Abaka.AI secures $8M Series A to expand multimodal AI data infrastructure
What's the deal? Abaka.AI, an enterprise data infrastructure startup, has secured $8 million in funding to expand its platform for preparing and evaluating multimodal AI data. The company helps businesses clean, label, review, and secure the datasets that power AI systems across text, image, audio, video, code, and 3D formats.
The funding will go toward expanding production pipelines, automating complex supervision workflows, and building stronger model-in-the-loop evaluation tools. A portion will also support benchmark and interoperability work through the 2077AI Foundation, an open-source initiative Abaka.AI co-founded.
Why now? As AI models grow more complex and multimodal, the data behind them has become a bigger business problem. Companies building systems for healthcare, finance, robotics, or autonomous vehicles need specialised data workflows — different annotation standards, privacy controls, quality checks, and review processes — that go well beyond basic text handling.
"The Abaka AI Data Platform is built to give enterprises of any size access to production-grade datasets and evaluation pipelines," said Yunfei Zhao, co-founder and chief operating officer. "We want to support the best possible AI outcomes, and those are only as strong as the datasets behind them."
What could go wrong? Data infrastructure is an increasingly crowded space. As more startups and cloud providers build tools for AI data preparation, Abaka.AI will need to differentiate on quality, specialisation, and compliance — especially in regulated industries where tracking how data is handled is critical.
The signal: The AI data infrastructure layer is attracting increasing attention as enterprises realise that model performance hinges on what goes in, not just how it's built. Abaka.AI's $8 million raise — at the early growth stage — underscores a widening gap between the billions flowing into foundation models and the comparatively modest investment in the tooling needed to make enterprise data production-ready. As multimodal AI moves from research into regulated sectors like healthcare and autonomous systems, startups that can handle the complexity of annotation, compliance, and quality assurance across formats stand to become critical picks-and-shovels providers.
Read more: Yahoo Finance
