Character AI Lands $150M in Cash and a Big Valuation

Andreessen Horowitz leads a monster round based on strong user growth

Andreessen Horowitz led a $150 million Series A funding round for Character AI, the company that enables anyone to create an interactive chatbot with a novel personality or one that mimics celebrities, tycoons, or cartoon characters. The company became well-known for its chatbots of Elon Musk, Mark Zuckerberg, and Tony Stark chatbots. Today, it has a $1 billion valuation just seven months after its founding.

Character AI launched in September and has been in open beta since then. Vociebot’s Eric Schwartz writes:

Character AI plans to invest in its core technology with the new capital, preparing to officially launch after coming out as a beta in September, a year after AI researchers Noam Shazeer and Daniel De Freitas founded the company.

Character AI hosts a conversational generative AI hub with chatbots mimicking a wide variety of real and fictional personalities as famous or personal as the designer chooses. The LLM filtered through that personality can then help users with specific services or general assistance depending on their choices.

100M Visits

The big valuation is based on proven interest. Character AI says it netted 100 million website visits in two recent months that accessed 2.7 million characters for an average of over two hours per day. That suggests somewhere north of 80 million hours of use per month.

Schwartz says, “The startup began by releasing a preview of the next iteration of its AI, called C1.2, which it claims is better at providing information, entertainment, and even sympathy.”

Self-Service Entertainment and Personalization

Sarah Wang, a general partner at Andreessen Horowitz, commented in a blog post about the investment:

Character.AI has trained their proprietary LLM from scratch, enabling their product to optimize for not only raw intelligence, but also a conversational empathy that captures and holds the consumer’s attention with humor, emotion, insight, and more. As our partners have written about extensively, we believe that end-to-end apps such as Character.AI, which both control their models and own the end customer relationship, have a tremendous opportunity to generate market value in the emerging AI value stack.

…

For nearly two decades, co-founders Noam Shazeer and Daniel De Freitas have been pivotal in the advancement of conversational AI and LLMs. The AI industry would not be where it is today without their massive contributions. Noam co-authored the definitive paper on the transformer algorithms that power LLMs and is one of the few people in the world who has touched every single building block needed to train a large language model. He pioneered many of Google’s AI breakthroughs over the last twenty years, including the language understanding models that became the basis of Google’s content advertising system. Daniel’s focus on conversational AI goes back even further to when he built his first chatbot at the age of seven. During his time at YouTube, and later Google Brain, Daniel created the Meena chatbot and led efforts on the LaMDA (Language Model for Dialogue Applications) model, the underlying technology for Google’s Bard.

Productivity applications are getting a lot of attention right now. However, generative AI may have an even bigger impact on entertainment and personalization. Immersive, interactive experiences are now far more practical than ever before. We know people will spend hours learning from their interactions with ChatGPT. They will also do this with an Elon Musk chatbot, as Wang outlines in her post.

In addition, personalization has been an elusive promise of the internet era. We see it on YouTube and Spotify, and a few other places. However, it is hard to do and to scale across services. Generative AI enables this at scale in a way that was previously impractical. Your life coach could be Elon Musk today, Princess Leia tomorrow, and Socrates next week. It can get to know you as well.

Synthetic Media is Just Getting Started

Character AI has some powerhouse founders and has clearly captured interest form everyday consumers. Key questions for the company now are whether the solution will be sticky and how it will generate revenue. Several anecdotes suggest this may be possible.

However, what is really going on here is that Character AI represents a class of large language model based solutions that are captivating consumer interest and are likely to drive a key shift in the attention-based economy.