API
(product)
. SpAItial surfaced a higher-quality Echo checkpoint and a programmable developer API for generating persistent 3DGS worlds from external tools and pipelines.
Source: SpAItial homepage.
https://spaitial.ai/
-
2026-05-01:
Developer API and startup program introduced a repeatable distribution motion
(gtm)
. SpAItial added programmable API access, developer docs, and a startup program with credits and priority access to seed adoption among builders and startups.
Source: SpAItial homepage.
https://spaitial.ai/
-
2026-04-28:
Echo-2 launched with broader real-time 3D world generation
(product)
. SpAItial released Echo-2, a higher-fidelity frontier model for physically grounded 3D world generation from text or images, with navigable outputs and broader use cases.
Source: SpAItial Blog.
https://spaitial.ai/blog/echo-2-release
-
2026-04-28:
Echo-2 expanded the product toward physically grounded world generation
(gtm)
. SpAItial released Echo-2 as a higher-quality model and framed it around physically grounded 3D world generation, with downstream applications in remodeling, digital twins, and robotics training.
Source: SpAItial blog.
https://spaitial.ai/blog/echo-2-release
Overworld
Website: over.world
Overworld is an AI research and development studio building real-time, local-first world models that generate interactive, playable environments you can actually control and explore.
Category: General & Spatial Worlds.
Funding stage: Pre Seed.
Founded: 2025.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits the General & Spatial Worlds category because its core product is a world model that maintains persistent state and generates interactive environments in real time, rather than static content or a single-industry workflow.
Company timeline
-
2026-07-01:
Team size observed
(hr)
. Public source shows a team size of 11-50 employees employees.. Team size observed: 11-50 employees employees.
Source: LinkedIn.
https://www.linkedin.com/company/over-world-ai
-
2026-04-09:
Waypoint-1.5 expanded local execution and browser access
(gtm)
. Overworld announced Waypoint-1.5, highlighting far more training data, local execution on consumer hardware, an EXE installer, and instant browser play via Overworld Stream.
Source: Overworld Blog.
https://over.world/blog/waypoint-1-5
-
2026-04-09:
Waypoint-1.5 expands local execution and streaming access
(product)
. Overworld released Waypoint-1.5 with higher visual quality, up to 720p60 output, two hardware tiers, a simpler EXE installer, local Biome runtime support, and browser play via Overworld Stream.
Source: Overworld blog.
https://over.world/blog/waypoint-1-5
-
2026-03-13:
Biome desktop client introduced for local play
(gtm)
. Overworld launched Biome as an open-source desktop client for running Waypoint locally on a user's own GPU, and the site added a download path plus browser-streaming fallback.
Source: Overworld Blog / Download Biome page.
https://over.world/blog/raising-a-biome-the-trials-and-tribulations-of-waypoint-at-home
-
2026-03-13:
Biome desktop client introduced
(product)
. Overworld launched Biome as an open-source desktop client for running Waypoint locally on a user’s own GPU.
Source: Overworld blog.
https://over.world/blog/raising-a-biome-the-trials-and-tribulations-of-waypoint-at-home
-
2026-02-20:
Prompt-to-worlds pipeline became product surface
(gtm)
. Overworld described its prompting pipeline as converting user input into structured signals and said it had become the other half of the product.
Source: Overworld Blog.
https://over.world/blog/transforming-prompt-to-worlds
Robotics & Embodied Intelligence
Models that help robots anticipate how their surroundings will change as they act within them.
One Robot
Website: onerobot.io
One Robot builds world-model-based simulation environments for robotics teams so they can train, evaluate, and debug manipulation policies with more realistic visuals and physics before deploying on real robots. ([onerobot.io](https://www.onerobot.io/))
Category: Robotics & Embodied Intelligence.
Founded: 2025.
Country: United States.
Employees: 2-10 employees.
