Together AI Raises $800 Million Series C at $8.3B Valuation as Open-Source Inference Demand Surges

The open-source AI infrastructure market just received its clearest vote of confidence yet.

Together AI announced on July 1, 2026, that it has closed an $800 million Series C funding round, vaulting the company’s valuation to $8.3 billion. The round was led by Aramco Ventures, the investment arm of Saudi Aramco, and included participation from Vista Equity Partners, General Catalyst, Emergence Capital, Nvidia, March Capital, Pegatron, and SentinelOne’s S Ventures, among others.

The raise is the largest in Together AI’s history and more than doubles the valuation from its previous $305 million Series B round at $3.3 billion, closed roughly 16 months ago. More strikingly, annual bookings have now surpassed $1.15 billion, suggesting the company has reached a scale where its infrastructure has become genuinely mission-critical for a wide range of enterprise AI deployments.

What Together AI Does

Together AI operates what the industry calls an AI neocloud: a hyperscale GPU compute provider built from the ground up for AI workloads, as opposed to traditional cloud providers that retrofitted AI capabilities onto general-purpose infrastructure. The company rents Nvidia GPU clusters and provides inference APIs specifically optimized for open-source models — including Meta’s Llama series, Mistral’s models, DeepSeek-R1, Qwen, and dozens of other open-weight models.

The pitch to enterprises is straightforward: open-source models have reached or exceeded closed-model quality on many benchmarks, but running them at production scale requires massive GPU infrastructure, model optimization expertise, and a reliable serving layer that most companies cannot build internally. Together AI packages all of that as a managed service, letting enterprises capture the flexibility and cost advantages of open-source AI without the infrastructure burden.

Why the Market Is Moving This Way

The surge in Together AI’s bookings reflects a structural shift in enterprise AI strategy that has been building for two years. When GPT-4 first launched, the gap between closed frontier models and open-source alternatives was wide enough that most enterprises defaulted to API-based access from OpenAI, Anthropic, and Google. That gap has narrowed dramatically.

Llama 4, DeepSeek-R1, and Mistral’s latest models now deliver performance competitive with GPT-5.5 and Claude Fable on a wide range of practical tasks, at costs that are significantly lower when run on dedicated GPU infrastructure rather than metered API access. For enterprises running millions of inference calls per day — customer service systems, document processing pipelines, internal knowledge assistants, coding tools — the cost differential is substantial enough to justify the integration work required to switch.

Open-source infrastructure also offers something closed APIs cannot: data privacy and model customization. Enterprises in regulated industries — financial services, healthcare, legal — often cannot send sensitive data to third-party API endpoints. Running open-source models on dedicated infrastructure controlled by the enterprise eliminates that constraint. Fine-tuning on proprietary data, which closed model providers allow in limited forms, is unrestricted on open-weight models.

Together AI’s annual bookings surpassing $1.15 billion in the last quarter — compared to OpenAI’s $47 billion annualized run rate and Anthropic’s reported trajectory toward similar figures — positions the company as the leading independent infrastructure layer underneath the open-source AI ecosystem.

The Strategic Role of Aramco Ventures

The choice of Aramco Ventures as lead investor is notable and carries strategic implications beyond the capital itself. Saudi Arabia has been investing aggressively in AI infrastructure as part of its Vision 2030 economic diversification program, with significant data center investment commitments to multiple AI companies. Aramco Ventures’ lead position in Together AI’s round suggests a deeper relationship: Together AI’s infrastructure footprint is expected to expand significantly in the Middle East, potentially including data center deployments in Saudi Arabia.

For Together AI, an anchor investor with both capital depth and regional infrastructure ambitions provides a template for international expansion that pure financial investors cannot offer.

Infrastructure Plans: 50× Scale in Five Years

Together AI plans to use the Series C proceeds to scale its infrastructure footprint approximately 50 times over the next five years — a buildout that implies moving from a few thousand GPUs today to potentially hundreds of thousands. The company has not specified which GPU architectures it plans to prioritize, but the implication is a major bet on continued demand for Nvidia’s Blackwell and successor architectures, combined with emerging open alternatives.

The scale target also signals Together AI’s ambition to compete not just with other AI neoclouds like CoreWeave, Lambda Labs, and Vast.ai, but potentially with the major hyperscalers — AWS, Azure, and GCP — as enterprise AI workloads increasingly migrate toward open-source models and specialized inference infrastructure.

The Broader Open-Source Infrastructure Market

Together AI’s raise follows a period of intense activity in AI infrastructure broadly. Rival neocloud CoreWeave went public in March at a $23 billion valuation, providing a precedent for AI infrastructure companies achieving durable public market value rather than being absorbed into hyperscalers.

The infrastructure layer is increasingly being recognized as one of the most durable positions in the AI value chain. Model capabilities are converging at the top — the performance gap between frontier models from major labs is narrowing — but the compute infrastructure required to serve those models at enterprise scale remains an enormous operational challenge. Companies that solve that problem reliably, at scale, and for open-source models specifically, occupy a position that is difficult to replicate quickly.

With $1.15 billion in annual bookings, a fresh $800 million in capital, and a clear mandate to scale its infrastructure 50-fold, Together AI is positioned to be the defining open-source AI infrastructure company of this generation. The race to build the rails underneath the open-source AI revolution is on in earnest.