Together AI raises $800M Series C with 500 MW compute commitments to accelerate open-source AI inference

Together AI has raised $800 million in a Series C funding round led by a consortium of strategic and financial investors, the company announced on July 1, 2026. The round brings Together AI fresh capital to expand its cloud infrastructure and deepen its focus on making open-source AI inference faster and more affordable.

What's new

The $800 million round drew investment from a wide field of backers: Aramco Ventures, NVIDIA, Vista Equity Partners, General Catalyst, Emergence Capital, Schneider Electric, Pegatron, Salesforce Ventures, March Capital, DTCP Growth, Lux Capital, Geodesic, and PSP Partners. Together AI did not disclose a post-money valuation.

Beyond the capital itself, the company secured more than 500 megawatts of committed compute capacity from its investor base — a strategic anchor that gives Together AI direct infrastructure leverage rather than relying purely on third-party cloud providers.

Key investor highlights:

• NVIDIA: a strategic chip-level backer, signaling alignment between Together's inference optimization work and NVIDIA's Blackwell roadmap
• Salesforce Ventures and General Catalyst: enterprise distribution reach
• Aramco Ventures and Schneider Electric: infrastructure and energy angles, both relevant to large-scale data center build-out

Context

Together AI has positioned itself as the primary inference provider for teams running open-source models — Llama, DeepSeek, Mistral, and similar families — at production scale. The company argues that proprietary frontier models are becoming unnecessary for many use cases as open-weight alternatives close the quality gap, and that the real competitive moat in AI is inference cost and speed, not model access.

CEO Vipul Ved Prakash framed the funding around a problem the company sees acutely among its customers: "As usage grows, inference bills compound faster than budgets, forcing companies to ration intelligence just as demand for it accelerates." The company cites Decagon, a customer service AI company, as having reduced its inference costs sixfold after switching to Together AI.

The company's earlier $305 million Series B (announced in 2024) helped it build out its Together Inference platform and publish research on speculative decoding and continuous batching. This Series C is considerably larger and comes at a moment when enterprise AI spending is shifting from experimentation to production deployment, driving demand for cost-efficient inference at scale.

Why it matters

The round is one of the largest single raises for an AI infrastructure provider focused explicitly on open-source models. It signals that the market for open-weight model inference is large enough to attract major capital — and that open-source is no longer a cost-cutting afterthought but a deliberate architectural choice for many enterprises.

The 500 MW compute commitment from investors is notable: that is enough power to run tens of thousands of high-end GPUs continuously, giving Together AI the ability to compete with hyperscalers on raw infrastructure capacity while maintaining its pricing advantage.

For the broader AI ecosystem, this raise creates a well-funded alternative to OpenAI and Anthropic inference at a moment when both are raising their own capital and building proprietary infrastructure. It also validates the open-source inference stack as a durable business — not just a research playground — and may accelerate the trend of enterprises moving workloads off proprietary APIs onto open-weight models where they control cost and data.