BeatpulseLabs, a London-based AI data company, has raised £1.4 million in pre-seed funding led by Araya Ventures and Lighthouse Ventures, with participation from Alumni Ventures and Avalancha Ventures.

The firm has witnessed 10x revenue growth over the first half of 2026.

The emergence of multimodal AI systems used in the enterprise has created growing demand for data that reflects the complexity of the real world. As companies build increasingly sophisticated models, the limitation is no longer access to raw training data, but the ability to encode human judgement in the context of the specific use case. Beatpulse is positioned to become the foundation data infrastructure layer targeting this gap.

Founded by South African Jason Rieff and Bulgarian Nikolay Vitanov, the company is addressing a growing challenge in artificial intelligence: that most multimodal models are trained on poor training data, limiting their ability to perform reliably in real-world environments.

BeatpulseLabs combines two tightly integrated core offerings: dataset preparation and dataset provision. The company transforms existing multimedia content libraries into enterprise-grade training datasets by cleaning, structuring, labelling, validating, enriching, and formatting raw speech, music and video assets. It also provides ready-made and custom, rights-cleared datasets for companies that need high-quality training data without starting from their own archive.

The result is enterprise-grade, context-rich data built for model training, fine-tuning and reinforcement learning. This shortens training time, helps improve model accuracy and reduces hallucinations.