Startups Venture

Nvidia Invests In Deepgram For The Future Of Voice Recognition

Nvidia’s venture capital arm, Nvidia GPU Ventures, announced an investment of an undisclosed amount in Deepgram, a deep learning-based automatic speech recognition (ASR) platform.

Follow Crunchbase News on Twitter & Facebook

Deepgram previously raised seed funding from Boost VC in November 2015. The startup participated in Y Combinator’s winter 2016 batch after receiving a $120,000 investment from the accelerator. Later in September 2017, Deepgram raised $1.8 million in new financing led by Y Combinator and Compound with Zillionize Angel, Fenox Venture Capital, 1517 Fund, and Ludlow Ventures also participating in the round.

Deepgram was founded in 2015 by Scott Stephenson and Noah Shutty. With end-to-end deep learning technology built with GPUs, the startup aims to provide companies and individuals with ASR that is powered by advanced, multi-layered machine learning technology. Their current ASR product, Deepgram Brain, offers transcription and audio search capabilities through individual and enterprise solutions.

According to CEO Scott Stephenson, voice search is one capability powered by deep learning that differentiates the company from other services. The feature allows companies to search for keywords and phrases in collected audio data, potentially allowing them to gather insight into methodology.

Moving forward, Deepgram hopes to expand its partnerships in the B2B world in financial services, insurance, sales, and support, as well as with platform companies, like large call centers, which are primarily used by mid-market companies and SMBs. And that’s not the end of the road.

“Right now people are still sort of afraid of working with their speech data. But as you get a cost effective, easy, fast, accurate way to do transcription, that’s sort of the gateway drug,” he told Crunchbase News in an interview. Stephenson further believes that, in the future, companies interested in deep learning will move from collecting and analyzing data to interacting directly with their AI to improve their outcomes.

Further down the road, he sees greater potential in the consumer-driven market for deep learning-based audio technology: “It’s not going to be in the way that you talk to Siri and get frustrated with it or any of the other assistants. It’s going to be a much more natural process.”

In light of that potential for movement forward, Stephenson is very confident about Deepgram’s position in this space.

“The biggest misunderstanding is that a company like Deepgram can’t actually battle the big guys because they have all the data, they have all the technology, they have all the engineers. That’s just not true in the ASR space.” Stephenson explained. “They’re using old tech, [and] they have a lot of momentum in the wrong direction. And it’s going to take them time to switch over.”

Though the terms of its latest round have yet to be disclosed, Crunchbase News will keep you updated on further developments.

Illustration Credit: Li Anne Dias

Copy link