Together AI has announced raising $800 million in a Series C funding round, marking a new step in the company’s expansion as it builds the cloud infrastructure and high-performance computing capabilities needed to train, deploy, and scale next-generation AI models.
Together AI is developing a cloud platform that helps companies own their AI infrastructure instead of relying entirely on closed models. The company focuses on supporting the open AI ecosystem by providing scalable infrastructure for training, deploying, and running AI models.
The funding comes after strong growth for the company over the past period, with bookings exceeding $1.2 billion over the last year. Its usage also increased from 30 billion tokens per month to 400 trillion tokens.
Vipul Ved Prakash, co-founder and CEO of Together AI, said the company started four years ago based on the belief that AI should be open and widely available, rather than controlled by only a few companies.
He added that an entire generation of AI companies is realizing that building sustainable businesses cannot depend only on closed models, noting that companies using open infrastructure are achieving 6x to 20x cost savings with equal or better performance.
He also noted that Decagon reduced its inference costs sixfold after switching to Together AI.
Together AI’s platform relies on advanced research and engineering to improve the efficiency and performance of AI model deployment, through technologies such as FlashAttention, ThunderKittens, Together.compile, MegaKernels, and Adaptive Speculative Decoding.
The company also confirmed that it continues to build the open infrastructure layer supporting the next generation of AI companies, while further developing an end-to-end platform serving some of the fastest-growing companies in the sector.
The new investment follows an earlier investment in Together AI through Prosperity7 Ventures in 2024, as part of a broader commitment to supporting the infrastructure enabling the future of artificial intelligence.
Read the article in
