aid: habana-labs accessModel: pricing: unknown onboarding: unknown trial: false try_now: false public: false label: Unknown confidence: low source: [] generated: '2026-07-22' method: derived image: https://kinlane-images.s3.amazonaws.com/shared/apis-json/icons/habana-labs.png name: Habana Labs description: >- Habana Labs is an artificial-intelligence processor company, now part of Intel, that designs deep-learning accelerators for AI training and inference. Its Gaudi (training) and Goya (inference) processors — sold today as the Intel Gaudi 2 and Gaudi 3 AI accelerators — are programmed through the SynapseAI software suite, which integrates with PyTorch and DeepSpeed and ships reference models, containers, and installation tooling. Habana was founded in 2016 in Israel and acquired by Intel in 2019. Its public developer surface is not a hosted web API but an open-source SDK and toolchain: the HabanaAI GitHub organization (reference models, PyTorch bridge, setup/install, plus vLLM and Megatron forks) and the Intel Gaudi documentation at docs.habana.ai. This profile catalogs that developer footprint for the API Evangelist network. url: https://raw.githubusercontent.com/api-evangelist/habana-labs/refs/heads/main/apis.yml x-type: company x-source: vc-portfolio x-backed-by: - battery-ventures - bessemer-venture-partners x-tier: stub x-tier-reason: portfolio-lead specificationVersion: '0.20' created: '2026-07-17' modified: '2026-07-19' tags: - Company - Artificial Intelligence - Machine Learning - AI Accelerators - Deep Learning - Hardware - Semiconductors - SDK - Developer Tools - Intel apis: [] maintainers: - FN: Kin Lane email: kin@apievangelist.com - FN: APIs.json email: info@apis.io common: - type: Website url: https://www.habana.ai/ - type: Documentation url: https://docs.habana.ai/ - type: GitHubOrganization url: https://github.com/HabanaAI - type: DomainSecurity url: security/habana-labs-domain-security.yml x-enrichment: date: '2026-07-19' status: backfilled pass: local-v1 note: backfilled from .gitignore signal + verified work evidence