{"id":"5fe2b33b-9f6a-49a3-9de1-3917d174c611","slug":"spirit-brain","name":"Spirit AI Brain","description":"Spirit AI's general-purpose embodied AI model for humanoid robots. Trained on 'dirty data' — unstructured, real-world data pipelines. Raised $280M Series A at $1.2B valuation in February 2026. Develops universal robotic brains for the Moz1 humanoid platform.","brainType":"foundation-model","isOpen":false,"maturityStage":"pilot","architecture":"Embodied AI model trained on real-world 'dirty data'","reviewStatus":"reviewed","sources":[{"url":"https://www.prnewswire.com/news-releases/spirit-ai-lands-280m-to-scale-embodied-ai-through-dirty-data-302697085.html","date":"2026-02-25","title":"Spirit AI Lands $280M to Scale Embodied AI Through Dirty Data","sourceName":"PR Newswire / Spirit AI"},{"url":"https://www.spirit-ai.com/en/about","title":"About Us","sourceName":"Spirit AI"},{"url":"https://www.spirit-ai.com/en/product","title":"Product","sourceName":"Spirit AI"},{"url":"https://www.spirit-ai.com/en/blog/spirit-v1-5","title":"Spirit-v1.5: Clean Data Is the Enemy of Great Robot Foundation Models","sourceName":"Spirit AI"},{"url":"https://github.com/Spirit-AI-Team/spirit-v1.5","title":"Spirit-v1.5 official repository","sourceName":"Spirit AI"}],"keyFacts":[{"label":"Universal Brain dirty-data VLA (primary PR Newswire / Spirit AI)","value":"Spirit AI's Feb 25 2026 release describes building a universal robotic brain via Scaling Law-driven VLA architectures trained on diverse human video and wearable sensor data rather than over-curated clean datasets. Co-founder and Chief Scientist Yang Gao is quoted that dirty data is the key to scaling VLA models. The company frames peers as Google DeepMind and Physical Intelligence for massive-dataset physical reasoning (PR Newswire / Spirit AI)"},{"label":"Spirit v1.5 and industrial claim (primary PR Newswire / Spirit AI)","value":"The same release says Spirit AI amassed over 200,000 hours of interaction data (roadmap to exceed 1 million hours by end of 2026), reduced data-acquisition cost 90% via proprietary wearables versus teleoperation, and that Spirit v1.5 topped RoboChallenge in January 2026. It also claims CATL production-line agents handling flexible wire harnesses at 99%+ success matching skilled human cycle times. Treat benchmarks and rates as company-stated (PR Newswire / Spirit AI)"},{"label":"Universal embodied brain","value":"Spirit AI says it develops Universal Brains for general-purpose embodied models that bridge simulation and real-world robot deployment (About Spirit AI)."},{"label":"Spirit v1.5 platform","value":"Spirit AI identifies Spirit v1.5 as a technical implementation of its embodied intelligence work and says the platform is designed to support versatile physical interaction across environments (Spirit AI product)."},{"label":"Spirit-v1.5 architecture","value":"Spirit AI's official model post describes Spirit-v1.5 as a vision-language-action model for real-robot manipulation, evaluated on RoboChallenge Table30 across multiple robot embodiments (Spirit-v1.5)."},{"label":"Spirit-v1.5 open release","value":"Spirit AI links the Spirit-v1.5 implementation, runtime wrapper, model weights, and fine-tuned checkpoint for public reproduction and exploration (Spirit-v1.5 repository)."}],"aliases":[],"collisionRisk":"low","reviewNote":null,"builtOnBrainId":null,"createdAt":"2026-08-09T14:36:52.914Z","updatedAt":"2026-10-09T22:00:43.642Z","jsonLd":{"@context":"https://schema.org","@type":"SoftwareApplication","@id":"https://registry.deploy.report/brains/spirit-brain","url":"https://registry.deploy.report/brains/spirit-brain","name":"Spirit AI Brain","description":"Spirit AI's general-purpose embodied AI model for humanoid robots. Trained on 'dirty data' — unstructured, real-world data pipelines. Raised $280M Series A at $1.2B valuation in February 2026. Develops universal robotic brains for the Moz1 humanoid platform.","identifier":"5fe2b33b-9f6a-49a3-9de1-3917d174c611","applicationCategory":"foundation-model","publisher":{"@id":"https://deploy.report/#organization"}},"framework_metadata":{"framework_schema_version":"0.1.0","verification_status":"verified","maturity_stage":"pilot","lifecycle_state":null,"architectural_position":{"cohort":null,"sub_cohorts":[]},"within_cohort_verified_vs_claimed_pair":null,"cap_flags":[],"verification_depth":{"sources_count":5,"primary_source_types":["code-repository"]}}}