Deepgram Secures $130M Funding at $1.3B Valuation, Acquires YC Startup

Deepgram, a San Francisco-based company that builds tools for voice artificial intelligence, announced on Tuesday it raised $130 million in a Series C funding round at a $1.3 billion valuation. The money comes from a group of investors led by AVP, and it will help the company buy a Y Combinator-backed startup called OfOne while growing its work in real-time voice technology.

Background

Deepgram started in 2015 as a way to make computers understand and create human speech better. Its founders saw that voice would become a big part of how people talk to machines, from phones to customer service lines. Over the years, the company built APIs, which are sets of tools developers use to add voice features to apps. These include turning speech into text, text into speech, and handling talks where people interrupt each other.

The rise of AI has pushed voice tech forward. More businesses want to use voice for sales calls, support chats, and even taking orders at fast food spots. Deepgram powers parts of these systems for over 1,300 companies. Some known users include Granola, which takes notes in meetings, Vapi for voice agents, and Twilio for phone services. Last year, Deepgram became cash flow positive, meaning it made more money than it spent. Even so, leaders decided to take new money to grow faster.

Voice AI faces real-world tests. For example, Taco Bell tried voice ordering but stopped after mix-ups like a customer asking for 18,000 cups of water. Deepgram aims to fix issues like that with tech that handles noise, speed, and back-and-forth talk.

Key Details

The Series C round brings Deepgram's total funding to over $215 million. AVP, a firm that invests in tech growth across Europe and North America, led the deal. Existing backers like Alkeon, In-Q-Tel, Madrona, Tiger, Wing, Y Combinator, and BlackRock funds all joined in. New investors include Alumni Ventures, Princeville Capital, Twilio, ServiceNow Ventures, SAP, Citi Ventures, University of Michigan, Columbia University, and Stanford University.

Deepgram plans to use the cash for several steps. First, it bought OfOne, a Y Combinator startup that makes voice AI for quick-service restaurants. OfOne claims over 93% accuracy in taking orders, which could help chains avoid past errors.

Tech and Patents

Deepgram's main products stand out for speed and reliability. Nova-3 turns speech to text in real time with top accuracy. Aura-2 creates natural-sounding speech from text at low cost for big companies. Flux handles interruptions in conversations, a key problem for voice agents.

The funding will grow its patents, filed since 2016. Recent U.S. grants include one for end-to-end speech recognition using transformers, speeding up and improving accuracy. Another covers hardware tricks to cut delays and costs on huge voice data loads. A third protects ways to search and classify audio using deep learning states.

"As we rapidly approach a world where billions of simultaneous conversations are powered by Voice AI, enterprises and developers need real-time, reliable infrastructure capable of fully duplex, contextual conversations at scale – this is Deepgram," said Scott Stephenson, CEO and co-founder of Deepgram.

The company also launched 'Powered by Deepgram,' a program for partners, and opened a Voice AI Collaboration Hub in San Francisco.

What This Means

This raise puts Deepgram in a hot spot amid voice AI growth. Other firms like Seasame got $250 million, ElevenLabs $180 million, and Gradium $70 million last year. Investors see voice as a trillion-dollar market for business tools, much like payments tech changed money handling.

For Deepgram, the money means pushing into Europe and Asia-Pacific, adding languages, and targeting restaurants. It will build better models to pass what they call the Audio Turing Test by 2026, where AI voices fool humans. Enterprises get cheaper, faster voice interactions, cutting costs in call centers and sales.

Strategic backers like Twilio and SAP bring ties to big customers. Schools like Stanford investing show faith in the tech's future. Deepgram's focus on real-time, low-latency talk sets it apart as voice goes mainstream.

The acquisition of OfOne tests voice AI in tough spots like noisy kitchens. If it works, more chains might adopt it, changing how people order food. Overall, this funding speeds up a shift where voice becomes the main way machines talk to us, powering everything from support bots to global meetings.