Together AI secures $800M for GPU neocloud as bookings top $1.15B

Together AI is positioning itself as the primary alternative to closed-model giants like OpenAI by lowering the barrier to entry for open-source model deployment. This funding validates the 'neocloud' model, where specialized GPU infrastructure outperforms general-purpose hyperscalers for high-density AI workloads. For the streaming and video industry, where token costs for massive media processing tasks can be prohibitive, this shift toward open-source inference infrastructure provides a pathway to more predictable unit economics. As video-native AI workloads scale, expect the market to favor providers that offer direct GPU cluster management over managed API abstractions. Watch for Nvidia’s continued participation in these rounds as a signal of its effort to diversify its distribution beyond traditional big-tech cloud partners.

The rise of neoclouds like Together AI reflects a broader structural shift in cloud architecture. Per Synergy Research Group in April 2026, neocloud revenues reached $9 billion in Q4 2025 alone, representing a 223% year-over-year increase. Specialized providers like CoreWeave and Lambda Labs are challenging traditional hyperscalers by offering dedicated, high-performance GPU clusters that bypass the overhead and pricing premium associated with generalized computing services. Analysts at Synergy forecast this segment will reach $400 billion by 2031 as demand for GPU-centric computing continues to outstrip the capacity available from legacy cloud providers.

This capital influx is largely fueled by the enterprise transition from closed-source to open-source models. According to reporting from mlFlow in May 2026, open-source AI adoption in large organizations has reached 89%, with deployments showing a 25% higher return on investment than closed-source stacks. Enterprises increasingly favor open-weight models like Meta’s Llama family and Mistral for production environments to avoid vendor lock-in and linear API costs. In this environment, neoclouds provide the 'plumbing' necessary for these companies to own their inference stack, offering latencies under 100ms for small-batch tasks that typically take 500ms or more on proprietary APIs, per data from NeuralPress in June 2026.

Aramco Ventures’ lead role in this Series C also signals a strategic push by Middle Eastern sovereign-backed capital into the foundational layer of the global AI stack. In April 2026, Aramco Ventures CEO Mahdi Aladel stated the firm’s $7.5 billion portfolio is heavily focused on the entirety of the AI stack, including compute, orchestration, and modeling. By backing Together AI alongside established Western firms like Vista Equity Partners, Aramco is not only seeking financial gains but also securing a strategic foothold in the high-density data center infrastructure that will power the next phase of industrial and commercial generative AI.