Character.AI, a full stack conversational artificial intelligence (AI) platform that gives consumers access to their own deeply personalized superintelligence, today announced a strategic partnership with Google Cloud and has selected it as its preferred cloud infrastructure to build and train sophisticated AI models with advanced reasoning and greater accuracy.
The field of generative AI, particularly large language models (LLMs), has made remarkable advancements in recent months and has proven to be a game-changer in AI by serving as the foundational model that outperforms previous algorithms in most natural language and text generation tasks. Character.AI is using Google Cloud generative AI and LLM infrastructure to meet the needs of its rapidly growing community of creators. By combining its own AI capabilities with those of Google Cloud, Character.AI will enhance the customer experience by inspiring imagination, discovery and understanding.
As part of the partnership, Character.AI will use Google Cloud’s Tensor Processor Units (TPUs) to train and infer LLMs faster and more efficiently. Google Cloud TPUs are purpose-built to accelerate AI training and inference, speeding up the most computationally-intensive workloads with more than 100 petaflops of performance in a single pod. By leveraging Google Cloud TPUs, Character.AI can take advantage of the same reliable, secure, and highly scalable infrastructure that powers Google Search, YouTube, and other Google services. Character.AI will also employ Google Cloud’s new A3 VMs running on NVIDIA H100 Tensor Core GPUs, a combination that uniquely enables customers to tackle any AI workload with speed and flexibility and provides customer optionality between TPUs and GPUs to meet their needs.
“We’ve recognized the power and strength of Google Cloud’s technology from day one,” said Noam Shazeer, CEO, Character.AI. “As we continue our growth trajectory, working with Google Cloud’s AI technologies was the obvious choice, allowing us to rapidly expand our compute abilities so we can deliver new features and capabilities to millions of users.”
Character.AI today uses Google Cloud for core infrastructure and is leveraging Google Cloud’s PostgreSQL-compatible database, AlloyDB, for increased operational uptime, speed, and flexibility. Looking ahead, Character.AI will soon deploy Google Cloud Spanner, a high availability and high performance database, to help scale its service to additional regions.
“We’re offering Google Cloud’s industry leading infrastructure, Google foundation models and AI tooling to companies across industries so they can build, train and deploy the future of AI creatively, reliably and at scale,” said Thomas Kurian, CEO, Google Cloud. “Character.AI’s cutting-edge conversational AI technology is creating entirely new opportunities to transform how we interact with AI systems, and we are thrilled to be their partner as they continue to scale their vision.”
Character.AI was founded by Noam Shazeer and Daniel De Freitas, who are two of the world’s foremost experts in conversational AI. Noam’s previous work is central to the current revolution in LLMs, while Daniel’s is related to building large-scale NLP and deep learning programs. Bringing together their expertise with Google Cloud’s cutting-edge technology will allow Character.AI to scale and bring this intelligence to consumers across the world at scale.
Join the growing community of millions who are experiencing the value of personalized AI at https://beta.character.ai. To learn more about the AI infrastructure powering the Character.ai platform, visit cloud.google.com/tpu.
Stay Ahead of the Game with MTC Podcast, Your Go-To Source for Cutting-Edge Martech Insights, the Latest Martech News, and Expert Updates from Top Business Leaders!