AI NewsFunding & InvestmentJuly 1, 20265 min read

Together AI Raises $800 Million to Scale Open Source AI

Together AI Raises $800 Million to Scale Open Source AI

Together AI just raised one of 2026's largest AI infrastructure rounds, and the lead investor is a Saudi oil giant betting that open models, not closed APIs, will run enterprise AI.

Key Takeaways

  • 1Together AI raised 800 million dollars in a Series C led by Aramco Ventures at an 8.3 billion dollar post money valuation.
  • 2The valuation more than doubled from 3.3 billion dollars in February 2025, and the company reports annual bookings above 1.15 billion dollars.
  • 3The round signals that investors now price open weight inference platforms as core production infrastructure, not a cheaper experiment.

Together AI has raised 800 million dollars in a Series C round that values the company at 8.3 billion dollars, more than doubling its worth in under 18 months. The financing was led by Aramco Ventures, the venture arm of Saudi Arabia's state oil company, and confirmed by Reuters on July 1, 2026.

The San Francisco company does not build its own frontier model. Instead it runs the cloud infrastructure that lets enterprises train and serve open weight models cheaply, a wager that just attracted one of the largest AI infrastructure checks of the year.

What Together AI Announced

Together AI said the 800 million dollar Series C came at an 8.3 billion dollar post money valuation, according to the company's Business Wire announcement. The round drew Vista Equity Partners, General Catalyst, Emergence Capital, Nvidia, Salesforce Ventures, March Capital, Pegatron and SentinelOne's S Ventures.

Chief executive Vipul Ved Prakash framed the mission around cost, saying the goal is to make intelligence abundant, not expensive. That positioning matters because it targets the single biggest complaint from teams building on AI, namely that closed model pricing can consume an entire product margin.

Why Aramco and Nvidia Are Backing Open Models

The lead investor choice is the story within the story. The Next Web reported that Aramco Ventures leading the round signals growing Middle Eastern interest in the infrastructure underneath artificial intelligence rather than the models themselves.

Abhishek Shukla, a managing director at Aramco's Prosperity7 program, called AI infrastructure the biggest build in human history in the company statement. Nvidia's participation is notable too, since the chipmaker is backing a customer that resells access to the very GPUs it sells, tying its hardware to the open model ecosystem.

The Numbers Behind the Round

The valuation jump is steep. Together AI was valued at roughly 3.3 billion dollars in a February 2025 Series B led by General Catalyst, so the new figure represents more than a doubling in about 16 months, a pace Tech Startups tracked in its July 1 funding roundup.

The company told Let's Data Science that annual bookings crossed 1.15 billion dollars last quarter, that customers include Cursor, Cognition and Decagon, and that it has secured commitments for more than 500 megawatts of compute. Together AI plans to grow its infrastructure footprint roughly 50 fold over five years.

Reuters reported that the company will use the capital to expand into a full provider of inference, the process of actually running trained models in production. That focus is deliberate, because inference is the recurring cost that determines whether an AI product makes money once it is live, and it is where open weight models can undercut closed systems most sharply. Founded in 2022, Together AI reached a 1.25 billion dollar valuation in March 2024 before the 3.3 billion dollar mark a year later, so this round caps a rapid climb tied directly to the surge in enterprise AI deployment.

Why It Matters for AI Buyers

The round is really a price signal about open weight economics. Enterprises running production workloads are moving to open models such as DeepSeek, Nemotron, MiniMax and Kimi precisely because closed frontier APIs can eat margins, a tension that has shaped the wider debate over AI compute economics.

TechFundingNews noted that Together AI competes with Fireworks AI, which raised 250 million dollars in October 2025, and with Groq, which took a different path by licensing chip technology to Nvidia before rebuilding as an inference cloud. What separates Together AI is its bundling of raw compute with inference optimization software the company says can cut the cost of running popular models sharply.

What to Watch Next

The constraint to watch is capacity, not capability. Together AI's promise to expand roughly 50 fold depends on securing power and GPUs at a moment when hyperscalers are racing to add AI serving capacity as fast as they can build it.

The other signal is strategic capital. With Aramco leading and Nvidia, Pegatron and Schneider Electric's venture fund all in the syndicate, this round reads less like a bet on one company and more like energy, hardware and cloud players buying a seat in the open model supply chain. Quartz framed the deal as confirmation that the plumbing of AI, not the flashiest model, is where a growing share of 2026 capital is flowing.

What Changed

Together AI closed an 800 million dollar Series C on July 1, 2026, led by Saudi Aramco's venture arm and backed by Nvidia, more than doubling its valuation to 8.3 billion dollars. The company sells cloud infrastructure for training and running open weight models rather than a proprietary frontier model of its own.

Why It Matters

The round is a bet that enterprises will keep shifting production workloads to open models like DeepSeek and Kimi to escape closed API pricing that can erase software margins. Aramco leading the round also shows energy and compute capital moving directly into the AI infrastructure layer.

Suggested Actions

Operators running production AI should benchmark open weight inference on a platform like Together against their current closed API bill, since the gap is now large enough to change unit economics. Track whether committed compute capacity, not model quality, becomes the real constraint on switching.

Related Tags

Related News