Sometimes it starts in a university lab. Other times, in a garage. SuperAnnotate? It started in the friction between theory and reality, a pair of PhD candidates who saw the bottleneck in AI, said “enough,” and made the jump from academia to enterprise. What Vahan Petrosyan and Tigran Petrosyan built wasn’t just a startup. AI isn’t starving for models. It’s starving for good data and the infrastructure to scale it.
Fast forward to July 15, 2025, and the signal just got a hell of a lot louder. SuperAnnotate has secured a $13.5 million Series B extension led by Dell Technologies Capital, stacking their total funding to over $67 million. No loud victory laps. Just methodical execution, and another vote of confidence from serious operators. These aren’t your “spray and pray” investors. They bet where it counts.
Let’s be clear: dataset creation isn’t sexy. But it’s absolutely essential. SuperAnnotate’s platform is where AI actually gets built, where raw, chaotic inputs become training gold. From image to audio, video to 3D, it’s one unified pipeline with human- and agent-in-the-loop annotation, full CI / CD orchestration, QA automation, and real-time evaluation. Fortune 50 aren’t guessing when they pick this stack. They’re optimizing for impact.
Co-founder & CTO Davit Badalyan brought the hardware edge. Jason Liang, Co-founder & SVP of Business Development, added the enterprise punch. This isn’t a bedroom project that lucked into scale. This is a global operation with offices in San Francisco, Yerevan, and Stockholm, and partnerships with Databricks, Canva, Motorola Solutions, IBM, and Qualcomm. When they say they’re running at enterprise velocity, believe them. Software revenue 5× last year. Customer base 3×. Over 3,000 data scientists in the stack. You don’t hit those numbers selling pipe dreams.
What’s next? Expansion into healthcare, financial services, and autonomous systems. More R&D, more annotation capacity, and a doubling of their engineering and go-to-market teams. Real-time pipelines, RAG dataset support, agent workflows, because in this game, latency kills, and quality controls win championships.
If you’re in the race to build the next-gen GenAI infrastructure, or just tired of slapping datasets together, SuperAnnotate isn’t optional. It’s foundational. Congrats to Vahan Petrosyan, Tigran Petrosyan, Davit Badalyan, and Jason Liang. And props to Dell Technologies Capital for spotting the real play.
The AI arms race isn’t about who trains faster. It’s about who trains better. And right now, SuperAnnotate is the quiet assassin in the room, scaling the part of AI that no one wants to touch, but everyone needs to win.
