San Francisco-based AI Acceleration Cloud provider Together AI has raised USD 305 million in a Series B funding round led by General Catalyst and co-led by Prosperity7, pushing its valuation to USD 3.3 billion. The round saw participation from a group of global institutional and strategic investors, including Salesforce Ventures, DAMAC Capital, NVIDIA, Kleiner Perkins, March Capital, Emergence Capital, Lux Capital, SE Ventures, Greycroft, Coatue, Definition, Cadenza Ventures, Long Journey Ventures, Brave Capital, Scott Banister, and John Chambers.
With this investment, Together AI plans to expand its AI Acceleration Cloud, with large-scale deployments of Nvidia Blackwell GPUs. “This strategic investment will further accelerate Together AI’s position as the leading end-to-end platform for building with open source models,” the AI company said on February 20.
The platform, which offers enterprise-grade inference and fine-tuning capabilities, serves over 450,000 AI developers and global enterprises, including Salesforce, Zoom, SK Telecom, Hedra, Cognition, Zomato, Krea, Cartesia, and The Washington Post, enabling them to build AI applications.
“Together AI’s mission to be the full stack AI cloud is truly inspiring, and General Catalyst brings the go-to-market expertise and ambition to supercharge this goal,” said Marc Bhargava, managing director at General Catalyst.
Together AI claims to deliver the fastest DeepSeek-R1 and Llama inference at production scale through its infrastructure and research innovations.
“AI is transforming every industry, creating unprecedented efficiencies and enabling entirely new classes of products. 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,” said Vipul Ved Prakash, CEO of Together AI.
“Our AI Acceleration Cloud uniquely provides organizations with the performance, security, and functionality required to train frontier models and build production-scale AI applications with incredible cost efficiency. With this investment, we will accelerate our mission to make open-source AI accessible for AI developers and customers globally,” Prakash added.
According to the official release, Together AI’s platform spans the AI lifecycle, delivering enterprise-grade inference solutions, training and fine-tuning for frontier foundational models, agentic workflows with built-in code interpretation, and synthetic data generation.
Supporting over 200 open-source models across all modalities—chat, image, audio, vision, code, and embeddings—the platform is powered by Together AI’s proprietary inference engine and built on FlashAttention-3 kernels and advanced quantization techniques. It delivers 2-3x faster inference than today’s hyperscaler solutions, the company stated.
Together AI has secured 200 MW of power capacity and is deploying optimized Nvidia Blackwell GPU clusters across North America. In partnership with Hypertec, it is also co-building a 36,000 Nvidia GB200 NVL72 GPU cluster.
Additionally, Together AI’s Kernel Collection, a proprietary technology stack, delivers 24 percent faster training operationswhile significantly reducing costs for customers like Pika Labs.
“Recent developments in open source AI illustrate the importance of Together AI’s work in making access to models like these secure, accessible, and powerful for organizations here in the US and around the world,” said Prosperity7 Ventures Managing Director Abhishek Shukla. “Together AI’s ability to do this while accelerating the process, increasing efficiencies and reducing costs makes Together AI an exceptional resource for even large, global businesses.”
Furthermore, Together AI’s research lab continues to drive innovation with open-source contributions like Mixture of Agents, Medusa, Sequoia, Hyena, and Mamba.
In 2024, Together AI grew its user base to over 450,000 AI developers and collaborated with Dell, Hypertec, Nvidia, and Meta. It deployed DeepSeek models in North American data centers with full opt-out privacy controls, launched the Together Enterprise Platform and AWS Marketplace availability, and partnered with Cartesia to enable ultra-low-latency voice AI through Sonic model integration.
Additionally, Together AI recently acquired CodeSandbox and plans to integrate its capabilities directly into its platform for built-in code interpretation—an industry first, the company said.
Support Telecomtalk