Landscape fit: It fits Robotics & Embodied Intelligence because its product helps robots anticipate how their environment will change under action by learning physics/visual dynamics and generating realistic rollouts for hard manipulation tasks. ([onerobot.io](https://www.onerobot.io/))
Funding rounds
-
Accelerator/Incubator
announced 2026-02-27
: 500000 USD
; investors: Y Combinator
; source: Dealroom.co
; https://app.dealroom.co/companies/one_robot
Company timeline
-
2026-02-27:
Accelerator/Incubator
(funding)
. Amount: 500000 USD. Investors: Y Combinator
Source: Dealroom.co.
https://app.dealroom.co/companies/one_robot
Fern
Website: fern.bot
Fern builds high-fidelity, custom robot world models and RL environments that let robotics teams evaluate and train policies in software instead of on physical hardware.
Category: Robotics & Embodied Intelligence.
Funding stage: Pre Seed.
Founded: 2025.
Country: United States.
Employees: 2-10 employees.
Landscape fit: It fits the World Model Startups category because its core product is an action-conditioned learned world model that predicts future robot-observed frames to enable evaluation and training of robot policies.
Company timeline
-
2026-07-01:
Case study shows RL trained entirely in simulation and deployed on real robots
(product)
. Fern published a case study reporting that RL was trained entirely in Fern’s simulator and then deployed on real robots with improved throughput and task success.
Source: Making robots faster and more successful with RL trained entirely in simulation.
https://www.fern.bot/case-studies/rl-in-the-world-model
-
2026-07-01:
Case studies index highlights RL-in-simulation and closed-loop evaluation results
(product)
. Fern’s case-studies index lists July 2026 RL-in-simulation results and June 2026 closed-loop evaluation results, surfacing multiple proof points for the same core simulator workflow.
Source: Customer Case Studies.
https://www.fern.bot/case-studies
-
2026-06-01:
Case study shows closed-loop evaluation matching real-robot traces
(product)
. Fern published a case study on closing the loop for a vertical robotics company, describing policy evaluation end-to-end in Fern’s simulator with joint-level traces closely matching the real robot.
Source: Closing the loop for a vertical robotics company.
https://www.fern.bot/case-studies/closing-the-loop
-
2026-06-01:
Whitepaper introduces learned simulators and planned public benchmarks
(product)
. Fern published a whitepaper on learned, action-conditioned world models for evaluating general robot policies, including 256×256 RGB rollouts and a plan for public benchmarks and RL environments.
Source: Learned simulators for evaluating general robot policies.
https://www.fern.bot/whitepaper
-
2026-01-01:
Homepage positions Fern around custom robot world models for policy evaluation and training
(product)
. Fern’s homepage frames the product as high-fidelity custom robot world models for scaling robot policy evaluation and training in software instead of on hardware.
Source: Fern home.
https://www.fern.bot
Saturn Dynamics
Website: saturndynamics.com
Saturn Dynamics builds world models for embodied intelligence, using physics-aware generative simulation to help train and validate autonomous systems.
Category: Robotics & Embodied Intelligence.
Funding stage: Seed.
Founded: 2025.
Country: United Kingdom.
Landscape fit: It fits World Models / Robotics & Embodied Intelligence because its core product is a generative world model and simulation infrastructure for autonomy training, synthetic data, and validation in real-world physical systems.
Company timeline
-
2025-07-29:
Founding leadership appointed
(hr)
. UK company filings show Saturn Dynamics LTD appointed multiple officers/directors on the same date, marking the company’s initial leadership setup.
Source: GOV.UK Companies House.
https://find-and-update.company-information.service.gov.uk/company/16612741/officers
Stanhope AI
Website: stanhopeai.com
Stanhope AI is a London-based deep-tech startup building active-inference AI software that lets robots and embodied systems make autonomous decisions in the real world, including in drone and robotics use cases.
Category: Robotics & Embodied Intelligence.
Funding stage: Seed.
Founded: 2021.
Country: United Kingdom.
Landscape fit: Builds active-inference world models that update from sensor observations and support embodied decision-making at the edge.
