Together AI raises $800M Series C led by Aramco Ventures
Together AI, the San Francisco-based AI neocloud specializing in hosting open-source models like Llama, GLM, and DeepSeek on rented Nvidia GPU clusters, announced an $800 million Series C on July 1 at an $8.3 billion post-money valuation. The round was led by Aramco Ventures — the venture arm of Saudi Aramco — with participation from Nvidia, Vista Equity Partners, General Catalyst, Emergence Capital, March Capital, Pegatron, and SentinelOne's S Ventures.
The raise more than doubles the company's $3.3 billion valuation from its Series B just 16 months earlier. Together AI reports annual bookings exceeding $1.15 billion as of its last quarter, driven by enterprises shifting workloads from expensive proprietary model APIs to competent, far cheaper open-source alternatives hosted on neocloud infrastructure. Customers include Cursor, Cognition, and Decagon.
The funding highlights a structural shift in AI infrastructure economics. Open-source model usage has tripled in the past year according to data from AI gateway OpenRouter, as teams realize that for most production workloads, self-hosted models like GLM-5.2 or DeepSeek V4 deliver frontier-adjacent performance at 5-10x lower cost than closed-model APIs. Together AI's platform handles the complexity of deployment — autoscaling, inference optimization, and model hosting — so developers can focus on application logic.
Aramco Ventures leading the round is strategically significant for the MENA region. It signals that Saudi capital is positioning itself at the infrastructure layer of the AI economy, not just in consumer applications. This matters for regional builders: the same economics driving Together AI's growth apply to any MENA team that can deploy and manage open-source models locally.
Aramco Ventures leading this round is the real story for MENA builders. It means Saudi capital is buying infrastructure seats at the AI table — not just consuming models but owning the layer where compute happens. For regional teams, the open-source hosting economics that drove Together AI's growth are replicable: anyone who can deploy GLM-5.2 or DeepSeek on managed GPUs locally can undercut proprietary API pricing dramatically.
What is an AI neocloud?
An AI neocloud is a specialized cloud provider that rents GPU compute infrastructure optimized for AI workloads — particularly hosting and running open-source language models for inference, fine-tuning, and training.