SuperAnnotate
SuperAnnotate is a data annotation and AI development platform founded in August 2018 by brothers Tigran and Vahan Petrosyan, who dropped out of their respective PhD programs at ETH Zurich and KTH Royal Institute of Technology to start the company.
Profile
SuperAnnotate provides a platform for labeling, curating, and evaluating multimodal data used to train, fine-tune, and test AI models.
SuperAnnotate is a data annotation and AI development platform founded in August 2018 by brothers Tigran and Vahan Petrosyan, who dropped out of their respective PhD programs at ETH Zurich and KTH Royal Institute of Technology to start the company. The platform originally focused on image annotation for computer vision, but has since expanded to support multimodal data labeling, RLHF, SFT, RAG, and agent evaluation workflows. Headquartered in Silicon Valley after joining Berkeley's SkyDeck accelerator in 2019, the company has raised a total of $53.5 million across seed ($3M, June 2020), Series A ($14.5M, July 2021), and Series B ($36M, November 2024) rounds.
The Series B was led by Socium Ventures (backed by Cox Enterprises), with participation from NVIDIA, Databricks Ventures, and Lionel Messi's Play Time Ventures. SuperAnnotate claims to be the #1 ranked data labeling platform on G2. As of early 2026, the company is pivoting toward agentic AI evaluation and fine-tuning workflows, offering workshops and solutions for AI agent performance insights.
No recent revenue or headcount figures are publicly available; the company is privately held and has not disclosed financial performance. The platform integrates with AWS, GCP, and Azure, and counts customers in autonomous driving, healthcare, and enterprise AI. The company's CTO is Davit Badalyan; Jason Liang serves as SVP of Business Development.
Who buys this
- Enterprise AI teams building custom large language models (LLMs) and multimodal models
- Autonomous vehicle companies requiring high-volume image and video annotation
- Healthcare and life sciences organizations needing domain-specific data labeling for medical imaging and NLP
- AI research labs and academic institutions working on computer vision and NLP benchmarks
- Companies deploying AI agents or RAG pipelines that require human-in-the-loop evaluation
Strengths and what to watch
Strengths
- Backed by strategic investors including NVIDIA and Databricks Ventures, providing both capital and ecosystem integration potential
- Platform covers the full AI data lifecycle from annotation through model evaluation, reducing toolchain fragmentation for enterprise customers
- Claims top G2 ranking in data labeling category, suggesting strong user satisfaction relative to competitors
Watch for
- No publicly disclosed revenue or customer count makes it difficult to assess traction or growth trajectory
- The data annotation market is crowded with well-funded competitors (Scale AI, Labelbox, Sama) and open-source alternatives, pressuring differentiation
- Company has pivoted from pure computer vision annotation to broader AI evaluation workflows; execution risk in capturing enterprise agentic AI budgets remains unproven
Key Information
- Industry
- Data Generation & Labelling
- Founded
- 2018
- Headquarters
- Silicon Valley
Frequently Asked Questions
What is SuperAnnotate and what does it do?
SuperAnnotate is a platform for labeling, curating, and evaluating multimodal data used to train, fine-tune, and test AI models. Founded in 2018, it supports image, video, text, and audio annotation, plus RLHF, SFT, RAG, and agent evaluation workflows.
Who founded SuperAnnotate and how much funding has it raised?
SuperAnnotate was founded in August 2018 by brothers Tigran and Vahan Petrosyan, who dropped out of PhD programs at ETH Zurich and KTH. The company has raised $53.5 million total across seed, Series A, and Series B rounds, the latest led by Socium Ventures.
What types of data annotation does SuperAnnotate support?
SuperAnnotate supports multimodal data labeling, including images, video, text, and audio. It covers computer vision, natural language processing, RLHF, SFT, RAG, and agent evaluation workflows, helping enterprises build custom LLMs and multimodal AI models.
Who are SuperAnnotate's main customers and investors?
Customers include enterprise AI teams, autonomous vehicle companies, healthcare organizations, and AI research labs. Strategic investors include NVIDIA, Databricks Ventures, and Lionel Messi's Play Time Ventures, with Series B led by Socium Ventures.
How does SuperAnnotate compare to competitors like Scale AI or Labelbox?
SuperAnnotate claims the #1 ranking on G2 for data labeling and covers the full AI data lifecycle from annotation to model evaluation. However, the market is crowded with well-funded competitors and open-source alternatives, making differentiation a key challenge.
Is SuperAnnotate pivoting toward agentic AI evaluation?
Yes, as of early 2026, SuperAnnotate is pivoting toward agentic AI evaluation and fine-tuning workflows. It offers workshops and solutions for AI agent performance insights, though execution risk in capturing enterprise agentic AI budgets remains unproven.
Sources
- www.superannotate.com — Founding story, funding rounds, accelerator program, platform milestones, leadership team
- techcrunch.com — Series B funding announcement ($36M), investor details (Socium Ventures, NVIDIA, Databricks Ventures, Play Time Ventures)
- www.superannotate.com — Current product positioning (RLHF, SFT, agents, RAG, evaluation), G2 ranking claim, integration with AWS/GCP/Azure
- investors.fiscalnote.com — Context on FiscalNote's Q1 2026 results (unrelated company, included in dossier but not relevant to SuperAnnotate)