Funding rounds
-
seed
announced 2026-02-01
: 8000000 USD
; investors: Frontline Ventures, Paladin Capital Group, Auxxo Female Catalyst Fund, UCL Technology Fund, MMC Ventures
; source: Stanhope AI press release
; https://www.stanhopeai.com/press-releases/stanhope-ai-raises-8m-as-real-world-ai-model-gains-momentum-beyond-llms
-
seed
announced 2024-03-20
: 2300000 GBP
; investors: UCL Technology Fund, Creator Fund, MMC Ventures, Moonfire Ventures, Rockmount Capital, angel investors
; source: Stanhope AI press release
; https://www.stanhopeai.com/press-releases/stanhope-ai-raises-23-million-for-ai-inspired-by-neuroscience-that-teaches-machines-to-make-truly-human-like-decisions-for-the-first-time-ever
Company timeline
-
2026-02-01:
seed
(funding)
. Amount: 8000000 USD. Investors: Frontline Ventures, Paladin Capital Group, Auxxo Female Catalyst Fund, UCL Technology Fund, MMC Ventures
Source: Stanhope AI press release.
https://www.stanhopeai.com/press-releases/stanhope-ai-raises-8m-as-real-world-ai-model-gains-momentum-beyond-llms
-
2025-12-01:
Careers page adds Chief of Staff and AI Developer openings
(hr)
. Stanhope AI’s careers page listed two full-time London roles: Chief of Staff and AI Developer – Robotics & Embodied Active Inference Systems, both with December 2025 commencement, plus a junior internship programme.
Source: Stanhope AI Careers.
https://www.stanhopeai.com/careers
-
2024-09-12:
First Active Inference-controlled AI agent demo at Drone Exhibition Day
(product)
. Stanhope AI held an Active Inference Drone Exhibition Day and showed its first AI agent controlled by Active Inference in a drone context.
Source: Events.
https://www.stanhopeai.com/events
-
2024-03-20:
seed
(funding)
. Amount: 2300000 GBP. Investors: UCL Technology Fund, Creator Fund, MMC Ventures, Moonfire Ventures, Rockmount Capital, angel investors
Source: Stanhope AI press release.
https://www.stanhopeai.com/press-releases/stanhope-ai-raises-23-million-for-ai-inspired-by-neuroscience-that-teaches-machines-to-make-truly-human-like-decisions-for-the-first-time-ever
1X
Website: 1x.tech
AI and robotics company building general-purpose humanoid robots, led by NEO for home use and EVE for industrial deployment. ([1x.tech](https://www.1x.tech/about))
Category: Robotics & Embodied Intelligence.
Founded: 2014.
Country: Norway.
Landscape fit: It fits world models / embodied intelligence because 1X is explicitly building embodied AI for humanoids and launched a World Model Lab to pretrain models that improve autonomous robot behavior in real environments. ([1x.tech](https://www.1x.tech/ja_jp/discover/1x-world-model-lab))
Funding rounds
-
Series B
announced 2024-01-13
: 100000000 USD
; investors: EQT Ventures, Samsung NEXT, Skagerak Capital, Nistad group
; source: www.1x.tech
; https://www.1x.tech/ja_jp/discover/1x-secures-100m-in-series-b-funding
-
Series A2
announced 2023-03-23
: 23500000 USD
; investors: OpenAI Startup Fund, Tiger Global, Sandwater, Alliance Ventures, Thursday Ventures, Skagerak Capital, Trevor Blackwell, Type One
; source: 1X
; https://www.1x.tech/discover/1x-rasies-23-5m-in-series-a2-funding-led-by-open-ai
Company timeline
-
2026-07-09:
NEO’s Hands announced
(product)
. 1X announced 25-DOF tendon-driven hands for the NEO humanoid platform and described them as the robot’s API to the physical world.
Source: NEO’s Hands | An API to the Physical World.
https://www.1x.tech/discover/neos-hands
-
2026-07-01:
Announced Bill Nash as CFO
(hr)
. 1X announced Bill Nash as its Chief Financial Officer, adding senior finance leadership. ([1x.tech](https://www.1x.tech/discover/bill-nash-cfo?utm_source=openai))
Source: 1X.
https://www.1x.tech/discover/bill-nash-cfo
-
2026-06-01:
Announced Tom Sanocki as VP of Engineering
(hr)
. 1X announced Tom Sanocki as its new VP of Engineering. ([1x.tech](https://www.1x.tech/discover/tom-sanocki-vp-engineering?utm_source=openai))
Source: 1X.
https://www.1x.tech/discover/tom-sanocki-vp-engineering
-
2026-04-30:
Hayward NEO factory
(product)
. 1X detailed the Hayward NEO factory and its vertical manufacturing and scaling plans for home humanoid production.
