US-based Together AI, a cloud platform designed for running open-source artificial intelligence models, has secured an $800 million Series C funding round. The investment was led by Aramco Ventures and lifts the company’s valuation to a formidable $8.3 billion.
Quick Facts
- Funding Round: $800 million Series C
- New Valuation: $8.3 billion
- Lead Investor: Aramco Ventures
- Key Backers: Nvidia Corp., Vista Equity Partners, General Catalyst
Aramco Ventures and Nvidia Bet on Open-Source Infrastructure
The significant investment from Aramco Ventures signals strong MENA-based confidence in the growing market for open-source AI infrastructure. The round also saw participation from chipmaking giant Nvidia Corp., along with Vista Equity Partners, General Catalyst, and other institutional backers, underscoring the strategic importance of platforms that enable developers to build and deploy AI applications outside of closed ecosystems.
This deal places Together AI among the most valuable private companies in the AI infrastructure sector, reflecting a broader market shift where businesses are seeking more control, flexibility, and cost-efficiency than proprietary models from providers like OpenAI and Anthropic can offer.
Powering the Open-Source AI Shift
Founded in 2022, Together AI provides a comprehensive platform for developers to train, fine-tune, and run open-source AI models. The company’s core offerings include a serverless inference service that simplifies the deployment process by eliminating the need for users to manage complex hardware configurations. The company claims this service delivers about twice the performance of its closest competitor.
The platform also provides dedicated infrastructure for clients needing higher reliability and customization, as well as a cost-focused “Batch Inference” service that offers up to a 50% price reduction for workloads that are not time-sensitive.
Under the Hood: The ATLAS Engine
Together AI’s performance advantage is driven by its custom software engine, ATLAS, which runs on Nvidia hardware. The engine utilizes a technique called speculative decoding, where a smaller, lightweight AI model rapidly generates a draft response to a user prompt. The main, more powerful model then reviews, corrects, and finalizes the output. This two-step process is significantly faster than relying on the large model alone.
The company claims its ATLAS technology can speed up certain inference tasks by as much as 400% by automatically adapting the lightweight model to user requirements, overcoming a common accuracy decay problem in other systems.
Bookings Top $1.15 Billion Amid Surging Demand
The company’s rapid growth is backed by strong financial performance, with annual bookings surpassing $1.15 billion in the second quarter. Together AI serves several thousand organizations, including prominent names like LG’s AI research lab, Cohere Inc., and the Mozilla Foundation. Customers have reported significant cost savings, with Decagon co-founder Ashwin Sreenivas stating that moving to Together AI cost one-fifth to one-seventh of what closed models charged.
With the new capital, Together AI plans a massive expansion of its public cloud infrastructure, aiming for a 50-fold increase in capacity over the next five years. The funds will also be used to enhance the platform’s training and inference features.
About Together AI
Founded in 2022, Together AI is a cloud platform optimized for developing and running open-source artificial intelligence models. The company provides a full-stack solution, including serverless inference, dedicated clusters, and fine-tuning capabilities, to help organizations build and scale AI applications efficiently and cost-effectively.
Source: SiliconANGLE
