Together AI has raised $800 million in a Series C round led by Aramco Ventures, more than doubling the AI infrastructure startup's valuation to $8.3 billion. Announced July 1, 2026, the deal is one of the year's most telling funding signals — not because of its size alone, but because of who wrote the checks and what they are betting on: a future where enterprises run AI on open models rather than closed ones.
The Deal and the Backers
Founded in 2022, Together AI operates a cloud platform that lets companies train and run AI workloads on open models — including DeepSeek, MiniMax, and Kimi — at lower cost than closed, proprietary systems. That value proposition has drawn an unusually strategic investor roster.
Beyond lead investor Aramco Ventures, the venture arm of Saudi oil giant Saudi Aramco, the round pulled in Vista Equity Partners, General Catalyst, Emergence Capital, Nvidia, Salesforce Ventures, March Capital, Pegatron, and SentinelOne's S Ventures. The mix of an energy major, the dominant AI chipmaker, a leading enterprise software vendor, and a hardware manufacturer reads like a cross-section of the entire AI supply chain placing a coordinated bet.
The valuation trajectory underscores the momentum. Together AI was valued at $3.3 billion in February 2025 in a round led by General Catalyst, which itself had more than doubled a $1.25 billion mark from March 2024. The new raise doubles the figure again in under 18 months.
Why Energy and Chips Are Buying In
The investor list is the story. Aramco's involvement signals that energy and cloud players want a direct stake in the enterprise AI platform layer — a logical hedge for companies whose core business is the power and infrastructure that AI voraciously consumes. Nvidia's participation extends its now-familiar pattern of investing across the ecosystem that buys its chips, while Salesforce Ventures' presence points to enterprise software's appetite for flexible model deployment.
Together AI's pitch lands squarely on a live enterprise anxiety: vendor lock-in and cost. Running frontier workloads through a single closed API is expensive and inflexible. An open-model cloud promises governance, model portability, and lower unit economics — precisely the priorities CFOs are elevating as AI moves from experiment to line-item.
A Market Splitting in Two
The raise fits a broader 2026 pattern in which capital is concentrating where AI meets real operational leverage — infrastructure, agentic systems for regulated workflows, and specialized models. Investors are increasingly rewarding companies with hard revenue metrics and defensible positioning rather than consumer experimentation.
That discipline is reshaping the funding landscape:
- Infrastructure and platforms — cloud, inference, and model-serving companies like Together AI are commanding premium valuations as the "picks and shovels" of the AI economy.
- Agentic and vertical AI — startups automating high-stakes, regulated workflows are drawing strategic capital from incumbents in finance and enterprise software.
- Open-model tooling — the ecosystem enabling enterprises to run and customize open weights is emerging as a distinct, well-funded category.
Why It Matters
Together AI's raise is a vote for a multi-model, open future at a moment when the narrative is often dominated by a handful of closed frontier labs. If the largest energy, chip, and software players believe enterprises will increasingly run AI on open models, that reshapes assumptions about where value accrues in the stack — away from any single model provider and toward the infrastructure that makes many models usable, portable, and affordable.
For enterprises evaluating their own AI strategies, the signal is worth reading. The companies closest to the economics of AI — those selling the power, the chips, and the software it runs inside — are backing flexibility over exclusivity. That suggests the smart long-term posture is not to bet everything on one lab's roadmap, but to build on infrastructure that keeps options open as models improve and prices fall.
It also reinforces a defining theme of the 2026 AI economy: deployment, not model capability, is the differentiator. Enterprises rarely fail because a model is too weak; they fail because they cannot wire it economically into messy, real-world workflows. Platforms that lower the cost and friction of running the best available model — whichever one that happens to be this month — are becoming the layer businesses actually pay for. Together AI's $8.3 billion valuation is a wager that this layer, not the models themselves, is where durable enterprise value will settle.