Source: NEO Factory | Building Your NEO.
https://www.1x.tech/discover/neo-factory
-
2026-04-01:
Announced Amy Lentz as SVP of People
(hr)
. 1X announced Amy Lentz as its new SVP of People, expanding senior leadership in people operations. ([1x.tech](https://www.1x.tech/discover/amy-lentz-svp-people?utm_source=openai))
Source: 1X.
https://www.1x.tech/discover/amy-lentz-svp-people
-
2026-03-17:
NVIDIA robotics platform work for NEO
(product)
. 1X described using NVIDIA tools for simulation, training, onboard inference, and deployment for NEO.
Source: Developing NEO’s AI Using NVIDIA’s Robotics Platform.
https://www.1x.tech/discover/nvidia-gtc-2026
Earth, Energy & Industrial Systems
Models that learn how large physical systems evolve and use that understanding to support better operational decisions.
Jua.ai AG
Website: jua.ai
Jua builds a foundation model for the physical world, starting with weather and Earth-system simulation, to improve decision-making in energy and other climate-sensitive industries.
Category: Earth, Energy & Industrial Systems.
Funding stage: Series A.
Founded: 2022.
Country: Switzerland.
Employees: 11-50 employees.
Landscape fit: It fits World Model Startups because its core product is a model of how large physical systems evolve, used for forecasting and operational decisions in energy, climate, and related industrial domains.
Funding rounds
-
Series A
announced 2025-06-01
: 11000000 USD
; investors: Ananda Impact Ventures, Future Energy Ventures, Promus Ventures, 468 Capital, Kadmos Capital
; source: Jua Blog
; https://jua.ai/blog/series-a-announcement/
-
Seed
announced 2024-02-01
: 16000000 USD
; investors: 468 Capital, Green Generation Fund, Promus Ventures, Kadmos Capital, Flix Mobility founders, Session.vc, Virtus Resources Partners, Notion.vc, InnoSuisse
; source: TechCrunch
; https://techcrunch.com/2024/02/05/jua-raises-16m-to-build-a-foundational-ai-model-for-the-natural-world-starting-with-the-weather/
Company timeline
-
2026-07-21:
EPT-2.1 Europa European day-ahead model
(product)
. Jua launched Europa, an hourly-refreshed European ensemble forecast with ~6 km resolution, 16 ensemble members, and day-ahead trading focus.
Source: Jua blog.
https://jua.ai/blog/ept2-1-europa/
-
2026-07-21:
EPT-2.1 Europa extends the platform with day-ahead forecasting for European power traders
(gtm)
. On July 21, 2026, Jua launched EPT-2.1 Europa, an hourly-refreshed, high-resolution ensemble forecast aimed at day-ahead trading across wind, solar, temperature, and Dutch wind power.
Source: Jua Blog.
https://jua.ai/blog/
-
2026-07-15:
EPT-2.1 Helios solar model
(product)
. Jua introduced Helios, a European solar forecasting model with 30-minute updates, 48-hour horizon, and historical backtesting archive.
Source: Jua blog.
https://jua.ai/blog/ept2-helios/
-
2026-05-08:
Morning Briefings workflow feature
(product)
. Jua launched Morning Briefings, an automated per-zone summary that compares model agreement, forecast shifts, and prior-day performance for traders.
Source: Jua blog.
https://jua.ai/blog/morning-briefings/
-
2026-05-08:
Morning Briefings go live across most European zones
(gtm)
. Jua added Morning Briefings in May 2026, an automated per-zone trading-day summary that synthesizes model agreement, forecast changes, and prior-day performance for traders.
Source: Jua Blog.
https://jua.ai/blog/morning-briefings/
-
2026-01-30:
EPT-2e extended-range forecast update
(product)
. Jua expanded EPT-2e from a 20-day product to a 60-day forecast and signaled a path toward longer horizons.
Source: Jua blog.
https://jua.ai/blog/ept2e-subseasonal/
kausable
Website: kausable.ai
kausable is a Heidelberg-based AI startup building causal, adaptive world models that reason from first principles and update without retraining; its flagship model, TipPFN, is positioned to predict rare but consequential events in complex dynamic systems. ([kausable.ai](https://kausable.ai/?utm_source=openai))
Category: Earth, Energy & Industrial Systems.
