Together AI — an AI cloud platform that enables companies to train and deploy artificial intelligence models — has raised $305 million in Series B funding in a round led by General Catalyst, more than doubling its valuation to $3.3 billion from $1.25 billion last March. The funding comes amid growing demand for computing power to run advanced open-source models.
“We have built a cloud company for this AI-first world — combining state-of-the-art open source models and high performance infrastructure, with frontier research in AI efficiency and scalability,” Together AI CEO Vipul Ved Prakash said in the company’s announcement.
The San Francisco-based startup has seen explosive growth since its founding in 2022. According to Bloomberg, the company recently surpassed $100 million in annualized revenue, up from $30 million in February 2024.
Together AI’s funding announcement comes as the company reports increasing infrastructure demands for running advanced AI models. While some had speculated that newer models like DeepSeek-R1 might require less computing power, Together AI says it is seeing the opposite trend, stating that DeepSeek-R1 requires distribution across multiple servers and typically handles longer requests lasting two to three minutes.
Growing Demand
To meet growing demand, Together AI has secured 200 megawatts of power capacity and announced plans to deploy clusters of Nvidia’s latest Blackwell GPUs across North America. The company has partnered with Hypertec to build a cluster of 36,000 Nvidia GB200 NVL72 GPUs.
As for its real-world performance, Together AI claims that its platform can process DeepSeek-R1 queries at 85 tokens per second compared to Azure’s seven tokens per second, though such performance metrics can vary based on specific use cases and configurations. The company states that while new Blackwell chips cost approximately 25% more than previous generations, they deliver double the performance.
Currently serving over 450,000 AI developers and supporting more than 200 open-source models across various modalities — including chat, image, audio, vision, code, and embeddings — the company’s notable customers include Salesforce, Zoom, SK Telecom, and The Washington Post.
For organizations concerned about data security, Together AI also offers deployment options in virtual private cloud (VPC) and on-premises environments — a standard requirement for enterprise AI deployment. This allows businesses to maintain their existing privacy and compliance policies while accessing high-performance computing resources.
The funding round attracted significant interest from major players in the AI ecosystem, with participation from Salesforce Ventures, Nvidia, Kleiner Perkins and others. Saudi Arabia’s Prosperity7 Ventures co-led the round alongside General Catalyst.
The Broader Context
The rise of open-source models like DeepSeek represents a significant shift in the AI landscape — its emergence has pushed companies to rethink traditional approaches to AI deployment, particularly around computational efficiency and infrastructure requirements.
Together AI’s expansion also comes at a time when major cloud providers like Microsoft’s Azure are competing to provide infrastructure for increasingly sophisticated AI models. The company’s $3.3 billion valuation and substantial GPU deployment plans reflect a broader industry trend — despite the growth of efficient open-source models, the demand for AI computing infrastructure appears to be increasing.
The challenges ahead, however, remain significant. Together AI faces competition from both established cloud giants with extensive resources and other specialized AI infrastructure providers. As enterprise adoption of AI accelerates, success in this market will likely depend on delivering both performance and cost efficiency, while addressing enterprise requirements around privacy and security.
