SoundHound, a US-based developer of sound recognition and conversational artificial intelligence technology, raised $75m last week, as investors continue to put money into companies working to advance the way AI interacts with audio input.
Samsung Catalyst Fund, Nvidia GPU Ventures and RSI Fund, respective subsidiaries of consumer electronics manufacturer Samsung, graphics technology producer Nvidia and human resources firm Recruit Holdings, all took part in the round, as did insurance firm Sompo Japan Nipponkoa and financial services provider Nomura Holdings.
The investor list was rounded out by SharesPost, MKann and venture capital firms Kleiner Perkins Caufield & Byers, Global Catalyst Partners, Walden Venture Capital and TransLink Capital, and brought the company’s overall funding to $115m since it was founded.
SoundHound was initially known for its music recognition app, which took the template set by Shazam and added features such as the ability to find songs by inputting lyrics or to see lyrics linked to the music as it plays.
The company introduced Houndify, a platform that enabled developers to add complex voice interaction to their products, in late 2015, and added a voice-powered personal assistance app called Hound a few months later. It intends to build its products into a large-scale platform for voice and sound-based AI that can be used by a range of partners.
E-commerce and cloud services firm Amazon is of course also developing and extending Alexa, its own voice-powered AI system, and has formed a dedicated corporate venturing fund called Alexa Fund to back startups developing compatible products.
However, SoundHound CEO Keyvan Mohajer told TechCrunch that SoundHound may have an advantage in that companies can incorporate its technology without requiring users to log into an Amazon account or reveal personal data to a third party. That element of its offering could explain why so many corporates contributed to the round – none of the corporate participants were described as existing investors, SoundHound having last disclosed funding in 2011.
In addition to supporting international growth, the round will help fund the development of what SoundHound calls Collective AI, a mechanism that allows developers to collaborate on the creation of domains – programs that facilitate a natural and fully conversational language interface on specific topics without users needing to remember specific phrases – by linking with data contained on a number of existing platforms.
Collective AI currently provides access to data from more than 100 platforms covering subjects such as transport, weather, stocks, sports, local businesses, flights, hotels and mortgages, including local services review platform Yelp, ride hailing app Uber and travel search engine Expedia. Users may one day be able to search for and book a holiday, hotel and ride to the airport simply by chatting to their phone.
SoundHound is not the only company developing sound-focused AI to have raised funding of late. Advanced acoustic sensor developer Vesper received $15m in a December 2016 series A round that included Alexa Fund, while sound recognition technology developer Audio Analytic closed a $5.5m round last month.
Companies such as Conversica and Tact, both of which closed series B rounds in December 2016, are meanwhile continuing to develop text-based conversational AI technology.
Bringing AI up to the point where users can interact seamlessly with a bot has long been seen as an endpoint and while it may still be a way off, signs indicate the technology is getting nearer. SoundHound meanwhile is looking to apply the tools it develops for Houndify and Collective AI to text and image-based conversational search, meaning it could end up driving the entire space forward.