Funding stage: Seed.
Founded: 2025.
Country: Germany.
Landscape fit: It fits the World Model Startups category because the company explicitly builds a model that learns underlying dynamics and adapts to new evidence across complex, real-world systems such as energy grids, cybersecurity, and robotics, rather than a narrow task-specific model. ([kausable.ai](https://kausable.ai/?utm_source=openai))
Funding rounds
-
Seed round
announced 2026-07-23
: 12000000 EUR
; investors: UVC Partners, Entourage, HTGF, Mätch VC
; source: www.htgf.de
; https://www.htgf.de/en/kausable-seed-round/
Company timeline
-
2026-07-23:
kausable raises €12 million seed round
(funding)
. kausable announced a €12 million seed round led by UVC Partners and Entourage, with follow-on investment from HTGF and Mätch VC, to build reasoning-first frontier AI that adapts without retraining.
Source: www.htgf.de.
https://www.htgf.de/en/kausable-seed-round/
Archetype AI
Website: archetypeai.io
Archetype AI builds a Physical AI foundation model and platform that turns real-world sensor data into understanding, predictions, and actions for physical environments.
Category: Earth, Energy & Industrial Systems.
Funding stage: Series A.
Founded: 2023.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits the World Model Startups category because its core product learns how physical systems evolve from multimodal sensor data and uses that understanding to support operational decisions in the real world.
Funding rounds
-
series a
announced 2025-11-01
: 35000000 USD
; investors: IAG Capital Partners, Hitachi Ventures, Bezos Expeditions, Venrock, Amazon Industrial Innovation Fund, Samsung Ventures, Systemiq Capital, E12 Ventures, Higher Life Ventures, Gaingels, Plug and Play Ventures
; source: Archetype AI
; https://www.archetypeai.io/blog/archetype-ai-series-a
-
seed
announced 2024-04-01
: 13000000 USD
; investors: Venrock, Amazon Industrial Innovation Fund, Hitachi Ventures, Buckley Ventures, Plug and Play Ventures
; source: Business Wire
; https://www.businesswire.com/news/home/20240405306572/en/Archetype-AI-Introduces-Foundation-Model-to-Pioneer-Physical-AI
Company timeline
-
2026-06-24:
Newton Agents launch productizes the platform into ready-to-deploy workflows
(gtm)
. Archetype AI introduced Newton Agents as a portfolio of ready-to-deploy physical agents for machine intelligence, process intelligence, and workforce intelligence. ([archetypeai.io](https://www.archetypeai.io/blog/introducing-newton-agents?utm_source=openai))
Source: Archetype AI Blog.
https://www.archetypeai.io/blog/introducing-newton-agents
-
2026-06-24:
Newton Agents portfolio launched
(product)
. Archetype introduced Newton Agents as ready-to-deploy physical agents spanning operational state monitoring, anomaly discovery, rare event detection, task verification, and manual generation.
Source: Archetype AI Blog.
https://www.archetypeai.io/blog/introducing-newton-agents
-
2026-04-22:
Newton fine-tuning for customer operations
(product)
. Archetype introduced fine-tuning so customers can adapt Newton to their own sensor data and processes while keeping data and inference inside their infrastructure.
Source: Archetype AI Blog.
https://www.archetypeai.io/blog/newton-the-industrial-ai-foundation-model-built-for-physical-operations
-
2026-04-07:
Dong Lin joined as Principal Research Engineer
(hr)
. Archetype AI announced Dong Lin as Principal Research Engineer leading the Foundation Model Team.. Team size observed: 11-50 employees employees.
