Molecula’s data virtualization platform processes data a thousand times faster than traditional methods, which is the kind of progress you can only make by accident.
Five years ago, H.O. Maycotte had tasked his team at Umbel to find ways to reduce the number of servers and time required to process customer data. Most requests took 20 seconds and 40 servers to process. It's a problem many companies are facing, Maycotte said, and his company was just trying to make it a little bit better.
After patching together a solution, it took only a few milliseconds to complete and needed just two servers.
“We thought it was broken,” Maycotte told Built In.
After looking a little closer, they realized it had worked properly and that they had stumbled on the solution to one of the biggest challenges of the information age — processing all the data companies are accumulating.
Five years later, that discovery has become the foundation for a new company, Molecula, which announced today that it has raised $6 million in seed funding. The Austin-based company provides a data virtualization platform that is designed to speed up the process machine learning programs go through in analyzing data.
We’re a thousand times faster for machine learning models versus the traditional models.”
“We’re a thousand times faster for machine learning models versus the traditional models,” said Maycotte, who is CEO.
The current process requires making full copies of data files to index, cache and process them, Maycotte said. It takes up terabytes of space and is both slow and vulnerable to attack. As a result, machine learning is only able to access a fraction of the data a company has to make a decision, Maycotte said.
The solution his company discovered involves making representations of the data. These representations are a fraction of the size, but contain all the information a machine learning system needs to process a request and analyze it. In doing so, it will allow machine learning tools to access all of a company’s data to make decisions, Maycotte said.
“(Data virtualization) will be as ubiquitous as virtualized computing is today,” Maycotte said. “We want machines making decisions on 100 percent of data rather than fractions of it.”
With this round, Molecula plans to expand its go-to market efforts to reach more customers. Pilosa, Molecula’s open-source platform, is currently used by more than 1,800 organizations around the world, according to the company.
With data management expected to become a $146 billion industry by 2023, there is plenty of room to grow.
“We’re excited about where the product is today,” Maycotte said. “We have an ambitious roadmap ahead of us. A lot of it will be focused on resource utilization and making it incredibly efficient. We want to use the cloud to out cloud the cloud.”
Seraph Group, Lontra Ventures, Velar Capital, Capital Factory, Andrew Busey and Jason Dorsey all participated in the round.