Together AI has raised $800 million in a Series C funding round led by Aramco Ventures, valuing the company at more than $8 billion. The round also drew participation from Vista Equity Partners, General Catalyst, Emergence Capital, Nvidia, March Capital, Pegatron, and SentinelOne's S Ventures. Together AI says it now generates more than $1 billion in annual bookings and that usage of open-source models on its platform has tripled over the past year, according to data it cites from OpenRouter.
The valuation represents a sharp jump from its previous round. Together AI raised $305 million in a Series B in February 2025 at a valuation of roughly $3 billion, led by General Catalyst and Prosperity7, the venture arm of Aramco. Before that, it closed a $102 million Series A in November 2023 from Kleiner Perkins. The Information reported in March that the company was in talks to raise $1 billion at a valuation near $7.5 billion, underscoring the rapid pace of growth in the AI infrastructure market.
Together AI operates as a cloud platform for running open-source AI models, positioning itself between the hyperscalers and developers who want alternatives to closed systems from OpenAI or Anthropic. Its customers include Cursor, the AI coding tool; Cognition, the maker of the Devin AI software engineer; and Decagon, an AI customer support startup. The company says it will use the new capital to expand its compute infrastructure and accelerate development of its inference engine.
The founding team brings strong academic credibility. Co-founder and CEO Vipul Ved Prakash previously built Topsy, the social analytics company Apple acquired in 2013 for more than $200 million. Co-founders Percy Liang, a Stanford computer science professor, and Ce Zhang, who has held positions at ETH Zurich and the University of Chicago, bring deep research ties. The company was founded in 2022 and has quickly become a key player in the open-source AI ecosystem.
The raise lands in a market pouring capital into AI infrastructure at an unprecedented pace, with neocloud valuations soaring as demand for GPU compute continues to outstrip supply. Upscale AI closed a $190 million extension in June, bringing its total funding to $500 million at a $2 billion valuation. TensorWave, which builds its cloud on AMD chips rather than Nvidia's, raised $350 million in a Series B at a valuation near $1.5 billion. These rounds highlight the intense investor appetite for alternative cloud providers that can offer specialized AI compute services.
Together AI's open-source bet sets it apart from the neocloud pack. While rivals like Groq and RunPod focus primarily on renting out raw GPU capacity, Together AI bundles compute with its own inference optimization software, which it says can cut the cost of running popular models by up to 80%. That software layer is the moat the company is building around what would otherwise be a commodity hardware business. By optimizing model performance and reducing costs, Together AI aims to make open-source models more accessible to enterprises and developers alike.
The open-source AI movement has gained significant momentum in recent years, driven by the release of powerful models such as Meta's Llama, Mistral's models, and various fine-tuned versions from the community. Together AI's platform supports thousands of these models, allowing users to deploy them with managed infrastructure and optimized performance. The company's inference engine leverages techniques like flash attention, quantization, and speculative decoding to accelerate throughput while maintaining accuracy.
Together AI's success reflects a broader shift in the AI industry. While proprietary models from companies like OpenAI and Anthropic dominate the headlines, open-source models have quietly become the backbone of many enterprise AI deployments. Organizations value the ability to customize models, retain control over data, and avoid vendor lock-in. Together AI's platform addresses these needs by providing a scalable and cost-efficient way to run open-source models in production.
The funding round also signals growing Middle Eastern interest in the infrastructure underpinning artificial intelligence. Aramco Ventures leading the round is notable given the Saudi oil giant's deep pockets and strategic focus on diversifying beyond hydrocarbons. The kingdom has been investing heavily in AI through its sovereign wealth fund and various corporate arms, positioning itself as a hub for AI infrastructure investment. This aligns with similar moves by other Gulf states, such as the UAE's investment in AI through MGX and other entities.
Together AI's valuation of $8 billion places it among the most valuable private AI infrastructure companies. For comparison, CoreWeave, a leading neocloud provider, was valued at $19 billion in its 2024 funding round before filing for an IPO. Lambda, another GPU cloud provider, raised $320 million at a $1.5 billion valuation in 2024. Together AI's niche focus on open-source models and inference optimization gives it a differentiated position, but it faces stiff competition from both hyperscalers and well-capitalized rivals.
Hyperscalers such as Amazon Web Services, Google Cloud, and Microsoft Azure have been rapidly expanding their own inference capabilities. AWS offers SageMaker and Bedrock with optimization tools, Google Cloud provides Vertex AI with custom model serving, and Microsoft Azure has integrated Nvidia's GPUs and its own inference solutions. These platforms have massive scale, extensive developer ecosystems, and the ability to subsidize costs through other services. The question for Together AI is whether its open-source positioning and software edge can hold as the hyperscalers build out their own inference capabilities at a scale no startup can match.
Despite the competitive pressure, Together AI has demonstrated impressive growth. The company's annualized bookings exceeding $1 billion indicate strong demand from both startups and enterprises. The tripling of open-source model usage on its platform reflects a broader trend toward open ecosystems in AI. Developers are increasingly turning to open-source models not only for cost savings but also for transparency, customization, and community support.
Together AI's platform architecture is designed to maximize efficiency. It uses a multi-tenant GPU cluster with dynamic resource allocation, allowing customers to share compute while maintaining performance isolation. Its inference engine includes batching, caching, and model parallelism to reduce latency and increase throughput. The company also offers fine-tuning services and custom model deployment via APIs, making it easy for developers to integrate open-source models into their applications.
The company's leadership team has deep experience in both academia and industry. CEO Vipul Ved Prakash is a serial entrepreneur with a background in distributed systems and machine learning. Co-founders Percy Liang and Ce Zhang are respected researchers in natural language processing and systems. Liang, a Stanford professor, leads the Center for Research on Foundation Models and has published extensively on model evaluation and robustness. Zhang, a professor at ETH Zurich and previously at the University of Chicago, specializes in machine learning systems and data management. Their academic credentials lend credibility and attract top talent to the company.
Together AI was born out of the Stanford AI Lab and has maintained close ties with the research community. The company employs several researchers who contribute to open-source projects and collaborate with academic institutions. This research-driven approach helps Together AI stay at the forefront of inference optimization and model serving technology, which is critical for maintaining its competitive edge.
The $800 million Series C round is one of the largest in the AI infrastructure space this year. It brings Together AI's total funding to over $1.2 billion, providing substantial runway for expansion. The company plans to use the capital to procure additional GPU capacity, build out data centers, and hire engineering talent. It also intends to invest in its software stack, adding support for more model architectures and improving ease of use.
Looking ahead, Together AI faces several challenges. The rapid pace of model development means that inference optimization techniques must constantly evolve. New architectures like mixture-of-experts and multimodal models require different serving strategies. Additionally, the supply of Nvidia GPUs remains tight, and any disruption could impact the company's expansion plans. To mitigate these risks, Together AI has diversified its hardware partnerships and is exploring support for AMD and other chip vendors.
Despite these challenges, the company's trajectory has been remarkable. From its founding in 2022 to reaching $1 billion in annual bookings and an $8 billion valuation in just three years, Together AI has capitalized on the explosive growth of generative AI. Its success serves as a case study in how a startup can thrive by focusing on a specific niche within a massive market. By championing open-source models and providing a superior inference experience, Together AI has carved out a valuable position in the AI infrastructure ecosystem.