Source: Archetype AI Blog.
https://www.archetypeai.io/blog/archetype-ai-dong-lin-principal-research-engineer-foundation-model-team
-
2026-03-19:
T-Mobile and NVIDIA partnership broadens infrastructure-led distribution
(gtm)
. Archetype AI announced a partnership with T-Mobile and NVIDIA to support distributed Physical AI deployment closer to where physical data is generated and consumed. ([archetypeai.io](https://www.archetypeai.io/blog/archetype-ai-partners-with-t-mobile-and-nvidia-to-accelerate-ai-in-the-physical-world?utm_source=openai))
Source: Archetype AI Blog.
https://www.archetypeai.io/blog/archetype-ai-partners-with-t-mobile-and-nvidia-to-accelerate-ai-in-the-physical-world
-
2026-01-15:
Priya Shivakumar appointed Chief Product Officer
(hr)
. Archetype AI announced Priya Shivakumar as Chief Product Officer to oversee product strategy for Physical AI deployments and platform expansion.. Team size observed: 11-50 employees employees.
Source: Archetype AI Blog.
https://www.archetypeai.io/blog/archetypeai-priya-shivakumar-chief-product-officer
Causal Labs
Website: causallabs.ai
Causal Labs is building a Large Physics foundation Model to learn cause-and-effect in physical systems, starting with weather prediction and control.
Category: Earth, Energy & Industrial Systems.
Funding stage: Seed.
Founded: 2024.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits World Model Startups because it is explicitly training a model of a large physical system (the atmosphere) to predict future states and support better operational decisions in the real world.
Funding rounds
-
seed
announced 2025-03-12
: 6000000 USD
; investors: Kindred Ventures, Refactor, BoxGroup, Factorial, Otherwise, Karman Ventures, angel investors
; source: Causal Labs: Towards general causal intelligence
; https://blog.getcausal.ai/p/causal-labs-towards-causal-intelligence
Company timeline
-
2026-07-01:
Hiring push announced on LinkedIn
(hr)
. Dar Mehta posted that Causal Labs was hiring aggressively and asked impacted Meta Superintelligence Labs staff to send their resumes.. Team size observed: 11-50 employees employees.
Source: LinkedIn.
https://www.linkedin.com/posts/darmehta_to-all-of-the-brilliant-staff-at-meta-superintelligence-activity-7387169775849943041-Nqay
-
2025-03-12:
seed
(funding)
. Amount: 6000000 USD. Investors: Kindred Ventures, Refactor, BoxGroup, Factorial, Otherwise, Karman Ventures, angel investors
Source: Causal Labs: Towards general causal intelligence.
https://blog.getcausal.ai/p/causal-labs-towards-causal-intelligence
Biology
Models that represent living systems and predict how they respond when their internal state or external conditions change.
Noetik
Website: noetik.ai
Noetik is an AI-native biotechnology company that builds foundation models and multimodal data platforms to discover and develop precision cancer therapies.
Category: Biology.
Funding stage: Series A.
Founded: 2022.
Country: United States.
Employees: 11-50 employees.
Landscape fit: It fits the World Model Startups / Biology category because its core product is a model of the tumor microenvironment that can represent and simulate how living systems respond to changing conditions and therapies.
Funding rounds
-
series a
announced 2024-08-01
: 40000000 USD
; investors: Polaris Partners, Khosla Ventures, Wittington Ventures, Breakout Ventures, DCVC, Zetta Venture Partners, Catalio Capital Management, 11.2 Capital, Epic Ventures, Intermountain Ventures, North South Ventures, ApSTAT Technologies, Linearis Labs and Ventures Fund
; source: Business Wire
; https://www.businesswire.com/news/home/20240829428359/en/Noetik-Secures-%2440-Million-Series-A-Financing-to-Advance-Precision-Cancer-Therapies
-
seed
announced 2023-09-01
: 14000000 USD
; investors: DCVC, Zetta Venture Partners, 11.2 Capital, Catalio Capital Management, Epic Ventures, Intermountain Ventures, North South Ventures, Chau Khuong, CJNV BioVentures, Viswa Colluru, Hummingbird Nomads Fund, Chris Gibson, Michael Secora
; source: Business Wire
; https://www.businesswire.com/news/home/20230907509972/en/Noetik-Raises-%2414-Million-Seed-Financing-to-Revolutionize-Cancer-Immunotherapy-Using-Artificial-Intelligence
Company timeline
-
2026-07-27:
Team size observed
(hr)
. Public source shows a team size of 11-50 employees employees.. Team size observed: 11-50 employees employees.
