Character.ai's founders, Noam Shazeer and Daniel De Freitas, returned to Google with part of their research team. This return not only includes the two founders, but also brings 30 employees responsible for model training and speech AI. They will join Google's Gemini AI project and participate in broader generative AI R&D.
To facilitate the team's return, Google paid a hefty $2.5 billion to obtain a non-exclusive license to Character.AI's large language model technology. Notably, this valuation is much higher than Character.AI's previous valuation of $1 billion, although it is only half of the company's peak valuation of $5 billion.
Still, the majority of Character.AI's team will remain with the company to serve its growing user base. Currently, Character.AI has 4 million mobile users and 14.8 million monthly website visitors.
The whole incident shows the fierce competition and reshuffle in the current AI startup field. With the departure of the core team, the future of Character.AI is uncertain.
Google's $2.5 billion restructuring of Character.AI team: a recap
2021年,Noam Shazeer和Daniel de Freitas因不满谷歌的官僚作风,创立 Character.AI。 二人曾在谷歌领导构建LaMDA,Noam Shazeer更是Transformer论文的主要作者之一。
Character.AI quickly attracted users and investors after its establishment. In March 2023, it completed a $150 million financing at a valuation of $1 billion, led by A16Z and followed by other well-known investors.
However, after 2023, the company began to struggle. Although it has 20 million users, it is difficult to monetize, with only $200,000 in mobile revenue in July 2024 and an estimated $16.7 million in revenue for the full year. At the same time, the cost of self-built models is high, the capital chain is broken, and there are frequent news of poor financing.
In August 2024, Google paid Character.AI investors equity value at a valuation of $2.5 billion for a non-exclusive license to its large language model technology. Noam Shazeer and Daniel De Freitas return to Google DeepMind with 30 employees.
Character.AI临时CEO由总法律顾问Dominic Perella担任。
Character.AI will continue to operate and move to using open-source models such as Meta's Llama 3.1. Google pledged more funding to support Character.AI continue to build personalized AI products for users around the world. Previously, Character.AI open-sourced its core technology, Prompt Poet, to simplify the prompt design and management process.
Are AI chatbots a good business?
In its heyday, Character.AI became all the rage by allowing users to interact with personalized chatbots, attracting a large number of younger users.
However, the high cost of training conversational AI models and the low percentage of paying users make it difficult to stay afloat.
To survive, Character.AI has negotiated with several tech companies, including Meta and xAI. Eventually, Google helped Character.AI tide over the storm by providing cloud computing services and financing, and signed a partnership agreement.
In addition, due to strict regulation, many users' favorite interactive content has been blocked, which has further affected users' willingness to pay. Angry users even set off a wave of protests on social media, saying that AI models were emasculated as "wasters."
At the same time, Meta also announced the cessation of the celebrity AI chatbot project, showing the plight of AI chatbots in the current environment.
The future of AI chatbots still needs to be explored, and how to balance cost, user experience, and regulation is a key challenge.
A wave of reshuffle for AI start-ups
Since the beginning of this year, AI start-ups have faced increasingly fierce competition, and many star companies have had to "sell" large companies in order to survive. Companies such as Character.AI, Inflection and Adept have all opted to license their technology to tech giants, and founders and core teams have joined these big companies.
Microsoft paid $650 million in March to bring in the founders and team of Inflection to form a new division.
Amazon, on the other hand, recruited several of Adept's co-founders and employees in June.
类似的，Stability AI和Perplexity也有被收购的传闻流传。
Industry insiders point out that many AI start-ups do not have a stable source of income and struggle to maintain operations. Although innovative projects continue to emerge, they are severely burned and difficult to monetize, resulting in high valuations and difficult financing.
Funding and resources in the AI field are gradually concentrating on leading companies, and many startups can only survive through acquisitions. Faced with the resources and technological advantages in the hands of giants, it is difficult for start-ups to survive on their own under the pressure of capital and profitability.
Of course, there are many small and beautiful companies that live well, and AI software that can really find pain points, solve problems, and improve efficiency, while balancing costs and benefits, is the general trend of the future.
If you find this article helpful, please like, bookmark, retweet and share. In the meantime, please follow me for more updates and insights on artificial intelligence!
Reference: