en.Wedoany.com Reported - Together AI has completed an $800 million Series C funding round, led by Aramco Ventures, the venture capital arm of Saudi Aramco, valuing the company at $8.3 billion. The round attracted participation from multiple institutions including Vista Equity Partners, General Catalyst, Emergence Capital, NVIDIA, March Capital, Pegatron, and S Ventures. The infrastructure provider, which focuses on open models, stated that its annualized bookings in the previous quarter exceeded $1.15 billion.
Together AI is dedicated to providing training and inference services for open models to enterprises and developers, supporting models including DeepSeek, Nemotron, MiniMax, and Kimi. The company claims that compared to the pricing of closed models with equivalent or better performance, customers can reduce inference costs by 6 to 60 times using its platform. For instance, customer Decagon reportedly saw a six-fold reduction in inference costs after migrating to Together AI. Currently, the company serves thousands of paying customers, including Cursor, Cognition, and Decagon, with the usage of open-source models across the industry tripling within 12 months.
This funding round reveals several major trends in the AI infrastructure sector. Inference is becoming the primary cost center for production-grade AI, and open models are demonstrating commercial appeal in cost reduction. The composition of the investor list is noteworthy, including not only venture capital but also energy-related capital and hardware ecosystem participants, indicating that AI infrastructure is increasingly intertwined with energy strategy and industrial policy. The lead investment by Aramco Ventures suggests that energy infrastructure—such as electricity, data center sites, and cooling required for computing capacity—is becoming a key constraint on AI platform expansion.
Together AI plans to use the funds to expand its products and features, enhance inference capabilities, and increase its infrastructure capacity and footprint by approximately 50 times over the next five years. CEO Vipul Ved Prakash views intelligence as a foundational resource similar to electricity or bandwidth, believing that the infrastructure market will ultimately reward scale, reliability, and distribution. However, for enterprise procurement teams, the decision-making process is more deliberate, with key inquiries focusing on security controls, data processing, model provenance, indemnification, regional availability, audit trails, and service-level commitments. In sectors such as finance, healthcare, energy, and public services, regulators may also pose more stringent questions.
It should be noted that performance comparisons largely depend on workload, latency tolerance, model size, prompting strategies, tuning, retrieval architecture, and evaluation quality. "Equivalent or better" is not a universal condition but an engineering outcome based on specific scenarios. Enterprises still need to conduct independent benchmarking, develop governance plans, and prepare contingency strategies.
Research from McKinsey, cited in the announcement, indicates that nearly three-quarters of organizations expect to increase their use of open-source AI. While open models can reduce vendor lock-in and lower costs, building and operating large-scale computing clusters still requires significant capital investment. Whether Together AI can translate its cost advantages into a lasting infrastructure edge, as GPU supply, energy constraints, and enterprise risk controls tighten simultaneously, will be a true test.
Schneider Electric's venture capital division also participated in this round. CEO Olivier Blum views efficiency as the link between AI and energy infrastructure, noting that more efficient inference impacts not only software margins but also electricity demand and data center planning. The promise of open AI does not automatically equate to a decentralized market; it may give rise to another set of gatekeepers in the infrastructure space.