Source: LinkedIn company page.
https://www.linkedin.com/company/noetik-inc
-
2026-06-15:
TARIO-2 white paper and validation launch routine-pathology response prediction
(product)
. Noetik released a white paper showing TARIO-2 could identify BOT/BAL responders and improved response and survival outcomes using only pretreatment H&E images.
Source: Noetik TARIO-2.
https://www.noetik.ai/tario2
-
2026-01-08:
GSK licenses Noetik foundation models in anchor partnership
(product)
. Noetik announced that GSK licensed its AI foundation models in an anchor partnership to transform cancer therapeutic research and development.
Source: Noetik News.
https://www.noetik.ai/news
-
2025-11-10:
SITC 2025 introduces virtual-cell simulations for treatment outcomes
(product)
. Noetik presented OCTO-vc research showing virtual-cell simulations can decode spatial biology and support target discovery, biomarker development, and patient stratification.
Source: Noetik SITC 2025.
https://www.noetik.ai/sitc-2025
-
2025-06-17:
Entered collaboration with Agenus on predictive biomarkers
(gtm)
. Noetik announced a collaboration with Agenus to develop AI-enabled predictive biomarkers for BOT/BAL using foundation models of virtual cell biology.
Source: Noetik News.
https://www.noetik.ai/news
-
2025-06-17:
Agenus collaboration adds predictive biomarker use case
(product)
. Noetik and Agenus entered a collaboration to develop AI-enabled predictive biomarkers for BOT/BAL using foundation models of virtual cell biology.
Source: Noetik News.
https://www.noetik.ai/news
GenBio AI
Website: genbio.ai
GenBio AI builds multiscale AI foundation models for biology, centered on an AI-driven digital organism that can simulate, predict, and program biological systems across DNA, RNA, protein, cell, and tissue scales.
Category: Biology.
Funding stage: Seed.
Founded: 2024.
Country: United States.
Employees: 51-200 employees.
Landscape fit: It fits this category because its core product is a biological world model/AIDO that represents living systems and predicts responses to perturbations, such as gene knockdowns, drugs, and cellular changes, across multiple biological scales. GenBio AI says it is building a programmable virtual cell and models to simulate biology at every scale. ([genbio.ai](https://genbio.ai/home/?utm_source=openai))
Company timeline
-
2026-06-23:
Announced NVIDIA collaboration for virtual-cell scale-up
(gtm)
. GenBio AI said it would collaborate with NVIDIA and integrate BioNeMo, Megatron, NIM microservices, and the BioNeMo Agent Toolkit to accelerate construction and deployment of virtual-cell systems.
Source: GenBio AI.
https://genbio.ai/advancing-agentic-ai-for-building-virtual-cells-powered-by-nvidia/
-
2026-06-23:
GenBio AI announced NVIDIA collaboration for virtual-cell world models
(product)
. GenBio AI said it was collaborating with NVIDIA to accelerate virtual-cell world models and integrate BioNeMo, Megatron, NIM microservices, and the BioNeMo Agent Toolkit.
Source: Genbio AI.
https://genbio.ai/advancing-agentic-ai-for-building-virtual-cells-powered-by-nvidia/
-
2026-03-26:
AIDO.Tissue launched for spatially aware tissue modeling
(product)
. GenBio AI released AIDO.Tissue, a spatial transcriptomics framework that learns from neighboring cells to model tissue context.
Source: Genbio AI.
https://genbio.ai/aido-tissue/
-
2025-07-19:
AIDO.StructureDiffusion launched for protein design
(product)
. GenBio AI introduced AIDO.StructureDiffusion, a generative module for protein backbone generation, monomers, complexes, binders, and antibodies.
Source: Genbio AI.
https://genbio.ai/aido-structurediffusion-the-aido-module-for-molecular-design/
-
2025-06-17:
Added Emma Lundberg as co-founder and chief scientific advisor
(gtm)
. GenBio AI announced Emma Lundberg as co-founder and chief scientific advisor, with the role focused on data strategy and scientific initiatives for generative digital organism work.
Source: PR Newswire / GenBio AI.
https://www.prnewswire.com/news-releases/genbio-ai-welcomes-emma-lundberg-as-co-founder-and-chief-scientific-advisor-302483814.html
-
2025-06-17:
Emma Lundberg joins as co-founder and chief scientific advisor
(hr)
. GenBio AI announced that Emma Lundberg officially joined as Co-Founder and Chief Scientific Advisor to lead data strategy and scientific initiatives.. Team size observed: 51-200 employees employees.
