Mathpix joins DataVerge colocation facility to support AI workflows

Author: Isha Jain

Mathpix, an AI-powered document automation and scientific communication company, has joined DataVerge, a carrier-neutral interconnection facility in Brooklyn. DataVerge surpassed Mathpix’s criteria, which included robust and redundant power, a fast connection to AWS US-East-1, scalability and proximity to its Brooklyn, New York headquarters, making it the colocation facility of choice.

As more companies rely on AI for their business, colocation and data centres must deliver greater than ever levels of uninterrupted power and connectivity to support high-density AI workloads. Though many companies with a thriving presence in the New York metropolitan area, are now seeking to reap the benefits of AI, few New York area data centres are equipped with the abundance of power required to meet their AI needs. In addition to power, DataVerge, which is the largest interconnection facility in Brooklyn, New York, offers more than 50,000ft² of customisable data centre space, along with secure IT infrastructure and rapid deployment and connection to more than 30 carriers as well as 24/7 access and support.

Mathpix enables organisations to quickly and accurately convert PDFs and other types of documents, including handwritten text and images, into searchable, exportable and machine readable text used in large language models (LLMs) and other applications.

According to Nico Jimenez, CEO of MathPix, “DataVerge enables us to colocate our own servers, which are equipped with our own GPUs. This setup provides us the opportunity to select the hardware we need to build and configure our servers so that they significantly reduce latency, and at a considerably lower price point than what the hyperscalers charge for colocation.”

“AI is essential to how many New York area companies run their businesses,” says Jay Bedovoy, CTO of DataVerge. “Our colocation and data centres provide them with uninterrupted power and connectivity, as well as the scalability, high performance and proximity needed to avoid latency issues, which is critical for AI and other real-time interaction in decision-making applications.”