Nvidia Backs Neocloud Startups as Market Crowds

NVIDIA argues that AI factories, which cost roughly $60 million per megawatt, maximize returns through three factors: productive systems that deliver high throughput per megawatt and low cost per token, durable hardware that remains economically viable years after deployment, and fungible platforms that run diverse workloads. The company claims its latest Vera Rubin NVL72 systems deliver over 30x higher throughput per megawatt than GB300 NVL72 systems and up to 45x lower cost per million tokens on certain models, while older generations like the A100 GPU from 2020 remain in commercial service six years later.

Flow Engineering, an AI startup focused on applying AI agents to hardware design, raised funding at a $750M valuation. The round was backed by Valor Equity Partners, Atreides Management, and Sequoia Capital. Roelof Botha joined as an angel investor and board member.

DeepSeek announced Wednesday that it will open-source software tools customized for Huawei's AI chips, marking an escalation in the Chinese AI firm's push to rely on domestic semiconductors for model training and deployment. The move includes a Huawei-compatible version of TileLang, a software framework. This effort reflects broader Chinese industry strategy to reduce dependence on foreign chip technology amid U.S. export restrictions.

GMI Cloud, a five-year-old GPU cloud provider specializing in Nvidia chip servers, raised $668 million in combined equity and debt financing. The funding round, which includes participation from Nvidia itself, signals continued investment by the chipmaker in smaller cloud providers that can expand Nvidia's market reach among enterprises. The raise underscores Nvidia's strategy of supporting alternative cloud infrastructure players beyond hyperscalers.