Source: GenBio AI.
https://genbio.ai/emma-lundberg-co-founder-chief-scientific-advisor/
Bioptimus
Website: bioptimus.com
Bioptimus is a French AI-biotech company building a universal foundation model for biology to connect and predict across molecular, cellular, tissue, and organism-level data.
Category: Biology.
Funding stage: Seed.
Founded: 2023.
Country: France.
Landscape fit: It fits the Biology world-model category because its core product is a multi-scale foundation model intended to represent living systems and predict biological behavior across changing internal states and conditions.
Funding rounds
-
Series A
announced 2025-01-01
: 41000000 USD
; investors: Cathay Innovation, Sofinnova Partners, Bpifrance, Bpifrance Large Venture, Andera Partners, Hitachi Ventures, Boom Capital Ventures, Pomifer Capital, Sunrise, Emmanuel Cassimatis, Thomas Wolf
; source: Bioptimus
; https://www.bioptimus.com/news/bioptimus-hits-76m-funding
-
Seed
announced 2024-02-01
: 35000000 USD
; investors: Sofinnova Partners, Bpifrance Large Venture, Frst, Cathay Innovation, Headline, Hummingbird, NJF Capital, Owkin, Top Harvest Capital, Xavier Niel
; source: Bioptimus
; https://www.bioptimus.com/news/ex-google-deepmind-and-owkin-scientists-team-up-to-create-bioptimus-to-build-the-first-universal-ai-foundation-model-for-biology
Company timeline
-
2026-05-20:
M-Optimus-1 public results deepen the multimodal story
(gtm)
. Bioptimus published results showing M-Optimus-1 as a first iteration of its world model, emphasizing multimodal prediction, routine-data leverage, and STELA as the scaling engine.
Source: Bioptimus.
https://www.bioptimus.com/introducing-m-optimus
-
2026-05-20:
M-Optimus introduced as Bioptimus’s multimodal world model for biology
(product)
. Bioptimus introduced M-Optimus-1 as a multimodal, multi-scale model combining pathology, bulk RNA-seq, and spatial transcriptomics to predict molecular geography and support drug discovery workflows.
Source: Bioptimus Blog.
https://www.bioptimus.com/introducing-m-optimus
-
2026-03-25:
STELA launches as a data-and-partnership engine for M-Optimus
(gtm)
. Bioptimus launched STELA with 10x Genomics and Broad Clinical Labs to build a clinically linked spatial biology atlas and support M-Optimus with large-scale multimodal data.
Source: Bioptimus.
https://www.bioptimus.com/news/bioptimus-announces-stela-the-worlds-largest-clinically-linked-spatial-biology-atlas-alongside-key-partners-10x-genomics-and-broad-clinical-labs
-
2026-03-25:
STELA launches as Bioptimus’s clinically linked spatial biology data engine
(product)
. Bioptimus launched STELA with 10x Genomics and Broad Clinical Labs to generate the multimodal, clinically linked datasets that feed M-Optimus.
Source: Bioptimus News.
https://www.bioptimus.com/news/bioptimus-announces-stela-the-worlds-largest-clinically-linked-spatial-biology-atlas-alongside-key-partners-10x-genomics-and-broad-clinical-labs
-
2025-12-17:
M-Optimus unveiled with early-access commercial motion
(gtm)
. Bioptimus unveiled M-Optimus and opened early access to a limited number of pioneer clients while also expanding access to H-Optimus-1 through Amazon SageMaker AI.
Source: Bioptimus.
https://www.bioptimus.com/news/bioptimus-unveils-m-optimus-a-world-model-for-biology
-
2025-09-11:
Scientific Advisory Board adds external credibility
(gtm)
. Bioptimus formed a scientific advisory board with senior leaders from Cambridge, MIT, Broad, and other institutions to guide its biology model strategy.
Source: Bioptimus.
https://www.bioptimus.com/news/bioptimus-assembles-elite-scientific-advisory-board-to-build-next-gen-ai-models-for-biology
